Compare commits

...
254 Commits
Author SHA1 Message Date
steffen911 fae508d820 chore: release v3.45.2 2025-03-26 09:59:47 +01:00
Steffen SchmitzandGitHub a6e581d349 fix: use correct column name for filtering traces bar chart (#6215) 2025-03-26 08:56:09 +00:00
Max DeichmannandGitHub 9e77738a2f fix: fix by model charts (#6208)
* fix

* push

* push

* push
2025-03-25 20:56:00 +00:00
Marc Klingen d15c0a27f5 chore: release v3.45.1 2025-03-25 19:49:44 +01:00
Marc KlingenandGitHub a8626841b5 chore(cloud): update blobstorage integration entitlement (#6209) 2025-03-25 18:40:57 +00:00
marliessophieandGitHub 5945a93b7d fix(evaluator-config): preview execution of dataset run items (#6206)
* chore(evaluator-config): preview execution of dataset run items

* push
2025-03-25 17:44:11 +00:00
Steffen SchmitzandGitHub 9c447de584 chore: auto-close stale issues (#6207) 2025-03-25 17:21:44 +00:00
Max DeichmannGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
48df0e04de fix: fix model charts (#6202)
* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* Update packages/shared/src/server/repositories/dashboards.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Update web/src/__tests__/async/repositories/dashboard-repository.servertest.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* fix

* fix

* fix

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-25 17:12:50 +00:00
Steffen SchmitzandGitHub 0897ac753b feat: add custom-query backend (#6006)
* feat: add self-serve dashboard backend poc

* chore: data model thoughts

* chore: create initial query builder example and test case

* chore: handle empty dimensions and metrics

* chore: lint

* chore: extend filter conditions to account for timestamps

* chore: refactor query builder into sub-functions

* chore: use template queries for user-supplied values

* chore: add query builder tests

* chore: handle time dimension

* chore: add trpc endpoint to execute custom clickhouse query

* chore: add a query playground component to test a couple of queries

* chore: handle multiple joins

* chore: add observations table to data model

* chore: fix bug in data model

* chore: add test case to compare with old dashboard results

* chore: add scaffold for additional views

* expand the users view and add segments to filter scores subviews

* chore: add test cases for score views

* chore: drop users and sessions for now

* chore: add scores aggregate and observations cost tests

* chore: add sql injection tests

* chore: lint

* chore: fix dashboard test cases

* chore: drop users query test

* chore: add order by logic

* chore: fill timeseries values

* chore: use new query function on dashboard

* chore: convert TracesBarListChart.tsx to new query endpoint

* chore: make chart data compile

* chore: update tests

* chore: add tags to custom queries

* chore: typing

* chore: suffix join condition with sql

* chore: add typing in queryBuidler

* chore: limit playground to cloud admin users

* chore: add util to map legacy dashboard columns to new model

* chore: separate time filter state for TracesBarListChart.tsx

* chore: pass timestamps directly into chart components

* chore: add userId and sessionId on observations

* chore: pick auto time granularity based on hours
2025-03-25 14:54:37 +00:00
Marc KlingenandGitHub d00c63100e chore(cloud): remove rate limits from projects api as it is used by auth_check in sdks (#6193)
* chore(cloud): remove rate limits from projects api as it is used by auth_check in sdks

* address feedback

* improve comment
2025-03-25 14:52:23 +00:00
steffen911 e8970bc11c chore: release v3.45.0 2025-03-25 15:52:33 +01:00
Steffen SchmitzGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
eecb253093 feat: add blob storage integration (#6143)
* feat: add page to configure blob storage integration

* chore: remove blob storage logo and include three config options

* chore: add blob storage integration table

* chore: update router for blobstorage

* chore: placeholder for password and type update

* chore: create scaffolding for blobstorage queue processors

* chore: add scheduling logic for blob storage integration

* chore: prepare setup in processing job

* chore: implement extraction fucntion

* chore: add test case and fix generator setup

* chore: lint

* chore: overwrite default region to empty

* chore: remove details from auditlog and add beta

* Update web/src/pages/project/[projectId]/settings/integrations/blobstorage.tsx

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Update worker/src/__tests__/blobStorageIntegrationProcessing.test.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* chore: correct env name in tests

* Update worker/src/ee/integrations/blobstorage/handleBlobStorageIntegrationProjectJob.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* chore: update doc urls

* chore: open docs in new tab

* chore: ensure consistent block id length azure

* chore: revert test change

* chore: remove dummy error log

* chore: type annotation

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-25 13:39:33 +00:00
marliessophieGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
4e4fbb5101 fix(prompt-experiments): show form error if name is duplicated (#6189)
* fix(prompt-experiments): show form error if name is duplicated

* chore: fix typo

* Update web/src/ee/features/experiments/hooks/useExperimentNameValidation.tsx

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* eslint

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-25 09:07:16 +00:00
marliessophieandGitHub bb4823aa27 fix(evaluator-templates): control editing state (#6187) 2025-03-24 13:29:19 +00:00
marliessophieandGitHub 174d96c100 feat(evals): preview number of historic items to be evaluated (#6116)
* feat(evals): preview number of historic items to be evaluated

* refactor: split up large evaluator form file

* chore: drop optional pagination

* chore: fix eslint

* chore: rename test
2025-03-24 13:05:50 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
8e59b544c9 chore(deps): bump next from 14.2.21 to 14.2.25 (#6175)
Bumps [next](https://github.com/vercel/next.js) from 14.2.21 to 14.2.25.
- [Release notes](https://github.com/vercel/next.js/releases)
- [Changelog](https://github.com/vercel/next.js/blob/canary/release.js)
- [Commits](https://github.com/vercel/next.js/compare/v14.2.21...v14.2.25)

---
updated-dependencies:
- dependency-name: next
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2025-03-24 12:57:26 +00:00
Hassieb PakzadandGitHub cffd84f0e7 chore(azure-llm-connection): update api-version to support reasoning models (#6186) 2025-03-24 14:06:55 +01:00
Hassieb PakzadandGitHub 845faee385 fix(cost-tracking): null values in OpenAI usage schema (#6184) 2025-03-24 13:42:20 +01:00
Hassieb PakzadGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
e45c4af30f chore(prompts): add error logs for resolvePromptGraph (#6181)
* chore(prompts): add error logs for resolvePromptGraph

* Update packages/shared/src/server/services/PromptService/index.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* fix

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-24 11:31:29 +00:00
Marc KlingenandGitHub 616544a48c chore: apply prettier (#6176) 2025-03-24 09:46:13 +00:00
steffen911 75d64b38b3 chore: release v3.44.0 2025-03-24 09:35:15 +01:00
Steffen SchmitzandGitHub 65cdd75282 chore: drop queue backup option in bullmq (#6141)
* chore: drop queue backup option in bullmq

* chore: lint
2025-03-24 08:26:32 +00:00
Steffen SchmitzandGitHub 2744028f4c feat: extract top-level metadata from langfuse.metadata (#6173)
* feat: extract top-level metadata from langfuse.metadata

* chore: extract metadata based on prefix

* chore: add support for resourceAttributes
2025-03-24 08:09:04 +00:00
Max DeichmannandGitHub 3c02c54de2 fix: fix model paring from redis (#6157)
* push

* fix

* push

* fix

* fix
2025-03-21 18:42:06 +00:00
marliessophieandGitHub 974d2c1007 fix(ui): reference errors in prompt detail screen (#6126)
* chore(ui): convert `action-button` to use forwardRef for compatibility w/ buttons

* fix(ui): ensure prompt item ref points to `command-item`
2025-03-21 17:07:21 +00:00
Max DeichmannandGitHub 008045c771 perf: cache model not found (#6148)
* push

* push

* fix
2025-03-21 14:57:17 +00:00
Marc KlingenandGitHub c6790c6c0e chore(ui): do not auto-generate new api keys in setup-tracing step (#6149)
* chore(ui): do not auto-generate new api keys in setup-tracing step

* fix
2025-03-21 14:17:01 +00:00
Marc KlingenandGitHub 909445ce5b fix(prompts): trim prompt names (#6145)
* fix(prompts): trim prompt names

* push
2025-03-21 13:38:04 +00:00
Max DeichmannandGitHub f4ed4a2635 chore: improve model match observability (#6147)
push
2025-03-21 13:18:39 +00:00
Max DeichmannandGitHub d6206b5269 chore: increase model match observability (#6144) 2025-03-21 13:45:36 +01:00
Max DeichmannandGitHub 62fa43aa78 feat: cache LLM models in Redis (#6129)
* push

* fix

* fix

* fix

* push

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* push

* push

* push

* fix

* fix

* fix

* fix

* push
2025-03-21 11:26:11 +00:00
Steffen SchmitzandGitHub 3c9239a4ab chore: update .env.prod.example (#6140) 2025-03-21 10:24:48 +00:00
marliessophieandGitHub 476cddb4e1 chore(ui): clamp page-header title to one line (#6139) 2025-03-21 10:01:45 +00:00
Steffen SchmitzandGitHub a6c1678da8 feat: add additional input/output parsing for pydantic via OTel (#6138) 2025-03-21 10:00:33 +00:00
Steffen SchmitzandGitHub 129805ebde fix: safely parse big numbers in api responses (#6136)
* fix: safely parse big numbers in api responses

* feat: add additional input/output parsing for pydantic via OTel

* Revert "feat: add additional input/output parsing for pydantic via OTel"

This reverts commit 71d0f3abd951702393aac7ab79a875b888f089cf.
2025-03-21 09:30:24 +00:00
marliessophieandGitHub 59eafb42c3 fix(ui): require annotation queue selection for bulk action (#6127) 2025-03-20 21:08:57 +00:00
marliessophieandGitHub e4695ac717 fix(ui): dom nesting error data-table-row-height-switch (#6125) 2025-03-20 20:12:52 +00:00
marliessophieandGitHub eebd6ea0e9 fix(ui): dom nesting error page-header (#6124) 2025-03-20 20:06:05 +00:00
Marc KlingenandGitHub b4a8b08d3d chore(cloud): add posthog for projects and orgs (#6117)
* chore(cloud): add posthog for projects and orgs

* push
2025-03-20 18:07:17 +00:00
Max DeichmannandGitHub ed44b663ff chore: add dd env variables to github build command (#6119) 2025-03-20 17:36:48 +01:00
Marc KlingenandGitHub 720ebe861e chore(cloud): cache response of cloudStatus api for 5 minutes (#6115)
chore(cloud): keep response of cloudStatus api cached across views for 5 minutes
2025-03-20 14:25:54 +00:00
Max DeichmannandGitHub 9a6f8fd75d refactor: refactor naming of database batch exports (#6111)
* refactor: refactor naming of database batch exports

* refactor: refactor naming of database batch exports
2025-03-20 12:56:38 +00:00
Steffen SchmitzandGitHub f0da1cfa30 chore: reduce cloudwatch flush frequency and batch metrics (#6108)
* chore: reduce cloudwatch flush frequency and batch metrics

* chore: simplify metrics reset

* chore: keep 30s flush

* chore: correct CW metrics
2025-03-20 12:05:07 +00:00
Marc KlingenandGitHub 0c9ebb3da7 feat(cloud): show status page menu item during incidents (#6100)
* feat(cloud): show status page menu item during incidents

* add to support menu
2025-03-20 09:59:49 +00:00
Hassieb PakzadandGitHub 3d16d1ac57 fix(model-params-ui): duplicate model names leading to rendering issues (#6107) 2025-03-20 10:40:59 +01:00
Marc KlingenandGitHub fa462a8bc4 chore(ui): improve padding in peek view header (#6099) 2025-03-19 18:59:29 +00:00
David Alexander PfeifferandGitHub 695b65fe0d chore(tests): fully unpin node version from temporary fix #6063 (#6096) 2025-03-19 15:38:36 +00:00
Hassieb PakzadGitHubjake goldenellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
955388e846 feat(llm-connections): add atla adapter (#6088)
* feat: enable Atla integration (#6022)

* push

* push

* Update packages/shared/src/server/llm/fetchLLMCompletion.ts

* Update web/src/features/public-api/components/CreateLLMApiKeyForm.tsx

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Revert "Update web/src/features/public-api/components/CreateLLMApiKeyForm.tsx"

This reverts commit efb2d3b359a85680224876481421d31bfbd44fd3.

---------

Co-authored-by: jake golden <jackson.golden@gmail.com>
Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-19 12:36:12 +00:00
David Alexander PfeifferandGitHub c6864ed8d9 fix(tests): disable --experimental-require-module (default since Node 20.19.0) since it breaks jest-based testing (#6075)
* pin node to problematic version for working on a fix

* added cross-env and disabled the feature flag causing the testing issue - long term fix should probably be switching to vitest

* unpin node version since problematic feature is now disabled fo affected tests
2025-03-19 11:04:59 +00:00
marliessophieandGitHub 917516fa91 chore(ui): control for more specific onClick actions in table with peek view (#6086) 2025-03-19 10:25:03 +00:00
David Alexander PfeifferandGitHub 19017a6709 fix(ui): OpenAiMessageView now renders messages where content is an empty string instead of ignoring them (#6068)
* fix(ui): IOPreview now renders messages where content is an empty string instead of ignoring them

* removed debug logs

* OpenAiMessageView first filters valid messages to render and then executes the "show x more ..." logic, this way we ensure that there can never be a mismatch between messages items and the amount of fields we support displaying

* Render empty string quotation marks if content is empty but not null or undefined

* Change length check to falsy check for conditional display of empty quotes if content is empty

* cleaner falsy check for rendering empty string quotation marks
2025-03-19 09:41:32 +00:00
marliessophieandGitHub ca5e6569f9 fix(peek-view): ensure onClickRow action smoothly keeps peek view open; onClick of select cells independent of peek view state (#6079)
* fix(ui): stop propagation on traces table checkbox

* fix(peek-ui): handle onRowClick action correctly for peek view

* feat(peek): do not close upon any screen button or checkbox clicks

* push
2025-03-19 00:41:36 +00:00
marliessophieandGitHub 101942b98b fix(ui): padding on single prompt page (#6070) 2025-03-18 18:35:32 +00:00
marliessophieandGitHub 1188a21b02 feat(traces): add peek view on traces table (#6048)
* feat: adjust table to new style

* style: filter and column options

* fixup: add peek view

* feat: add detail navigation

* chore: remove duplicated code

* chore: style settings tables

* chore: pin first column on traces table

* style: adjust animation speed of peek view

* fix: padding on score configs table

* style: onClick cursor in table

* fix: eslint

* style

* chore: fix toggle for timeline

* push

* style: add padding
2025-03-18 17:25:52 +00:00
Marc KlingenandGitHub ec72dd840d chore(ui): render min-level hint in a single line (#6051) 2025-03-18 14:55:54 +00:00
steffen911 c2d800ff0e chore: release v3.43.0 2025-03-18 15:15:41 +01:00
Steffen SchmitzandGitHub dcf159ef82 chore: lock dd-trace to 5.36.0 (#6064) 2025-03-18 14:13:30 +00:00
Marc KlingenandGitHub 998984b9ed chore(cloud): improved pricing/packaging (#5664)
* wip

* simplify

* actionbutton for upgrade prompt

* count events for free plan limit

* add sso settings

* rename lite to core

* add product id

* nit

* nits

* fix rate limit test

* ci run on teams plan

* fix

* fix names of pipeline steps

* test tmp

* add better error logs to makeAPICall

* push

* clone to throw better errors

* improve error

* improve error

* fix

* remove comment

* refactor

* push

* next attempt

* nit

* create new auth for failing test suites

* push

* fix rate limit test
2025-03-18 13:47:03 +00:00
Marc KlingenandGitHub 022b234e57 chore: pin node 20.18.3 to avoid issue introduced by 20.19.0 (#6063)
* test commit

* chore: pin node 20.18.3 to avoid issue introduced by 20.19.0

* docker

* revert test commit

* push
2025-03-18 13:30:58 +00:00
b8777aa5d1 chore: better error message when uploading existing dataset item (#6053)
chore: better docstring when uploading existing dataset item to different dataset

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-18 13:07:11 +00:00
Steffen SchmitzandGitHub 2e84feeab7 feat: support external S3 endpoint for batch exports (#6055)
* feat: add media and export external endpoint config

* chore: add docs string

* chore: add external endpoint support for azure

* chore: drop feature for media
2025-03-18 10:57:14 +00:00
Steffen SchmitzandGitHub a508485f88 chore: enable datadog profiling for web (#6020)
* chore: enable datadog profiling for web

* chore: update dockerfile
2025-03-18 07:26:53 +00:00
marliessophieandGitHub d8bb283a44 fix(ui): make models detail page scrollable (#6050) 2025-03-17 19:48:27 +00:00
Steffen SchmitzandGitHub 75ec05e206 Delete .devcontainer/devcontainer.json 2025-03-17 15:01:41 +01:00
Steffen SchmitzandGitHub 19f710abe2 chore: add devcontainer config 2025-03-17 14:56:35 +01:00
David Alexander PfeifferandGitHub ad2d6bee66 chore(ui): remove padding of "Additional Input" field on OpenAiMessageView (#6037) 2025-03-17 10:06:32 +00:00
David Alexander PfeifferandGitHub 40d0acae5f chore(ui): alignment and sizing consitency in ObservationTree / GroupedScoreBadge (#6031)
* Unified design of grouped score badge with trace detail score badges, except that text is no longer medium/semibold

* Made padding of root item of ObservationTree consistent with other items
2025-03-17 09:58:23 +00:00
Marc KlingenandGitHub 4d387cc35c chore(ui): standardize the prompt variable list preview across views (#6027)
* chore(ui): standardize the prompt variable list preview across views

* use ctx prisma
2025-03-17 09:34:44 +00:00
690885c900 fix(ui): fixed scrollbar appearing on IO preview mode switcher in trace and observation details (#5995)
* fixed scrollbar appearing on IO preview mode switcher in trace and observation details

* fix

---------

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-14 19:32:46 +00:00
David Alexander PfeifferandGitHub ef2baa94be fix(ui): fixed misalignment of badges in trace and observation details (#5994)
* fixed alignment issues of badges wrapped in links in trace and observation detail views

* Fixed tooltip of aggregated usage data badge (in trace details) rendering even if there is no underlying data
2025-03-14 19:19:14 +00:00
db6b442832 feat: Add environment tag on trace / observation detail views (#5985)
* feat: Add environment tag on trace / observation detail views

* shorten label

---------

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-14 19:17:19 +00:00
Hassieb Pakzad 8231ebae4a chore: release v3.42.1 2025-03-14 17:05:33 +01:00
Marc KlingenandGitHub 326e8d3189 chore(ui): always show tag buttons in table cells instead of only on hover (#6025) 2025-03-14 15:23:32 +00:00
Hassieb PakzadandGitHub d2001d110d chore(docker-compose): bind services to localhost (#6023) 2025-03-14 14:16:36 +01:00
Hassieb PakzadandGitHub 3f81c6b9d5 feat(cost-tracking): add OpenAIResponseUsageSchema support (#6003) 2025-03-14 12:08:28 +00:00
Marc KlingenandGitHub 991f837ee5 fix(ui): hide jump to playground button when not entitled (#6014) 2025-03-14 00:01:40 +00:00
Marc Klingen 062a7f27b8 chore: release v3.42.0 2025-03-13 19:23:52 +01:00
Steffen SchmitzandGitHub 924b326db6 chore: remove unused prisma query builder (#6005) 2025-03-13 17:21:23 +00:00
Andres CarrilloandGitHub 2780e66c13 feat(auth): add custom Gitlab SSO URL option (#5984) (#6004) 2025-03-13 17:50:16 +01:00
Marc KlingenandGitHub c2f51b8d2d feat(ui): add csv/json export for scores table (#5949)
* feat(ui): add csv/json export for scores table

* Merge branch 'main' into marc/lfe-3466-export-scores

* push

* add test

* fix cols

* clean up diff

* fix
2025-03-13 16:23:30 +00:00
Steffen SchmitzandGitHub 3a3021a176 chore: reduce host port mapping for internal docker compose components (#6002) 2025-03-13 15:36:34 +00:00
Marc KlingenandGitHub 5dd5a39f90 feat(api): add api for annotation queues (#5945)
* feat(api): add api for annotation queues

* fix

* push

* strict zod interfaces

* fix

* fix spec and types

* add additional tests
2025-03-13 15:24:19 +00:00
Steffen SchmitzandGitHub 9ee4544e1d docs: add security notes to docker compose deployment (#5999) 2025-03-13 14:27:10 +00:00
Hassieb PakzadandGitHub 136f8a3e83 fix(llm-connections): show bedrock error (#6000) 2025-03-13 15:06:42 +01:00
Hassieb Pakzad eecf0d680b chore: release v3.41.1 2025-03-13 11:07:07 +01:00
Hassieb PakzadandGitHub d50dbed797 feat(llm-connections): allow extraHeaders for azure (#5996) 2025-03-13 09:46:38 +00:00
Steffen SchmitzandGitHub bb1dbdf824 chore: reproduce clickhouse missing surrogate key error (#5980)
* chore: reproduce clickhouse missing surrogate key error

* chore: add fix candidate

* chore: adjust test cases

* chore: adjust test case
2025-03-13 07:47:31 +00:00
Hassieb Pakzad 0991d44525 chore: release v3.41.0 2025-03-12 18:28:25 +01:00
Max DeichmannandGitHub 449c9360ac fix: search past observations with correct timestamp in evals (#5986)
fix: use readonly for the column definitions
2025-03-12 16:41:14 +00:00
Hassieb PakzadandGitHub cc4a88d6da fix(prompts): block label deletion if dependent exists (#5983)
* fix(prompts): block label deletion if dependent exists

* fix

* Update web/src/features/prompts/server/actions/updatePrompts.ts
2025-03-12 16:05:46 +00:00
Steffen SchmitzandGitHub 11b2411a5b chore: prevent sorting by environment in UI tables (#5978) 2025-03-12 13:45:25 +00:00
d8a5ebc42e chore(cloud): Make UI recognize cloud:enterprise plan and add corresponding access rights and rate limits (#5975)
* chore(cloud): Make UI recognize cloud:enterprise plan and add corresponding access rights and rate limits

- label on project selector
- rate limits same as team
- access rights same as team
- no stripe product id for this (yet), cannot be purchased self-serve in product

* remove comments

---------

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-12 13:44:01 +00:00
Steffen SchmitzandGitHub 9042b005b8 feat: extract logfire.msg as span and trace name (#5977) 2025-03-12 13:23:33 +00:00
Hassieb PakzadandGitHub 7b70b5b678 feat(prompts): add composability (#5882) 2025-03-12 13:26:00 +01:00
Marc KlingenandGitHub c3ce043d47 chore: remove unnecessary comment (#5961) 2025-03-11 13:20:18 +00:00
Marc KlingenandGitHub 8c5737ebea chore: rename batch exports from generation to observation as table was generalized (#5950)
* chore: rename batch exports from generation to observation as table was generalized

* fix test
2025-03-11 13:15:36 +00:00
steffen911 cb928bf3fc chore: release v3.40.0 2025-03-11 13:49:18 +01:00
Steffen SchmitzandGitHub dc8a997324 feat: add score deletion to UI (#5956)
* feat: enable score deletion via Langfuse UI

* chore: add test cases

* chore: typo

* chore: implement deleteAnnotationScore

* chore: lint

* feat: add score deletion to UI

* chore: add trpc tests

* chore: allow multi-score deletion from table view

* chore: typoes and mock resets

* chore: update mock handling
2025-03-11 12:44:17 +00:00
Steffen SchmitzandGitHub 31da4bc5a5 chore: refactor score deletions into async queue (#5939)
* feat: enable score deletion via Langfuse UI

* chore: add test cases

* chore: typo

* chore: implement deleteAnnotationScore

* chore: lint
2025-03-11 08:02:13 +00:00
Max DeichmannandGitHub c0189eecad fix: use readonly for the column definitions (#5948)
* fix: use readonly for the column definitions

* fix: use readonly for the column definitions
2025-03-10 22:13:28 +00:00
Marc KlingenandGitHub a3f85e75dc docs: add zod type strict() to public api cursor rule (#5947) 2025-03-10 21:56:41 +00:00
Marc KlingenandGitHub 88ee2f3823 docs: improve public api cursor rule (#5946) 2025-03-10 21:35:38 +00:00
Max DeichmannandGitHub 0b485ad5d8 fix: fix eval status on details page (#5944)
* fix: fix eval status on details page

* push

* fix: fix eval status on details page

* fix: fix eval status on details page
2025-03-10 20:18:04 +00:00
Marc KlingenandGitHub 4e95550411 feat(datasets): show run name on runitems table (#5943)
* feat(datasets): show run name on runitems table

* perf

* push
2025-03-10 19:50:00 +00:00
Marc KlingenandGitHub ba9e26181b chore: remove models page notification (#5942) 2025-03-10 19:28:48 +00:00
Marc KlingenandGitHub 289a7210ee chore(cloud): add more data masking for posthog (#5941) 2025-03-10 19:27:36 +00:00
Max Deichmann fe0c4dc98d chore: release v3.39.0 2025-03-10 17:09:30 +01:00
Hassieb PakzadandGitHub 61c2ec268a perf(playground): memoizate ChatMessages (#5937)
* perf(playground): memoizate ChatMessages

* push

* add callbacks
2025-03-10 16:07:28 +00:00
Max DeichmannandGitHub 1df2781b59 fix: correctly extract jsonpath form observations for evals (#5938)
* fix: correctly extract jsonpath form observations for evals

* fix: correctly extract jsonpath form observations for evals
2025-03-10 15:42:44 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
5478e870b3 chore(deps): bump axios from 1.7.7 to 1.8.2 (#5915)
Bumps [axios](https://github.com/axios/axios) from 1.7.7 to 1.8.2.
- [Release notes](https://github.com/axios/axios/releases)
- [Changelog](https://github.com/axios/axios/blob/v1.x/CHANGELOG.md)
- [Commits](https://github.com/axios/axios/compare/v1.7.7...v1.8.2)

---
updated-dependencies:
- dependency-name: axios
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2025-03-10 14:49:50 +00:00
Max DeichmannandGitHub 80c298fa1e chore: limit event id length (#5934) 2025-03-10 14:48:51 +00:00
Steffen SchmitzandGitHub 330cca6978 feat: add trace delete API endpoints (#5931)
* feat: add trace delete API endpoints

* chore: add api docs

* chore: update docs

* chore: typo

* chore: update audit logs

* chore: lint

* chore: update fern docs

* chore: fix return type
2025-03-10 14:13:33 +00:00
Steffen SchmitzandGitHub 286b78a1fe feat: add api key actions to auditlog table (#5932)
* feat: add api key actions to auditlog table

* chore: propagate audit logs to all relevant API routes

* chore: lint
2025-03-10 12:35:10 +00:00
Marc KlingenGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
f0c37c8309 chore: add initial cursor rules (#5926)
* chore: add some initial cursor rules

* Update .cursor/rules/frontend-features.mdc

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Update .cursor/rules/public-api.mdc

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-03-09 22:54:01 +00:00
Marc KlingenandGitHub 326b27cdda fix: ui config in localstorage should persist signing out (#5916)
* fix: ui config in localstorage should persist signing out

* push

* push
2025-03-07 17:50:04 +00:00
Marc KlingenandGitHub 886e352c39 fix: select/highlight text in Compare Dataset view (#5914)
* fix: select/highlight text in Compare Dataset view

* clean diff
2025-03-07 15:59:53 +00:00
steffen911 1cd76a49f1 chore: release v3.38.0 2025-03-07 16:48:21 +01:00
Marc KlingenandGitHub 0639b78468 feat(datasets): add multi-delete for dataset runs table (#5912) 2025-03-07 15:29:50 +00:00
Steffen SchmitzandGitHub 2d6e51e219 chore: test processing of special characters in S3 (#5907)
* chore: test processing of special characters in S3

* chore: reject carriage-return character in IDs

* Update ingestion-api.servertest.ts
2025-03-07 15:20:11 +00:00
Yohan GonçalvesandGitHub 1ed6358738 feat: add customizable claim mapping for SSO (#5858) 2025-03-07 16:19:51 +01:00
4fc339dbac chore: remove redundant readme files (#5900)
chore: remoe redundant readme files

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-06 19:55:34 +00:00
Steffen SchmitzandGitHub 1e64d97b75 feat: create new environment materialized views for quick lookups (#5895)
* feat: create new environment materialized views for quick lookups

* chore: update repository

* chore: lint

* chore: add final to environments query

* chore: always include default in array
2025-03-06 19:09:28 +00:00
Steffen SchmitzandGitHub 4d06482d78 fix: enforce ints on usageDetails (#5889)
* fix: enforce ints on usageDetails

* chore: make test actually test a  generation

* chore: lint
2025-03-06 17:05:10 +00:00
Marc KlingenandGitHub 603e5ca7c9 chore(ui): add env label on env multi-select (#5890) 2025-03-06 16:37:22 +00:00
Steffen SchmitzandGitHub 61c6560f24 chore: return 200 instead of 207 as OTel success status (#5883) 2025-03-06 14:58:40 +00:00
Steffen SchmitzandGitHub 6bff022a88 chore: log dropped data items correctly in clickhouse writer (#5876)
* chore: log dropped data items correctly in clickhouse writer

* chore: update test case
2025-03-06 14:43:11 +00:00
Steffen SchmitzandGitHub 1f7cac2dd8 feat: add environment column to UI tables (#5877)
* feat: add environment column to UI tables

* feat: add environment on sessions table

* chore: add to observations

* chore: add to scores

* chore: add to users table

* chore: lint
2025-03-06 14:08:41 +00:00
Hassieb PakzadandGitHub 49bccb8fa0 fix(ui-graph-view): handle single top most node (#5881) 2025-03-06 15:15:03 +01:00
Marc KlingenandGitHub fea8a3d8d1 chore: add noindex header (#5880) 2025-03-06 13:46:00 +00:00
Max DeichmannandGitHub 08fcdb68be chore: fix contributingmd file (#5879) 2025-03-06 14:34:46 +01:00
f18a120bd8 feat: add environment filter (#5765)
* feat: add environment filter on dashboard

* chore: add a useEffect and dashboard columns for filtering

* chore: add dashboard filters

* fixup: add environment filter hook and barebone code

* chore: add traces table filter

* chore: add sessions environment filter

* chore: do observations, scores, users

* chore make filter bar dynamic and change UI placement

* chore: map environment column specially in dashboard

* chore: add column

* chore: lint

* chore: remove overflow on multi-select

* chore: only load downstream calls once environment is available

* chore: only use isLoading condition

---------

Co-authored-by: Marlies Mayerhofer <74332854+marliessophie@users.noreply.github.com>
2025-03-06 10:45:18 +00:00
Max DeichmannandGitHub 2b3ca85344 chore: fix contributingmd file (#5870)
* chore: fix contributingmd file

* chore: fix contributingmd file

* chore: fix contributingmd file

* chore: fix contributingmd file

* chore: fix contributingmd file
2025-03-06 08:03:39 +00:00
Hassieb PakzadandGitHub 8b408858c4 feat(batch-exports): increase row limit to 1.5M (#5872) 2025-03-06 08:47:56 +01:00
89ad92c0ff chore: consistent casing of README files (#5865)
* docs: fix broken link in CONTRIBUTING.md

* chore: rename files to fix casing

---------

Co-authored-by: Marc Klingen <git@marcklingen.com>
2025-03-05 23:23:54 +00:00
Marc Klingen 16e1d19917 chore: release v3.37.0 2025-03-05 23:51:35 +01:00
Marc KlingenandGitHub 18d5ba1bfa feat(auth): add AUTH_WORKOS_ORGANIZATION_ID and AUTH_WORKOS_CONNECTION_ID envs (#5866) 2025-03-05 21:31:45 +00:00
Steffen SchmitzandGitHub 95d6b9fe13 chore: add name mappings to sync prisma schema with reality (#5863) 2025-03-05 16:56:47 +00:00
Max DeichmannandGitHub 23183a71c1 chore: add missing clickhouse writer metrics (#5851)
* chore: add missing clickhouse writer metrics

* chore: add missing clickhouse writer metrics

* Update worker/src/services/ClickhouseWriter/index.ts
2025-03-05 16:37:00 +00:00
5eb0a275a5 feat: add S3_CONCURRENT_READS and S3_CONCURRENT_WRITES config options (#5824)
Co-authored-by: Steffen Schmitz <steffen@langfuse.com>
2025-03-05 16:58:26 +01:00
Steffen SchmitzandGitHub 4c5366fa20 perf: reduce trace scans for scores counts (#5860)
* perf: reduce trace scans for sessions counts

* chore: add countAll test

* chore: add session for testing
2025-03-05 15:18:10 +00:00
Marc Klingen 4ebaf0a24a chore: release v3.36.0 2025-03-05 13:32:42 +01:00
Marc KlingenandGitHub b98b90e635 feat(auth): add workos as idp (#5828)
* feat(auth): add workos as idp

* add button for org and connection based sign in

* fix

* remove client auth method and checks

* remove unnecessary diff
2025-03-05 12:29:26 +00:00
Marc Klingen 36f11a645f chore: release v3.35.1 2025-03-04 22:32:19 +01:00
Marc KlingenandGitHub 3c8b26aa79 chore(api): update api reference and postman collection (#5842) 2025-03-04 22:31:38 +01:00
Marc Klingen 86840e4119 chore: release v3.35.0 2025-03-04 22:23:24 +01:00
Steffen SchmitzandGitHub 4b412fb6ca feat: map pydantic logfire events to langfuse datamodel (#5841)
* feat: map pydantic logfire events to langfuse datamodel

* chore: map events to empty if parsing fails
2025-03-04 20:58:42 +00:00
Steffen SchmitzandGitHub 271f75aa17 chore: reduce default batch size of background migrations (#5838)
* chore: reduce default batch size of background migrations

* chore: update the cli defaults
2025-03-04 15:58:11 +00:00
Marc KlingenandGitHub 669ad4cb92 feat(datasets): deletion of dataset runs via api (#5834)
* public api

* improve public api
2025-03-04 13:41:27 +00:00
Steffen SchmitzandGitHub 9b237d4041 fix: detect openinference LLM calls as generations (#5833)
feat: detect openinference LLM calls as generations
2025-03-04 11:13:42 +00:00
marliessophieandGitHub 15b4ca01ec fix(ui): overlapping status message on trace detail page (#5830) 2025-03-04 09:07:29 +00:00
Marc KlingenandGitHub 196b5319bf feat(datasets): allow for deletion of dataset items in ui and public API (#5816)
* push

* move into drop down menu

* fix

* move copy to the right side

* fern generate

* use trpc errors
2025-03-03 14:03:18 +00:00
Steffen SchmitzandGitHub d88e0456b8 perf: reduce memory consumption on observations query (#5811)
* perf: reduce memory consumption on observations query

* chore: drop skip index list for lint
2025-03-03 13:35:40 +00:00
Marc KlingenandGitHub 78463a9ce3 feat(datasets): copy items between datasets (#5800)
* wip

* push

* push
2025-03-03 12:51:06 +00:00
marliessophieandGitHub 0bd7219eba style(ui): trace view alignment improvements (#5789)
* style(ui): trace view alignment improvements

* fix: test
2025-03-03 12:25:15 +00:00
Steffen SchmitzandGitHub a307270d8c chore: upgrade fern and add java sdk example (#5814) 2025-03-03 11:09:41 +00:00
Max DeichmannandGitHub f1555190c1 fix: use aes-gcm 12 bytes for IV (#5764)
* fix: use aes-gcm 12 bytes for IV

* fix: use aes-gcm 12 bytes for IV

* fix: use aes-gcm 12 bytes for IV
2025-03-03 10:32:22 +00:00
Max DeichmannandGitHub 56c77e0de8 chore: reduce log level for frequent logs (#5804) 2025-03-02 17:54:07 +00:00
Marc KlingenandGitHub 48cbbb694f feat(ui): add LLM connections settings page separate from api keys (#5798) 2025-03-01 19:04:33 +00:00
Marc KlingenandGitHub 168dfd2caa feat(datasets): add archive/unarchive functionality for dataset items on single item view (#5797)
* feat(datasets): add archive/unarchive functionality for dataset items on single item view

* imrpove copy
2025-03-01 19:04:23 +00:00
Marc KlingenandGitHub ae66dd9e6f fix(ui): add playsInline attribute to video players (#5795) 2025-03-01 18:33:46 +00:00
Marc KlingenandGitHub d081aab38f feat(datasets): support adding dataset items to multiple datasets (#5796)
* feat(datasets): support adding dataset items to multiple datasets

* feat(datasets): optimize bulk dataset item creation with createMany
2025-03-01 18:09:46 +00:00
Marc KlingenandGitHub a86f0ab263 fix(cloud): only active evals should count towards plan limit (#5793) 2025-03-01 12:42:39 +00:00
Marc KlingenandGitHub 738df01ab7 feat(ui): add settings deep links to cmd+k (#5680)
* proejct settings

* push

* add org and small fixes
2025-03-01 10:12:04 +00:00
marliessophieandGitHub 549eac6fd3 fix(ui): overflow behaviour on trace timeline (#5792) 2025-03-01 07:50:55 +00:00
Marc KlingenandGitHub 6d125ec274 chore(ui): remove redundant className props from splash screens (#5791) 2025-03-01 00:44:46 +00:00
Marc KlingenandGitHub 4f54577f83 chore(ui): improve perf of sessions hasAny (#5783)
chore: improve performance of sessions hasAny
2025-03-01 00:31:29 +00:00
Steffen SchmitzandGitHub 1ec1bdf849 feat: add environment properties on api routes (#5781)
* feat: add environment properties on api routes

* chore: add env filter to traces

* chore: add observation env and daily metrics env filters

* chore: add env to score endpoints

* chore: filter sessions by environment

* chore: filter traces by environment

* chore: expand trace test assertions

* chore: lint
2025-02-28 16:12:18 +00:00
Hassieb PakzadandGitHub ed5a53b10e fix(ui): llm api key form model name delete (#5786) 2025-02-28 15:57:18 +00:00
marliessophieandGitHub 5b1d9b1dd1 style(page-header): margins (#5784) 2025-02-28 14:42:19 +00:00
marliessophieandGitHub adf4508ffd feat(traces): redesign single trace UI (#5718)
* feat(traces-ui): refactor preview for traces and observations

* fix: typo in use prompt code

* feat(ui): add redesigned timeline

* feat: support all settings on timeline view

* feat: support langgraph trace

* style: langgraph view

* chore: fix other usages of single trace components

* style: comment icon color

* fix: eslint

* style: optimise for reponsive design

* chore: refactor common functions to helpers

* chore(globals): remove comments

* style: expand clickable area; adjust search string

* style: small improvements

* fix: styling of trace preview page

* fix: eslint

* fix: remove duplicated cost in observation preview

* fix: session IO formatting

* style

* style

* style: final margins
2025-02-28 13:37:07 +00:00
Steffen SchmitzandGitHub 45f8b2e97b chore: skip tracing for LLM Api key routes (#5780) 2025-02-28 10:49:40 +00:00
marliessophieandGitHub 39c49b65a4 fix(ui): overlap on dataset runs screen (#5778) 2025-02-28 10:15:26 +00:00
steffen911 d905623cb3 chore: release v3.34.1 2025-02-28 11:11:43 +01:00
Steffen SchmitzandGitHub a87a3505b8 chore: add env.HOSTNAME support for worker container (#5779) 2025-02-28 10:06:41 +00:00
Steffen SchmitzGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
591de20a72 feat: simplify otel input/output and accept string token values (#5777)
* feat: simplify otel input/output and accept string token values

* chore: typo

* Update web/src/features/otel/server/index.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* chore: also beautify event attribute fallback

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-02-28 09:49:49 +00:00
Marc Klingen 33d3ab80cb chore: release v3.34.0 2025-02-27 22:54:18 +01:00
Marc KlingenandGitHub 6b9773caec feat(models): add openai gpt-4.5-preview (#5772) 2025-02-27 21:47:24 +00:00
Steffen SchmitzandGitHub c90b477d40 fix: disable tracing for resetPassword route (#5763) 2025-02-27 15:36:36 +00:00
Max DeichmannGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
0f448ccdde fix: always return 422 for signup error handler (#5761)
* fix: always return 422 for signup error handler

* Update web/src/features/auth-credentials/server/signupApiHandler.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Update web/src/features/auth-credentials/server/signupApiHandler.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-02-27 15:33:41 +00:00
marliessophieandGitHub 74f93aad79 chore: move dashboard button to the right, unify eval headers (#5762)
chore: move dashboard button to the right
2025-02-27 15:15:40 +00:00
Max DeichmannandGitHub 9650e87f64 fix: standardize way to get sso blocked domains (#5760)
fix: standardise way to get sso blocked domains
2025-02-27 15:15:10 +00:00
Steffen SchmitzandGitHub c5fc30a9d0 feat: map otel environment config to langfuse environment (#5757) 2025-02-27 14:46:43 +00:00
steffen911 4a26f47971 chore: release v3.33.1 2025-02-27 15:34:24 +01:00
Steffen SchmitzGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
e3a9e61d07 perf: filter relevant observation for data set run item stats (#5752)
* perf: filter relevant observation for data set run item stats

* Update web/src/features/datasets/server/service.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-02-27 14:08:07 +00:00
Hassieb Pakzad a12c2a3a97 fix(model-prices): bedrock claude 3-7 version 2025-02-27 13:14:05 +01:00
Hassieb PakzadandGitHub 80681e08af fix(model-prices): bedrock claude sonnet 3-7 version (#5753) 2025-02-27 13:10:08 +01:00
steffen911 4e97494390 chore: release v3.33.0 2025-02-27 11:58:10 +01:00
Steffen SchmitzandGitHub 346e5fe7b6 feat: map environment to score within eval service (#5749)
* feat: map environment to score within eval service

* chore: adjust tests

* chore: drop unused imports
2025-02-27 10:50:36 +00:00
Steffen SchmitzandGitHub 16482f588a fix: create traces via otel for empty parent span buffer (#5748)
* fix: create traces via otel for empty parent span buffer

* chore: test range for session duration
2025-02-27 10:10:51 +00:00
Marc KlingenandGitHub f047e21e13 chore: improved domain handling and early return in signupApiHandler (#5744) 2025-02-26 21:56:20 +00:00
marliessophieandGitHub 534e832c13 fix(ui): overlapping variables for chat prompt (#5740) 2025-02-26 17:47:03 +00:00
marliessophieandGitHub 5dbfec24f0 fix(ui): experiment from the prompt screen -> window does not fit screen (#5738) 2025-02-26 17:21:05 +00:00
Hassieb PakzadandGitHub 700705be84 chore(llm-keys): update bedrock form description (#5737) 2025-02-26 17:58:49 +01:00
Hassieb Pakzad 611c31ce54 chore: release v3.32.1 2025-02-26 17:06:23 +01:00
Hassieb PakzadandGitHub bdb5ddadd9 fix(ingestion): environment to immutable keys (#5736) 2025-02-26 16:01:26 +00:00
Hassieb PakzadandGitHub 0fd1f065bd fix(api): add environment to ingestion spec (#5735) 2025-02-26 16:02:49 +01:00
Marc KlingenandGitHub 8aa063303a chore(ui): move "configure tracing" to the right in dashboard header (#5731) 2025-02-26 13:59:21 +00:00
Marc KlingenandGitHub d0ff8a87fe chore(ui): minor improvements to onboarding pages (#5730)
chore: minor improvements to onboarding pages
2025-02-26 13:56:39 +00:00
marliessophieandGitHub 0e61ffd621 style(prompts): always show edit label button for prompt versions (#5729) 2025-02-26 13:41:01 +00:00
steffen911 9f20926bc4 chore: release v3.32.0 2025-02-26 14:36:38 +01:00
Steffen SchmitzandGitHub 81a232e4a4 fix: fallback to "default" for missing environment (#5728) 2025-02-26 13:34:33 +00:00
Marc KlingenandGitHub c7e7d62f47 feat(ui): add onboarding screens for core features (#5706)
* feat: add in-ui-onboarding for tracing

* feat(ui): add onboarding screens for core features

* add annotation queues

* move videos to static, add eval onboarding

* push

* push

* push

* push

* push
2025-02-26 13:13:43 +00:00
steffen911 2aa8971dff chore: release v3.31.0 2025-02-26 13:53:58 +01:00
Steffen SchmitzandGitHub 85f4cb0c13 fix: show retention settings for self-hosters (#5727) 2025-02-26 12:34:14 +00:00
Hassieb PakzadandGitHub 4bbba9d983 feat(api): add environment property to traces, scores, obs (#5724) 2025-02-26 13:32:03 +01:00
Steffen SchmitzandGitHub 0b7dd5a1a7 feat: process environment in ingestion pipeline (#5671)
* feat: process environment in ingestion pipeline

* chore: remove unnecessary status

* chore: typing

* chore: more typing

* chore: score env prop

* chore: adjust apis to pass tests

* chore: adjust ingestion service test

* chore: restrict allowed prefix
2025-02-26 12:18:38 +00:00
Steffen SchmitzandGitHub b5cad2693f fix: exclude clickhouse default date from session duration calc (#5726) 2025-02-26 09:51:52 +00:00
Max Deichmann 58d85d66b3 chore: release v3.30.0 2025-02-26 10:30:59 +01:00
Max DeichmannandGitHub 55a5a896f1 chore: track distribution of ingested event body size (#5723)
push
2025-02-26 09:05:08 +00:00
Steffen SchmitzandGitHub 5437beec48 Revert "fix: return nulls for left-joins for correct session durations (#5697)" (#5717)
This reverts commit ce967bac12.
2025-02-25 18:14:26 +01:00
Steffen SchmitzandGitHub 418e33a333 fix: only delete media items with no remaining references (#5713)
* fix: only delete media items with no remaining references

* chore: expand filter condition
2025-02-25 15:18:02 +00:00
Hassieb PakzadandGitHub e9bbd4488f feat(playground+evals): add Google AI Studio support (#5711) 2025-02-25 14:48:22 +01:00
Steffen SchmitzandGitHub 668ca24d5e fix: support gzip encoding on otel endpoint (#5710)
* fix: support gzip encoding on otel endpoint

* chore: check for includes
2025-02-25 10:41:43 +00:00
Hassieb PakzadandGitHub 7c746c7cfc feat(models): add claude-3-7-sonnet-20250219 (#5709) 2025-02-25 10:07:41 +00:00
Hassieb Pakzad de0eb76153 chore: release v3.29.1 2025-02-24 20:53:46 +01:00
Hassieb PakzadandGitHub 85088574b2 fix(public-api): list observations with name filter (#5700) 2025-02-24 20:52:58 +01:00
Steffen SchmitzandGitHub ce967bac12 fix: return nulls for left-joins for correct session durations (#5697)
* fix: return nulls for left-joins for correct session durations

* chore: comment updates

* chore: adjust trace_count
2025-02-24 15:57:47 +00:00
Steffen SchmitzandGitHub b8d0a4bbe8 chore: remove prisma references to traces, observations, scores (#5672)
* chore: remove prisma references to traces, observations, scores

* chore: separate enum and type name

* chore: type overwrites

* chore: chore

* chore: drop views

* chore: reduce diff

* chore: diff

* chore: types gen

* chore: lint

* chore: adjust tests

* chore: remove unused objects in test utils

* chore: adjust tests

* chore: score type parsing

* chore: fix otelmapping

* chore: make observation level internal
2025-02-24 15:43:25 +00:00
marliessophieandGitHub 1e7aeb8fc5 chore(llmCompletion): support system/developer messages for o3 (#5661) 2025-02-24 09:25:33 +00:00
marliessophieandGitHub 9d5cc39cd4 style(ui): improvements to prompt details screen (#5678)
* fix: header

* fix: search always show current version

* chore: adjust code block

* chore: re-add copy in code editor if no title

* chore: do not match current version always

* adjust margins

* chore: remove breadcrumb

* fix: typo

* fix: final margin

* style: eval template

* small design nits
2025-02-22 09:16:44 +00:00
Marc KlingenandGitHub 5c5dfcbac6 chore(ui): improve consistency of prompt and eval template forms (#5679)
* chore(ui): consistency in prompt form labels

* eval prompt form
2025-02-21 18:10:12 +00:00
marliessophieandGitHub 628d180051 feat(ui): new prompt detail screen; adjust design of IO and message view components (#5537)
* style(ui): adjust design of IO and message view components

* chore: adjust final style and fix imports

* style

* nit: show "null" if no output in dataset compare view

* style: add header and tertiary colors

* chore: add Header levels

* chore: add Timeline component

* feat: make JSONView optionally scrollable

* chore: refactor TagManager styling

* style: button variants for header

* chore: ensure chatml message overflow behaves within chat message

* feat: new prompt screen outline

* chore: refactor `extractVariables` to return count

* chore: ensure robust mobile behaviour

* style: small component adjustments

* chore: refactor style of `Comment` button

* feat: add search functionality to prompt nodes

* style: restyle tags

* revert: variable extraction count

* style: prompts page overflow

* chore: control label behaviour

* labels: show buttons on hover

* fix: formatting of titles

* eslint

* push

* push

* fix: overflows

* push

* diabled view
2025-02-21 13:46:59 +00:00
Max Deichmann 57708f0219 chore: release v3.29.0 2025-02-21 13:42:45 +01:00
Max DeichmannandGitHub 5d30aefcb3 fix: do not allow to deactivate eval configs which are only on existing objects (#5675)
* fix: do not allow to deactivate eval configs which rano on historic traces only

* fix: do not allow to deactivate eval configs which rano on historic traces only

* fix: do not allow to deactivate eval configs which rano on historic traces only

* fix: do not allow to deactivate eval configs which rano on historic traces only

* fix: do not allow to deactivate eval configs which rano on historic traces only
2025-02-21 12:40:18 +00:00
Max DeichmannandGitHub dd14e11535 perf: reduce eval create queue concurrency (#5665) 2025-02-20 19:24:38 +00:00
Max DeichmannandGitHub 5f87980013 fix: do not return all executions for eval config list trpc (#5663)
* fix: do not return all executions for eval config list trpc

* fix: do not return all executions for eval config list trpc

* fix: do not return all executions for eval config list trpc

* fix: do not return all executions for eval config list trpc
2025-02-20 17:56:38 +00:00
Max DeichmannandGitHub 2edc2487ae perf: limit concurrency for eval job creations (#5660)
* push

* push
2025-02-20 16:05:41 +00:00
Steffen SchmitzandGitHub 18508187f6 chore: add environment column to trace_sessions (#5656)
* chore: add environment column to trace_sessions

* chore: update test
2025-02-20 15:06:39 +00:00
Max DeichmannGitHubellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
eb5ee957e6 feat: apply evals on history of datasets + traces (#5488)
* push

* push

* push

* push

* fix

* fix

* add tests

* add tests

* add tests

* add tests

* add tests

* add tests

* fic

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* push

* push

* Update web/src/ee/features/evals/components/evaluator-form.tsx

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* push

* fix: fix dompurify

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-02-20 13:59:14 +00:00
Steffen SchmitzandGitHub 582f150b72 chore: add environment columns in clickhouse (#5649)
* chore: add environment columns in clickhouse

* chore: replace logger.warning with logger.warn
2025-02-20 13:05:35 +00:00
Steffen SchmitzandGitHub 7ce4e24cab chore: replace logger.warning with logger.warn (#5650) 2025-02-20 12:12:31 +00:00
marliessophieandGitHub b2d3080790 chore(prompts): support developer message (#5639)
* chore(prompts): support developer message

* chore: fix eslint config

* push
2025-02-20 09:40:32 +00:00
steffen911 3590608d30 chore: increase event log deletion timeout to 5min 2025-02-19 15:56:11 +01:00
Steffen SchmitzandGitHub 78e6c882ba chore: reduce attempts and increase delay on trace delete errors (#5637) 2025-02-19 14:24:28 +00:00
Steffen SchmitzandGitHub d5ea44b33c perf: only validate response schema on dev servers (#5630) 2025-02-19 13:11:42 +00:00
Steffen SchmitzandGitHub 8b601eed20 fix: re-enable trace deletion and remove trace S3 files (#5594)
* fix: re-enable trace deletion and remove trace S3 files

* chore: lint

* chore: lint

* chore: increase query timeout

* chore: reduce query result to only return event log

* chore: keep condition pattern

* chore: revert entitlement changes

* chore: move permission check

* chore: explain semi join and shift entitlement comment
2025-02-19 12:45:37 +00:00
Steffen SchmitzandGitHub caebc92051 chore: bump static page generation timeout (#5621) 2025-02-18 14:44:33 +00:00
Steffen SchmitzandGitHub e41f1c5e5b build: notify slack on failed deployment (#5617) 2025-02-18 14:40:29 +01:00
Steffen SchmitzandGitHub c419fd69fa chore: remove superfluous container_name overwrites (#5615) 2025-02-18 13:03:45 +00:00
Steffen SchmitzandGitHub 485dc94444 feat: map langfuse data model to further otel attributes (#5614)
* feat: map langfuse data model to further otel attributes

* chore: expand tests
2025-02-18 11:29:17 +00:00
Steffen SchmitzandGitHub c233c7fa8d feat: parse mlflow attributes on otel spans (#5609)
* feat: parse mlflow attributes on otel spans

* chore: correct failing test
2025-02-18 09:07:12 +00:00
marliessophieandGitHub 0263b40987 perf(sessions): display only first 10 users; show rest in paginated popover (#5599)
* perf(sessions): show only up to 10 initial users and the rest in popover

* nit: fix header on traces page
2025-02-17 19:01:01 +00:00
Marc KlingenandGitHub 44621d0d1f chore: remove the unnecessary file (#5598)
chore: remove unnecessary file
2025-02-17 18:34:01 +00:00
marliessophieandGitHub 5bd891ddc7 style(traces): style traces delete button in more menu in table (#5596) 2025-02-17 14:59:12 +00:00
marliessophieandGitHub 8e02f86fa7 chore(page-header): add new tabsComponent (#5593)
* chore(page-header): add tabs component

* style: replace tabs component on datasets pages

* style: replace tabs component on evals pages

* nit: fix dataset item page formatting

* style: replace tabs component on prompts pages

* chore: adjust col behaviour on header

* eslint
2025-02-17 13:34:57 +00:00
Max DeichmannandGitHub d7582c66ee security: upgrade jsonpath plus (#5591) 2025-02-17 12:41:12 +00:00
marliessophieandGitHub 75bef08f0b style(page-header): optimise for mobile view (#5590)
* chore: fix demo banner

* fix(page-header): info icon relative positioning

* fix: header title wrap behaviour

* fix(page-header): have info icon follow title

* style: move dataset level buttons to more menu

* push

* push
2025-02-17 12:36:33 +00:00
Max Deichmann 5f7b50cea8 chore: release v3.28.3 2025-02-17 09:56:07 +01:00
Steffen SchmitzandGitHub a9757d144f chore: fallback to LLM http status for /api/chatCompletion (#5554) 2025-02-17 08:42:49 +00:00
Steffen SchmitzandGitHub 12fb26388c fix: accept further content type formats on otel endpoint (#5584) 2025-02-17 08:41:14 +00:00
Max Deichmann 95d9aef705 chore: release v3.28.2 2025-02-15 13:29:06 +01:00
Max Deichmann 15fea028c2 chore: release v3.28.1 2025-02-15 01:04:28 +01:00
Max DeichmannandGitHub ac6e94ee3c security: upgrade release-it (#5558)
* security: upgrade release-it

* security: upgrade release-it

* security: upgrade release-it
2025-02-14 18:59:53 +00:00
Max DeichmannandGitHub e7332bd1e5 security: upgrade dompurify (#5557) 2025-02-14 19:39:59 +01:00
508 changed files with 33895 additions and 11095 deletions
+6
View File
@@ -0,0 +1,6 @@
---
description: How to manage authorization and RBAC within full-stack langfuse features
globs: web/
alwaysApply: true
---
See [README.md](mdc:web/src/features/rbac/README.md) for details
+7
View File
@@ -0,0 +1,7 @@
---
description: How to enforce entitlements across full-stack features
globs: web/**
alwaysApply: true
---
Check [README.md](mdc:web/src/features/entitlements/README.md) to learn more about entitlements
+33
View File
@@ -0,0 +1,33 @@
---
description: How we build new full-stack features including frontend APIs, file structure, and components
globs: web/**
alwaysApply: true
---
## Framework
- Next.js
- Pages router
## File Structure
- We generally put all code related to a net-new feature into a folder within web/src/features.
- Checkout other features to learn about the common structure.
## API for frontend features
- We use TRPC.io to power full-stack features of the Langfuse Frontend
- Entry point for all trpc routes: [root.ts](mdc:web/src/server/api/root.ts)
- Authentication, see [authorization-and-rbac.mdc](mdc:.cursor/rules/authorization-and-rbac.mdc)
- Entitlements, see [entitlements.mdc](mdc:.cursor/rules/entitlements.mdc)
## Components
- We use Shadcn/ui
- Components in `@/src/components/ui`
- If a component is not installed yet, ask the user to install it for you
- When creating new custom components that generalize, we add them with a generalizable naming to `@/src/components`
## Styling
- We use Tailwind CSS
- We use a standard color palette which automatically handles light/dark mode, see [globals.css](mdc:web/src/styles/globals.css)
+16
View File
@@ -0,0 +1,16 @@
---
description: How to create new public api routes for Langfuse
globs:
alwaysApply: false
---
# Implementation
- All public api routes are in /web/src/pages/api/public
- New api routes should follow these guidelines:
- Use [withMiddlewares.ts](mdc:web/src/features/public-api/server/withMiddlewares.ts) as a wrapper
- Define types of api request and response in /web/src/features/public-api/types, use strict() for all zod objects
- Add end-to-end test, similar to [datasets-api.servertest.ts](mdc:web/src/__tests__/async/datasets-api.servertest.ts)
- Add fern configuration in /fern, learn more about structure here: https://buildwithfern.com/learn/api-definition/fern/overview
- Prompt user to regenerate the OpenAPI spec via the fern CLI
- Pagination starts at 1, query typing defined in publicApiPaginationZod, return meta in paginationMetaResponseZod
+40 -27
View File
@@ -84,6 +84,7 @@ OTEL_SERVICE_NAME="langfuse"
# AUTH_GITLAB_ISSUER=
# AUTH_GITLAB_CLIENT_AUTH_METHOD=
# AUTH_GITLAB_CHECKS=
# AUTH_GITLAB_URL=
# AUTH_AZURE_AD_CLIENT_ID=
# AUTH_AZURE_AD_CLIENT_SECRET=
# AUTH_AZURE_AD_TENANT_ID=
@@ -114,6 +115,11 @@ OTEL_SERVICE_NAME="langfuse"
# AUTH_KEYCLOAK_ALLOW_ACCOUNT_LINKING=false
# AUTH_KEYCLOAK_CLIENT_AUTH_METHOD=
# AUTH_KEYCLOAK_CHECKS=
# AUTH_WORKOS_CLIENT_ID=
# AUTH_WORKOS_CLIENT_SECRET=
# AUTH_WORKOS_ALLOW_ACCOUNT_LINKING=false
# AUTH_WORKOS_ORGANIZATION_ID=
# AUTH_WORKOS_CONNECTION_ID=
# AUTH_CUSTOM_CLIENT_ID=
# AUTH_CUSTOM_CLIENT_SECRET=
# AUTH_CUSTOM_ISSUER=
@@ -166,7 +172,32 @@ OTEL_SERVICE_NAME="langfuse"
# LANGFUSE_INIT_USER_NAME=User Name
# LANGFUSE_INIT_USER_PASSWORD=password
# Redis configuration
# REDIS_HOST=
# REDIS_PORT=
# REDIS_AUTH=
# REDIS_CONNECTION_STRING=
# REDIS_ENABLE_AUTO_PIPELINING=
# Cache configuration
# LANGFUSE_CACHE_API_KEY_ENABLED=
# LANGFUSE_CACHE_API_KEY_TTL_SECONDS=
# LANGFUSE_CACHE_PROMPT_ENABLED=
# LANGFUSE_CACHE_PROMPT_TTL_SECONDS=
# Clickhouse configuration
# CLICKHOUSE_URL=
# CLICKHOUSE_CLUSTER_NAME=default
# CLICKHOUSE_DB=default
# CLICKHOUSE_USER=
# CLICKHOUSE_PASSWORD=
# CLICKHOUSE_CLUSTER_ENABLED=true
# Ingestion configuration
# LANGFUSE_INGESTION_QUEUE_DELAY_MS=
# LANGFUSE_INGESTION_CLICKHOUSE_WRITE_BATCH_SIZE=
# LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=
# LANGFUSE_INGESTION_CLICKHOUSE_MAX_ATTEMPTS=
### START Enterprise Edition Configuration
@@ -235,19 +266,8 @@ OTEL_SERVICE_NAME="langfuse"
# Admin API
# ADMIN_API_KEY=
# Redis
# REDIS_HOST=
# REDIS_PORT=
# REDIS_AUTH=
# REDIS_CONNECTION_STRING=
# REDIS_ENABLE_AUTO_PIPELINING=
# Cache configuration
# LANGFUSE_CACHE_API_KEY_ENABLED=
# LANGFUSE_CACHE_API_KEY_TTL_SECONDS=
# LANGFUSE_CACHE_PROMPT_ENABLED=
# LANGFUSE_CACHE_PROMPT_TTL_SECONDS=
# LANGFUSE_CACHE_MODEL_MATCH_ENABLED=
# LANGFUSE_CACHE_MODEL_MATCH_TTL_SECONDS=
# Rate limiting
# LANGFUSE_RATE_LIMITS_ENABLED=
@@ -256,22 +276,15 @@ OTEL_SERVICE_NAME="langfuse"
# STRIPE_SECRET_KEY=
# STRIPE_WEBHOOK_SIGNING_SECRET=
# Betterstack Status Page
# BETTERSTACK_UPTIME_API_KEY=
# BETTERSTACK_UPTIME_STATUS_PAGE_ID=
### END Langfuse Cloud Config
## START Langfuse V3 Ingestion
### START Langfuse CI Config
# Clickhouse
# CLICKHOUSE_MIGRATION_URL=
# CLICKHOUSE_URL=
# CLICKHOUSE_USER=
# CLICKHOUSE_PASSWORD=
# CLICKHOUSE_DB=
# LANGFUSE_INIT_ORG_CLOUD_PLAN=
# Ingestion
# LANGFUSE_INGESTION_QUEUE_DELAY_MS=
# LANGFUSE_INGESTION_CLICKHOUSE_WRITE_BATCH_SIZE=
# LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=
# LANGFUSE_INGESTION_CLICKHOUSE_MAX_ATTEMPTS=
## END Langfuse V3 Ingestion
### END Langfuse CI Config
@@ -56,6 +56,12 @@ jobs:
--build-arg SENTRY_AUTH_TOKEN=${{ secrets.SENTRY_AUTH_TOKEN }} \
--build-arg SENTRY_ORG=${{ vars.SENTRY_ORG }} \
--build-arg SENTRY_PROJECT=${{ vars.SENTRY_PROJECT }} \
--build-arg NEXT_PUBLIC_DATADOG_APPLICATION_ID=${{ vars.NEXT_PUBLIC_DATADOG_APPLICATION_ID }} \
--build-arg NEXT_PUBLIC_DATADOG_CLIENT_TOKEN=${{ vars.NEXT_PUBLIC_DATADOG_CLIENT_TOKEN }} \
--build-arg NEXT_PUBLIC_DATADOG_SITE=${{ vars.NEXT_PUBLIC_DATADOG_SITE }} \
--build-arg NEXT_PUBLIC_DATADOG_ENVIRONMENT=${{ vars.NEXT_PUBLIC_DATADOG_ENVIRONMENT }} \
--build-arg NEXT_PUBLIC_DATADOG_SESSION_SAMPLE_RATE=${{ vars.NEXT_PUBLIC_DATADOG_SESSION_SAMPLE_RATE }} \
--build-arg NEXT_PUBLIC_DATADOG_SESSION_REPLAY_SAMPLE_RATE=${{ vars.NEXT_PUBLIC_DATADOG_SESSION_REPLAY_SAMPLE_RATE }} \
.
docker push $REGISTRY/$REPOSITORY:$IMAGE_TAG
- name: Render AWS ECS Task Definition
+16 -6
View File
@@ -89,7 +89,7 @@ jobs:
timeout 10 bash -c 'until curl -f http://localhost:3000/api/public/health; do sleep 2; done'
tests-web-sync:
timeout-minutes: 20
timeout-minutes: 30
runs-on: ubuntu-latest
needs:
- pre-job
@@ -134,14 +134,14 @@ jobs:
grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev.example > .env
echo "LANGFUSE_INGESTION_QUEUE_DELAY_MS=1" >> .env
echo "LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=1" >> .env
- name: Run + migrate
- name: Run dev containers
run: |
docker compose -f docker-compose.dev.yml up -d
sleep 5 # Wait for PostgreSQL to accept connections
docker compose ps
env:
POSTGRES_VERSION: ${{ matrix.postgres-version }}
- name: Seed DB
- name: Migrate DB
run: |
pnpm run db:migrate
pnpm --filter=shared ch:up
@@ -152,6 +152,7 @@ jobs:
env:
LANGFUSE_INIT_ORG_ID: "seed-org-id"
LANGFUSE_INIT_ORG_NAME: "Seed Org"
LANGFUSE_INIT_ORG_CLOUD_PLAN: "Team"
LANGFUSE_INIT_PROJECT_ID: "7a88fb47-b4e2-43b8-a06c-a5ce950dc53a"
LANGFUSE_INIT_PROJECT_NAME: "Seed Project"
LANGFUSE_INIT_PROJECT_PUBLIC_KEY: "pk-lf-1234567890"
@@ -165,7 +166,7 @@ jobs:
run: pnpm --filter=web run test-client
tests-web-async:
timeout-minutes: 20
timeout-minutes: 30
runs-on: ubuntu-latest
needs:
- pre-job
@@ -211,14 +212,14 @@ jobs:
grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev${{ matrix.blob-provider }}.example > .env
echo "LANGFUSE_INGESTION_QUEUE_DELAY_MS=1" >> .env
echo "LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=1" >> .env
- name: Run + migrate
- name: Run dev containers
run: |
docker compose -f docker-compose.dev${{ matrix.blob-provider }}.yml up -d
sleep 5 # Wait for PostgreSQL to accept connections
docker compose ps
env:
POSTGRES_VERSION: ${{ matrix.postgres-version }}
- name: Seed DB
- name: Migrate DB
run: |
pnpm run db:migrate
pnpm --filter=shared ch:up
@@ -229,6 +230,7 @@ jobs:
env:
LANGFUSE_INIT_ORG_ID: "seed-org-id"
LANGFUSE_INIT_ORG_NAME: "Seed Org"
LANGFUSE_INIT_ORG_CLOUD_PLAN: "Team"
LANGFUSE_INIT_PROJECT_ID: "7a88fb47-b4e2-43b8-a06c-a5ce950dc53a"
LANGFUSE_INIT_PROJECT_NAME: "Seed Project"
LANGFUSE_INIT_PROJECT_PUBLIC_KEY: "pk-lf-1234567890"
@@ -536,3 +538,11 @@ jobs:
platforms: |
linux/amd64
${{ startsWith(github.ref, 'refs/tags/') && 'linux/arm64' || '' }}
- name: Notify Slack
uses: ravsamhq/notify-slack-action@v2
if: always()
with:
status: ${{ job.status }}
notify_when: "failure"
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
+22
View File
@@ -0,0 +1,22 @@
name: Close inactive issues
on:
schedule:
- cron: "30 1 * * *"
jobs:
close-issues:
runs-on: ubuntu-latest
permissions:
issues: write
pull-requests: write
steps:
- uses: actions/stale@v9
with:
days-before-issue-stale: 30
days-before-issue-close: 14
stale-issue-label: "stale"
stale-issue-message: "This issue is stale because it has been open for 30 days with no activity."
close-issue-message: "This issue was closed because it has been inactive for 14 days since being marked as stale. Please reopen if the issue persists."
days-before-pr-stale: -1
days-before-pr-close: -1
repo-token: ${{ secrets.GITHUB_TOKEN }}
+14 -137
View File
@@ -53,117 +53,10 @@ A good first step is to search for open [issues](https://github.com/langfuse/lan
### Architecture Overview
**Langfuse v2**
```mermaid
flowchart TB
subgraph s4["Clients"]
subgraph s2["langfuse/langfuse-python"]
Python["Python low-level SDK"]
Decorator["observe() decorator"] -->|extends| Python
OAI["OpenAI drop-in replacement"] -->|extends| Python
Llamaindex["LlamaIndex Integration"] -->|extends| Python
LCPYTHON["Langchain Python Integration"] -->|extends| Python
Langflow -->|uses| LCPYTHON
LiteLLM -->|uses| Python
end
subgraph s3["langfuse/langfuse-js"]
JS["JS SDK"]
LCJS["Langchain JS Integration"] -->|extends| JS
Flowise -->|uses| LCJS
end
end
DB[Postgres Database]
Redis[Redis]
subgraph s1["Application (langfuse/langfuse/web)"]
API[Public HTTP API]
G[TRPC API]
I[NextAuth]
H[React Frontend]
Prisma[Prisma ORM]
H --> G
H --> I
G --> I
G --- Prisma
API --- Prisma
I --- Prisma
end
Prisma --- DB
JS --- API
Python --- API
```
**Langfuse v3 (work in progress, not released yet)**
> [!NOTE]
> Infrastructure will change in Langfuse version 3.0. More in the [GitHub Discussions](https://github.com/orgs/langfuse/discussions/1902).
> `langfuse/langfuse/worker` is under active development and not recommended for production use in Langfuse 2.x.
```mermaid
flowchart TB
subgraph s4["Clients"]
subgraph s2["langfuse/langfuse-python"]
Python["Python low-level SDK"]
Decorator["observe() decorator"] -->|extends| Python
OAI["OpenAI drop-in replacement"] -->|extends| Python
Llamaindex["LlamaIndex Integration"] -->|extends| Python
LCPYTHON["Langchain Python Integration"] -->|extends| Python
Langflow -->|uses| LCPYTHON
LiteLLM -->|uses| Python
end
subgraph s3["langfuse/langfuse-js"]
JS["JS SDK"]
LCJS["Langchain JS Integration"] -->|extends| JS
Flowise -->|uses| LCJS
end
end
subgraph s9 ["VPC (US and EU separated)"]
DB[Postgres Database]
Redis[Redis Cache/Queue]
Clickhouse[Clickhouse Database]
subgraph s1["Application (langfuse/langfuse/web)"]
API[Public HTTP API]
G[TRPC API]
I[NextAuth]
H[React Frontend]
ORM
H --> G
H --> I
G --> I
G --- ORM
API --- ORM
I --- ORM
end
subgraph s5["Application (langfuse/langfuse/worker)"]
Worker
end
Worker --- DB
Worker --- Redis
Worker --- Clickhouse
ORM --- DB
ORM --- Redis
ORM --- Clickhouse
end
JS --- API
Python --- API
```
See this [diagram](https://langfuse.com/self-hosting#architecture) for an overview of the architecture.
### Network Overview
> [!NOTE]
> This will change in Langfuse version 3.0. More in the [GitHub Discussions](https://github.com/orgs/langfuse/discussions/1902).
```mermaid
flowchart LR
Browser ---|Web UI & TRPC API| App
@@ -209,11 +102,12 @@ Requirements
**Steps**
1. Fork the repository and clone it locally
2. Run the development database
1. Install [golang-migrate](https://github.com/golang-migrate/migrate/tree/master/cmd/migrate#migrate-cli) as CLI
2. Fork the repository and clone it locally
```bash
pnpm run infra:dev:up
git clone https://github.com/langfuse/langfuse.git
cd langfuse
```
3. Create an env file
@@ -222,40 +116,23 @@ Requirements
cp .env.dev.example .env
```
4. Install dependencies
4. Run the entire infrastructure in dev mode
```bash
pnpm install
pnpm run dx
```
5. Run the migrations
You will be asked whether you want to reset Postgres and ClickHouse. Confirm both with 'Y' and press enter.
All database migrations and configs are in the `shared` package.
5. Open the web app in your browser to start using Langfuse:
```bash
pnpm --filter=shared run db:migrate
- [Sign up page, http://localhost:3000](http://localhost:3000)
- [Demo project, http://localhost:3000/project/7a88fb47-b4e2-43b8-a06c-a5ce950dc53a](http://localhost:3000/project/7a88fb47-b4e2-43b8-a06c-a5ce950dc53a)
# Optional: seed the database
# pnpm run db:seed
# pnpm run db:seed:examples
# pnpm --filter=shared run db:seed:load
```
6. Log in as a test user:
6. Start the development server
```bash
pnpm run dev
```
7. Open the web app in the browser:
http://localhost:3000
8. Log in as a test user (after you ran `db:seed` command):
Username: demo@langfuse.com
Password: password
- Username: `demo@langfuse.com`
- Password: `password`
## Monorepo quickstart
View File
-3
View File
@@ -80,8 +80,6 @@ services:
clickhouse:
image: clickhouse/clickhouse-server
user: "101:101"
container_name: clickhouse
hostname: clickhouse
environment:
CLICKHOUSE_DB: default
CLICKHOUSE_USER: clickhouse
@@ -101,7 +99,6 @@ services:
minio:
image: minio/minio
container_name: minio
entrypoint: sh
# create the 'langfuse' bucket before starting the service
command: -c 'mkdir -p /data/langfuse && minio server --address ":9000" --console-address ":9001" /data'
-3
View File
@@ -2,8 +2,6 @@ services:
clickhouse:
image: clickhouse/clickhouse-server
user: "101:101"
container_name: clickhouse
hostname: clickhouse
environment:
CLICKHOUSE_DB: default
CLICKHOUSE_USER: clickhouse
@@ -19,7 +17,6 @@ services:
azurite:
image: mcr.microsoft.com/azure-storage/azurite
container_name: azurite
command: azurite-blob --blobHost 0.0.0.0
ports:
- "10000:10000"
+6 -9
View File
@@ -2,8 +2,6 @@ services:
clickhouse:
image: clickhouse/clickhouse-server
user: "101:101"
container_name: clickhouse
hostname: clickhouse
environment:
CLICKHOUSE_DB: default
CLICKHOUSE_USER: clickhouse
@@ -12,14 +10,13 @@ services:
- langfuse_clickhouse_data:/var/lib/clickhouse
- langfuse_clickhouse_logs:/var/log/clickhouse-server
ports:
- "8123:8123"
- "9000:9000"
- 127.0.0.1:8123:8123
- 127.0.0.1:9000:9000
depends_on:
- postgres
minio:
image: minio/minio
container_name: minio
entrypoint: sh
# create the 'langfuse' bucket before starting the service
command: -c 'mkdir -p /data/langfuse && minio server --address ":9000" --console-address ":9001" /data'
@@ -27,8 +24,8 @@ services:
MINIO_ACCESS_KEY: minio
MINIO_SECRET_KEY: miniosecret
ports:
- "9090:9000"
- "9091:9001"
- 127.0.0.1:9090:9000
- 127.0.0.1:9091:9001
volumes:
- langfuse_minio_data:/data
healthcheck:
@@ -44,7 +41,7 @@ services:
command: >
--requirepass ${REDIS_AUTH:-myredissecret}
ports:
- 6379:6379
- 127.0.0.1:6379:6379
postgres:
image: postgres:${POSTGRES_VERSION:-latest}
@@ -60,7 +57,7 @@ services:
- POSTGRES_PASSWORD=postgres
- POSTGRES_DB=postgres
ports:
- 5432:5432
- 127.0.0.1:5432:5432
volumes:
- langfuse_postgres_data:/var/lib/postgresql/data
+25 -22
View File
@@ -1,3 +1,8 @@
# Make sure to update the credential placeholders with your own secrets.
# We mark them with # CHANGEME in the file below.
# In addition, we recommend to restrict inbound traffic on the host to langfuse-web (port 3000) and minio (port 9090) only.
# All other components are bound to localhost (127.0.0.1) to only accept connections from the local machine.
# External connections from other machines will not be able to reach these services directly.
services:
langfuse-worker:
image: langfuse/langfuse-worker:3
@@ -12,29 +17,29 @@ services:
clickhouse:
condition: service_healthy
ports:
- "3030:3030"
- 127.0.0.1:3030:3030
environment: &langfuse-worker-env
DATABASE_URL: postgresql://postgres:postgres@postgres:5432/postgres
SALT: "mysalt"
ENCRYPTION_KEY: "0000000000000000000000000000000000000000000000000000000000000000" # generate via `openssl rand -hex 32`
DATABASE_URL: postgresql://postgres:postgres@postgres:5432/postgres # CHANGEME
SALT: "mysalt" # CHANGEME
ENCRYPTION_KEY: "0000000000000000000000000000000000000000000000000000000000000000" # CHANGEME: generate via `openssl rand -hex 32`
TELEMETRY_ENABLED: ${TELEMETRY_ENABLED:-true}
LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES: ${LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES:-true}
CLICKHOUSE_MIGRATION_URL: ${CLICKHOUSE_MIGRATION_URL:-clickhouse://clickhouse:9000}
CLICKHOUSE_URL: ${CLICKHOUSE_URL:-http://clickhouse:8123}
CLICKHOUSE_USER: ${CLICKHOUSE_USER:-clickhouse}
CLICKHOUSE_PASSWORD: ${CLICKHOUSE_PASSWORD:-clickhouse}
CLICKHOUSE_PASSWORD: ${CLICKHOUSE_PASSWORD:-clickhouse} # CHANGEME
CLICKHOUSE_CLUSTER_ENABLED: ${CLICKHOUSE_CLUSTER_ENABLED:-false}
LANGFUSE_S3_EVENT_UPLOAD_BUCKET: ${LANGFUSE_S3_EVENT_UPLOAD_BUCKET:-langfuse}
LANGFUSE_S3_EVENT_UPLOAD_REGION: ${LANGFUSE_S3_EVENT_UPLOAD_REGION:-auto}
LANGFUSE_S3_EVENT_UPLOAD_ACCESS_KEY_ID: ${LANGFUSE_S3_EVENT_UPLOAD_ACCESS_KEY_ID:-minio}
LANGFUSE_S3_EVENT_UPLOAD_SECRET_ACCESS_KEY: ${LANGFUSE_S3_EVENT_UPLOAD_SECRET_ACCESS_KEY:-miniosecret}
LANGFUSE_S3_EVENT_UPLOAD_SECRET_ACCESS_KEY: ${LANGFUSE_S3_EVENT_UPLOAD_SECRET_ACCESS_KEY:-miniosecret} # CHANGEME
LANGFUSE_S3_EVENT_UPLOAD_ENDPOINT: ${LANGFUSE_S3_EVENT_UPLOAD_ENDPOINT:-http://minio:9000}
LANGFUSE_S3_EVENT_UPLOAD_FORCE_PATH_STYLE: ${LANGFUSE_S3_EVENT_UPLOAD_FORCE_PATH_STYLE:-true}
LANGFUSE_S3_EVENT_UPLOAD_PREFIX: ${LANGFUSE_S3_EVENT_UPLOAD_PREFIX:-events/}
LANGFUSE_S3_MEDIA_UPLOAD_BUCKET: ${LANGFUSE_S3_MEDIA_UPLOAD_BUCKET:-langfuse}
LANGFUSE_S3_MEDIA_UPLOAD_REGION: ${LANGFUSE_S3_MEDIA_UPLOAD_REGION:-auto}
LANGFUSE_S3_MEDIA_UPLOAD_ACCESS_KEY_ID: ${LANGFUSE_S3_MEDIA_UPLOAD_ACCESS_KEY_ID:-minio}
LANGFUSE_S3_MEDIA_UPLOAD_SECRET_ACCESS_KEY: ${LANGFUSE_S3_MEDIA_UPLOAD_SECRET_ACCESS_KEY:-miniosecret}
LANGFUSE_S3_MEDIA_UPLOAD_SECRET_ACCESS_KEY: ${LANGFUSE_S3_MEDIA_UPLOAD_SECRET_ACCESS_KEY:-miniosecret} # CHANGEME
LANGFUSE_S3_MEDIA_UPLOAD_ENDPOINT: ${LANGFUSE_S3_MEDIA_UPLOAD_ENDPOINT:-http://minio:9000}
LANGFUSE_S3_MEDIA_UPLOAD_FORCE_PATH_STYLE: ${LANGFUSE_S3_MEDIA_UPLOAD_FORCE_PATH_STYLE:-true}
LANGFUSE_S3_MEDIA_UPLOAD_PREFIX: ${LANGFUSE_S3_MEDIA_UPLOAD_PREFIX:-media/}
@@ -42,7 +47,7 @@ services:
LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS: ${LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS:-}
REDIS_HOST: ${REDIS_HOST:-redis}
REDIS_PORT: ${REDIS_PORT:-6379}
REDIS_AUTH: ${REDIS_AUTH:-myredissecret}
REDIS_AUTH: ${REDIS_AUTH:-myredissecret} # CHANGEME
REDIS_TLS_ENABLED: ${REDIS_TLS_ENABLED:-false}
REDIS_TLS_CA: ${REDIS_TLS_CA:-/certs/ca.crt}
REDIS_TLS_CERT: ${REDIS_TLS_CERT:-/certs/redis.crt}
@@ -53,11 +58,11 @@ services:
restart: always
depends_on: *langfuse-depends-on
ports:
- "3000:3000"
- 3000:3000
environment:
<<: *langfuse-worker-env
NEXTAUTH_URL: http://localhost:3000
NEXTAUTH_SECRET: mysecret
NEXTAUTH_SECRET: mysecret # CHANGEME
LANGFUSE_INIT_ORG_ID: ${LANGFUSE_INIT_ORG_ID:-}
LANGFUSE_INIT_ORG_NAME: ${LANGFUSE_INIT_ORG_NAME:-}
LANGFUSE_INIT_PROJECT_ID: ${LANGFUSE_INIT_PROJECT_ID:-}
@@ -72,18 +77,16 @@ services:
image: clickhouse/clickhouse-server
restart: always
user: "101:101"
container_name: clickhouse
hostname: clickhouse
environment:
CLICKHOUSE_DB: default
CLICKHOUSE_USER: clickhouse
CLICKHOUSE_PASSWORD: clickhouse
CLICKHOUSE_PASSWORD: clickhouse # CHANGEME
volumes:
- langfuse_clickhouse_data:/var/lib/clickhouse
- langfuse_clickhouse_logs:/var/log/clickhouse-server
ports:
- "8123:8123"
- "9000:9000"
- 127.0.0.1:8123:8123
- 127.0.0.1:9000:9000
healthcheck:
test: wget --no-verbose --tries=1 --spider http://localhost:8123/ping || exit 1
interval: 5s
@@ -94,16 +97,15 @@ services:
minio:
image: minio/minio
restart: always
container_name: minio
entrypoint: sh
# create the 'langfuse' bucket before starting the service
command: -c 'mkdir -p /data/langfuse && minio server --address ":9000" --console-address ":9001" /data'
environment:
MINIO_ROOT_USER: minio
MINIO_ROOT_PASSWORD: miniosecret
MINIO_ROOT_PASSWORD: miniosecret # CHANGEME
ports:
- "9090:9000"
- "9091:9001"
- 9090:9000
- 127.0.0.1:9091:9001
volumes:
- langfuse_minio_data:/data
healthcheck:
@@ -116,10 +118,11 @@ services:
redis:
image: redis:7
restart: always
# CHANGEME: row below to secure redis password
command: >
--requirepass ${REDIS_AUTH:-myredissecret}
ports:
- 6379:6379
- 127.0.0.1:6379:6379
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 3s
@@ -136,10 +139,10 @@ services:
retries: 10
environment:
POSTGRES_USER: postgres
POSTGRES_PASSWORD: postgres
POSTGRES_PASSWORD: postgres # CHANGEME
POSTGRES_DB: postgres
ports:
- 5432:5432
- 127.0.0.1:5432:5432
volumes:
- langfuse_postgres_data:/var/lib/postgresql/data
View File
+2 -2
View File
@@ -26,9 +26,9 @@
"dependencies": {
"@langfuse/shared": "workspace:*",
"@opentelemetry/api": ">=1.0.0 <1.10.0",
"axios": "^1.7.7",
"axios": "^1.8.2",
"https-proxy-agent": "^7.0.6",
"next": "^14.2.21",
"next": "^14.2.25",
"next-auth": "^4.24.11",
"zod": "^3.23.8"
},
@@ -0,0 +1,165 @@
# yaml-language-server: $schema=https://raw.githubusercontent.com/fern-api/fern/main/fern.schema.json
imports:
commons: ./commons.yml
pagination: ./utils/pagination.yml
service:
auth: true
base-path: /api/public
endpoints:
listQueues:
docs: Get all annotation queues
method: GET
path: /annotation-queues
request:
name: GetAnnotationQueuesRequest
query-parameters:
page:
type: optional<integer>
docs: page number, starts at 1
limit:
type: optional<integer>
docs: limit of items per page
response: PaginatedAnnotationQueues
getQueue:
docs: Get an annotation queue by ID
method: GET
path: /annotation-queues/{queueId}
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
response: AnnotationQueue
listQueueItems:
docs: Get items for a specific annotation queue
method: GET
path: /annotation-queues/{queueId}/items
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
request:
name: GetAnnotationQueueItemsRequest
query-parameters:
status:
type: optional<AnnotationQueueStatus>
docs: Filter by status
page:
type: optional<integer>
docs: page number, starts at 1
limit:
type: optional<integer>
docs: limit of items per page
response: PaginatedAnnotationQueueItems
getQueueItem:
docs: Get a specific item from an annotation queue
method: GET
path: /annotation-queues/{queueId}/items/{itemId}
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
itemId:
type: string
docs: The unique identifier of the annotation queue item
response: AnnotationQueueItem
createQueueItem:
docs: Add an item to an annotation queue
method: POST
path: /annotation-queues/{queueId}/items
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
request: CreateAnnotationQueueItemRequest
response: AnnotationQueueItem
updateQueueItem:
docs: Update an annotation queue item
method: PATCH
path: /annotation-queues/{queueId}/items/{itemId}
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
itemId:
type: string
docs: The unique identifier of the annotation queue item
request: UpdateAnnotationQueueItemRequest
response: AnnotationQueueItem
deleteQueueItem:
docs: Remove an item from an annotation queue
method: DELETE
path: /annotation-queues/{queueId}/items/{itemId}
path-parameters:
queueId:
type: string
docs: The unique identifier of the annotation queue
itemId:
type: string
docs: The unique identifier of the annotation queue item
response: DeleteAnnotationQueueItemResponse
types:
AnnotationQueueStatus:
enum:
- PENDING
- COMPLETED
AnnotationQueueObjectType:
enum:
- TRACE
- OBSERVATION
AnnotationQueue:
properties:
id: string
name: string
description: optional<string>
scoreConfigIds: list<string>
createdAt: datetime
updatedAt: datetime
AnnotationQueueItem:
properties:
id: string
queueId: string
objectId: string
objectType: AnnotationQueueObjectType
status: AnnotationQueueStatus
completedAt: optional<datetime>
createdAt: datetime
updatedAt: datetime
PaginatedAnnotationQueues:
properties:
data: list<AnnotationQueue>
meta: pagination.MetaResponse
PaginatedAnnotationQueueItems:
properties:
data: list<AnnotationQueueItem>
meta: pagination.MetaResponse
CreateAnnotationQueueItemRequest:
properties:
objectId: string
objectType: AnnotationQueueObjectType
status:
type: optional<AnnotationQueueStatus>
docs: Defaults to PENDING for new queue items
UpdateAnnotationQueueItemRequest:
properties:
status:
type: optional<AnnotationQueueStatus>
DeleteAnnotationQueueItemResponse:
properties:
success: boolean
message: string
+12
View File
@@ -38,6 +38,9 @@ types:
public:
type: optional<boolean>
docs: Public traces are accessible via url without login
environment:
type: optional<string>
docs: The environment from which this trace originated. Can be any lowercase alphanumeric string with hyphens and underscores that does not start with 'langfuse'.
TraceWithDetails: # GET /traces
extends: Trace
properties:
@@ -79,6 +82,9 @@ types:
id: string
createdAt: datetime
projectId: string
environment:
type: optional<string>
docs: The environment from which this session originated.
SessionWithTraces:
extends: Session
properties:
@@ -145,6 +151,9 @@ types:
costDetails:
type: optional<map<string, double>>
docs: The cost details of the observation. Key is the name of the cost metric, value is the cost in USD. The total key is the sum of all (non-total) cost metrics or the total value ingested.
environment:
type: optional<string>
docs: The environment from which this observation originated. Can be any lowercase alphanumeric string with hyphens and underscores that does not start with 'langfuse'.
ObservationsView:
extends: Observation
@@ -249,6 +258,9 @@ types:
queueId:
type: optional<string>
docs: Reference an annotation queue on a score. Populated if the score was initially created in an annotation queue.
environment:
type: optional<string>
docs: The environment from which this score originated. Can be any lowercase alphanumeric string with hyphens and underscores that does not start with 'langfuse'.
NumericScore:
extends: BaseScore
properties:
@@ -37,8 +37,21 @@ service:
type: optional<integer>
docs: limit of items per page
response: PaginatedDatasetItems
delete:
docs: Delete a dataset item and all its run items. This action is irreversible.
method: DELETE
path: /dataset-items/{id}
path-parameters:
id:
type: string
response: DeleteDatasetItemResponse
types:
DeleteDatasetItemResponse:
properties:
message:
type: string
docs: Success message after deletion
CreateDatasetItemRequest:
properties:
datasetName: string
+11
View File
@@ -41,6 +41,14 @@ service:
datasetName: string
runName: string
response: commons.DatasetRunWithItems
deleteRun:
method: DELETE
docs: Delete a dataset run and all its run items. This action is irreversible.
path: /datasets/{datasetName}/runs/{runName}
path-parameters:
datasetName: string
runName: string
response: DeleteDatasetRunResponse
getRuns:
method: GET
docs: Get dataset runs
@@ -72,3 +80,6 @@ types:
properties:
data: list<commons.DatasetRun>
meta: pagination.MetaResponse
DeleteDatasetRunResponse:
properties:
message: string
+22 -4
View File
@@ -47,6 +47,7 @@ service:
body:
id: abcdef-1234-5678-90ab
timestamp: "2022-01-01T00:00:00.000Z"
environment: "production"
name: "My Trace"
userId: "1234-5678-90ab-cdef"
input: "My input"
@@ -74,6 +75,7 @@ service:
id: abcdef-1234-5678-90ab
traceId: "1234-5678-90ab-cdef"
startTime: "2022-01-01T00:00:00.000Z"
environment: "test"
response:
body:
successes:
@@ -92,6 +94,7 @@ service:
traceId: "1234-5678-90ab-cdef"
name: "My Score"
value: 0.9
environment: "default"
response:
body:
successes:
@@ -168,6 +171,7 @@ types:
statusMessage: optional<string>
parentObservationId: optional<string>
version: optional<string>
environment: optional<string>
CreateEventBody:
extends: OptionalObservationBody
@@ -232,6 +236,7 @@ types:
level: optional<commons.ObservationLevel>
statusMessage: optional<string>
parentObservationId: optional<string>
environment: optional<string>
TraceBody:
properties:
@@ -246,6 +251,7 @@ types:
version: optional<string>
metadata: optional<unknown>
tags: optional<list<string>>
environment: optional<string>
public:
type: optional<boolean>
docs: Make trace publicly accessible via url
@@ -259,6 +265,7 @@ types:
id: optional<string>
traceId: string
name: string
environment: optional<string>
value:
type: commons.CreateScoreValue
docs: The value of the score. Must be passed as string for categorical scores, and numeric for boolean and numeric scores. Boolean score values must equal either 1 or 0 (true or false)
@@ -387,16 +394,27 @@ types:
successes: list<IngestionSuccess>
errors: list<IngestionError>
OpenAIUsageSchema:
OpenAICompletionUsageSchema:
docs: OpenAI Usage schema from (Chat-)Completion APIs
properties:
prompt_tokens: integer
completion_tokens: integer
total_tokens: integer
prompt_tokens_details: optional<map<string, integer>>
completion_tokens_details: optional<map<string, integer>>
prompt_tokens_details: optional<map<string, optional<integer>>>
completion_tokens_details: optional<map<string, optional<integer>>>
OpenAIResponseUsageSchema:
docs: OpenAI Usage schema from Response API
properties:
input_tokens: integer
output_tokens: integer
total_tokens: integer
input_tokens_details: optional<map<string, optional<integer>>>
output_tokens_details: optional<map<string, optional<integer>>>
UsageDetails:
discriminated: false
union:
- map<string, integer>
- OpenAIUsageSchema
- OpenAICompletionUsageSchema
- OpenAIResponseUsageSchema
+4
View File
@@ -29,6 +29,10 @@ service:
type: optional<string>
allow-multiple: true
docs: Optional filter for metrics where traces include all of these tags
environment:
type: optional<string>
allow-multiple: true
docs: Optional filter for metrics where events include any of these environments
fromTimestamp:
type: optional<datetime>
docs: Optional filter to only include traces and observations on or after a certain datetime (ISO 8601)
@@ -33,6 +33,10 @@ service:
type: optional<string>
traceId: optional<string>
parentObservationId: optional<string>
environment:
type: optional<string>
allow-multiple: true
docs: Optional filter for observations where the environment is one of the provided values.
fromStartTime:
type: optional<datetime>
docs: Retrieve only observations with a start_time or or after this datetime (ISO 8601).
+3
View File
@@ -128,6 +128,9 @@ types:
commitMessage:
type: optional<string>
docs: Commit message for this prompt version.
resolutionGraph:
type: optional<map<string, unknown>>
docs: The dependency resolution graph for the current prompt. Null if prompt has no dependencies.
ChatMessage:
properties:
+12
View File
@@ -37,6 +37,10 @@ service:
toTimestamp:
type: optional<datetime>
docs: Optional filter to only include scores created before a certain datetime (ISO 8601)
environment:
type: optional<string>
allow-multiple: true
docs: Optional filter for scores where the environment is one of the provided values.
source:
type: optional<commons.ScoreSource>
docs: Retrieve only scores from a specific source.
@@ -91,6 +95,9 @@ types:
docs: The value of the score. Must be passed as string for categorical scores, and numeric for boolean and numeric scores. Boolean score values must equal either 1 or 0 (true or false)
observationId: optional<string>
comment: optional<string>
environment:
type: optional<string>
docs: The environment of the score. Can be any lowercase alphanumeric string with hyphens and underscores that does not start with 'langfuse'.
dataType:
type: optional<commons.ScoreDataType>
docs: The data type of the score. When passing a configId this field is inferred. Otherwise, this field must be passed or will default to numeric.
@@ -113,10 +120,12 @@ types:
dataType: "NUMERIC"
configId: "9203-4567-89ab-cdef"
traceId: "cdef-1234-5678-90ab"
environment: "test"
- value:
name: "toxicity"
value: "not toxic"
traceId: "cdef-1234-5678-90ab"
environment: "production"
- value:
name: "correctness"
value: "partially correct"
@@ -147,6 +156,9 @@ types:
tags:
type: optional<list<string>>
docs: A list of tags associated with the trace referenced by score
environment:
type: optional<string>
docs: The environment of the trace referenced by score
GetScoresResponseDataNumeric:
extends: commons.NumericScore
+4
View File
@@ -25,6 +25,10 @@ service:
toTimestamp:
type: optional<datetime>
docs: Optional filter to only include sessions created before a certain datetime (ISO 8601)
environment:
type: optional<string>
allow-multiple: true
docs: Optional filter for sessions where the environment is one of the provided values.
response: PaginatedSessions
get:
docs: Get a session. Please note that `traces` on this endpoint are not paginated, if you plan to fetch large sessions, consider `GET /api/public/traces?sessionId=<sessionId>`
+28
View File
@@ -15,6 +15,15 @@ service:
type: string
docs: The unique langfuse identifier of a trace
response: commons.TraceWithFullDetails
delete:
docs: Delete a specific trace
method: DELETE
path: /traces/{traceId}
path-parameters:
traceId:
type: string
docs: The unique langfuse identifier of the trace to delete
response: DeleteTraceResponse
list:
docs: Get list of traces
method: GET
@@ -50,13 +59,32 @@ service:
release:
type: optional<string>
docs: Optional filter to only include traces with a certain release.
environment:
type: optional<string>
allow-multiple: true
docs: Optional filter for traces where the environment is one of the provided values.
response: Traces
deleteMultiple:
docs: Delete multiple traces
method: DELETE
path: /traces
request:
name: DeleteTracesRequest
body:
properties:
traceIds:
type: list<string>
docs: List of trace IDs to delete
response: DeleteTraceResponse
types:
Traces:
properties:
data: list<commons.TraceWithDetails>
meta: pagination.MetaResponse
DeleteTraceResponse:
properties:
message: string
Sort:
properties:
id: string
+7
View File
@@ -19,6 +19,13 @@ groups:
pydantic_config:
require_optional_fields: false
use_str_enums: false
# - name: fernapi/fern-java-sdk
# version: 2.20.1
# output:
# location: local-file-system
# path: ../../../../langfuse-java/src/main/java/com/langfuse/client/
# config:
# client-class-name: LangfuseClient
- name: fernapi/fern-postman
version: 0.0.45
output:
+1 -2
View File
@@ -1,5 +1,4 @@
{
"organization": "langfuse",
"version": "0.43.7"
"version": "0.56.0"
}
+3 -3
View File
@@ -1,6 +1,6 @@
{
"name": "langfuse",
"version": "3.28.0",
"version": "3.45.2",
"author": "engineering@langfuse.com",
"license": "MIT",
"private": true,
@@ -29,12 +29,12 @@
"prepare": "husky"
},
"devDependencies": {
"@release-it/bumper": "^6.0.1",
"@release-it/bumper": "^7.0.1",
"braces": "3.0.3",
"dotenv-cli": "^7.4.2",
"husky": "^9.0.11",
"prettier": "^3.3.3",
"release-it": "^17.3.0",
"release-it": "^18.1.2",
"turbo": "^1.13.4"
},
"release-it": {
@@ -0,0 +1,3 @@
ALTER TABLE traces ON CLUSTER default DROP COLUMN IF EXISTS environment;
ALTER TABLE observations ON CLUSTER default DROP COLUMN IF EXISTS environment;
ALTER TABLE scores ON CLUSTER default DROP COLUMN IF EXISTS environment;
@@ -0,0 +1,3 @@
ALTER TABLE traces ON CLUSTER default ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
ALTER TABLE observations ON CLUSTER default ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
ALTER TABLE scores ON CLUSTER default ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
@@ -0,0 +1,7 @@
-- Drop materialized views
DROP VIEW IF EXISTS project_environments_traces_mv ON CLUSTER default;
DROP VIEW IF EXISTS project_environments_observations_mv ON CLUSTER default;
DROP VIEW IF EXISTS project_environments_scores_mv ON CLUSTER default;
-- Drop the project_environments table
DROP TABLE IF EXISTS project_environments ON CLUSTER default;
@@ -0,0 +1,30 @@
-- Create the project_environments table with AggregatingMergeTree engine
CREATE TABLE project_environments ON CLUSTER default (
`project_id` String,
`environments` SimpleAggregateFunction(groupUniqArrayArray, Array(String))
) ENGINE = ReplicatedAggregatingMergeTree
ORDER BY (project_id);
-- Create materialized view for traces
CREATE MATERIALIZED VIEW project_environments_traces_mv ON CLUSTER default TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM traces
GROUP BY project_id;
-- Create materialized view for observations
CREATE MATERIALIZED VIEW project_environments_observations_mv ON CLUSTER default TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM observations
GROUP BY project_id;
-- Create materialized view for scores
CREATE MATERIALIZED VIEW project_environments_scores_mv ON CLUSTER default TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM scores
GROUP BY project_id;
@@ -0,0 +1,3 @@
ALTER TABLE traces DROP COLUMN IF EXISTS environment;
ALTER TABLE observations DROP COLUMN IF EXISTS environment;
ALTER TABLE scores DROP COLUMN IF EXISTS environment;
@@ -0,0 +1,3 @@
ALTER TABLE traces ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
ALTER TABLE observations ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
ALTER TABLE scores ADD COLUMN environment LowCardinality(String) DEFAULT 'default' AFTER project_id;
@@ -0,0 +1,7 @@
-- Drop materialized views
DROP VIEW IF EXISTS project_environments_traces_mv;
DROP VIEW IF EXISTS project_environments_observations_mv;
DROP VIEW IF EXISTS project_environments_scores_mv;
-- Drop the project_environments table
DROP TABLE IF EXISTS project_environments;
@@ -0,0 +1,30 @@
-- Create the project_environments table with AggregatingMergeTree engine
CREATE TABLE project_environments (
`project_id` String,
`environments` SimpleAggregateFunction(groupUniqArrayArray, Array(String))
) ENGINE = AggregatingMergeTree
ORDER BY (project_id);
-- Create materialized view for traces
CREATE MATERIALIZED VIEW project_environments_traces_mv TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM traces
GROUP BY project_id;
-- Create materialized view for observations
CREATE MATERIALIZED VIEW project_environments_observations_mv TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM observations
GROUP BY project_id;
-- Create materialized view for scores
CREATE MATERIALIZED VIEW project_environments_scores_mv TO project_environments AS
SELECT
project_id,
groupUniqArray(environment) AS environments
FROM scores
GROUP BY project_id;
+1 -2
View File
@@ -1,6 +1,5 @@
import { ObservationRecordReadType } from "@langfuse/shared/src/server";
import { prisma } from "../../src/db";
import { redis } from "@langfuse/shared/src/server";
import { ObservationRecordReadType, redis } from "../../src/server";
import { prepareClickhouse } from "../../scripts/prepareClickhouse";
import { createDatasets } from "../../prisma/seed";
import { queryClickhouse } from "../../src/server/repositories/clickhouse";
+5 -3
View File
@@ -33,7 +33,7 @@
"scripts": {
"build": "tsc",
"dev": "tsc --watch",
"lint": "eslint . --ext .js,.jsx,.ts,.tsx --max-warnings 106",
"lint": "eslint . --ext .js,.jsx,.ts,.tsx --max-warnings 64",
"lint:fix": "eslint . --ext .js,.jsx,.ts,.tsx --fix",
"db:migrate": "DISABLE_ERD=false dotenv -e ../../.env -- npx prisma migrate dev",
"db:push": "DISABLE_ERD=false dotenv -e ../../.env -- npx prisma db push",
@@ -47,7 +47,7 @@
"ch:down": "bash clickhouse/scripts/down.sh",
"ch:drop": "bash clickhouse/scripts/drop.sh",
"ch:reset": "pnpm run ch:down && pnpm run ch:up && pnpm run ch:seed",
"ch:seed": "dotenv -e ../../.env -- tsx clickhouse/scripts/seed.ts",
"ch:seed": "dotenv -e ../../.env -- ts-node -r tsconfig-paths/register -r dotenv/config --compiler-options '{\"module\":\"CommonJS\"}' clickhouse/scripts/seed.ts",
"load:setup": "dotenv -e ../../.env -- tsx scripts/load-seed.ts"
},
"prisma": {
@@ -64,6 +64,7 @@
"@langchain/anthropic": "^0.3.12",
"@langchain/aws": "^0.1.3",
"@langchain/core": "^0.3.37",
"@langchain/google-genai": "^0.1.9",
"@langchain/google-vertexai": "^0.1.8",
"@langchain/openai": "^0.3.17",
"@opentelemetry/api": ">=1.0.0 <1.10.0",
@@ -71,7 +72,7 @@
"@react-email/components": "^0.0.19",
"@react-email/render": "^0.0.15",
"@types/bcryptjs": "^2.4.6",
"axios": "^1.7.7",
"axios": "^1.8.2",
"bcryptjs": "^2.4.3",
"bullmq": "^5.34.10",
"dd-trace": "^5.36.0",
@@ -83,6 +84,7 @@
"langchain": "^0.3.15",
"langfuse-langchain": "3.30.3",
"lodash": "^4.17.21",
"lossless-json": "^4.0.2",
"next-auth": "^4.24.11",
"nodemailer": "^6.9.15",
"prisma-extension-kysely": "^2.1.0",
+122 -145
View File
@@ -12,25 +12,25 @@ export const Role = {
NONE: "NONE"
} as const;
export type Role = (typeof Role)[keyof typeof Role];
export const ObservationType = {
export const LegacyPrismaObservationType = {
SPAN: "SPAN",
EVENT: "EVENT",
GENERATION: "GENERATION"
} as const;
export type ObservationType = (typeof ObservationType)[keyof typeof ObservationType];
export const ObservationLevel = {
export type LegacyPrismaObservationType = (typeof LegacyPrismaObservationType)[keyof typeof LegacyPrismaObservationType];
export const LegacyPrismaObservationLevel = {
DEBUG: "DEBUG",
DEFAULT: "DEFAULT",
WARNING: "WARNING",
ERROR: "ERROR"
} as const;
export type ObservationLevel = (typeof ObservationLevel)[keyof typeof ObservationLevel];
export const ScoreSource = {
export type LegacyPrismaObservationLevel = (typeof LegacyPrismaObservationLevel)[keyof typeof LegacyPrismaObservationLevel];
export const LegacyPrismaScoreSource = {
ANNOTATION: "ANNOTATION",
API: "API",
EVAL: "EVAL"
} as const;
export type ScoreSource = (typeof ScoreSource)[keyof typeof ScoreSource];
export type LegacyPrismaScoreSource = (typeof LegacyPrismaScoreSource)[keyof typeof LegacyPrismaScoreSource];
export const ScoreDataType = {
CATEGORICAL: "CATEGORICAL",
NUMERIC: "NUMERIC",
@@ -59,6 +59,11 @@ export const CommentObjectType = {
PROMPT: "PROMPT"
} as const;
export type CommentObjectType = (typeof CommentObjectType)[keyof typeof CommentObjectType];
export const AuditLogRecordType = {
USER: "USER",
API_KEY: "API_KEY"
} as const;
export type AuditLogRecordType = (typeof AuditLogRecordType)[keyof typeof AuditLogRecordType];
export const JobType = {
EVAL: "EVAL"
} as const;
@@ -75,6 +80,12 @@ export const JobExecutionStatus = {
CANCELLED: "CANCELLED"
} as const;
export type JobExecutionStatus = (typeof JobExecutionStatus)[keyof typeof JobExecutionStatus];
export const BlobStorageIntegrationType = {
S3: "S3",
S3_COMPATIBLE: "S3_COMPATIBLE",
AZURE_BLOB_STORAGE: "AZURE_BLOB_STORAGE"
} as const;
export type BlobStorageIntegrationType = (typeof BlobStorageIntegrationType)[keyof typeof BlobStorageIntegrationType];
export type Account = {
id: string;
user_id: string;
@@ -132,9 +143,11 @@ export type AuditLog = {
id: string;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
user_id: string;
type: Generated<AuditLogRecordType>;
api_key_id: string | null;
user_id: string | null;
org_id: string;
user_org_role: string;
user_org_role: string | null;
project_id: string | null;
user_project_role: string | null;
resource_type: string;
@@ -181,6 +194,23 @@ export type BillingMeterBackup = {
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
};
export type BlobStorageIntegration = {
project_id: string;
type: BlobStorageIntegrationType;
bucket_name: string;
prefix: string;
access_key_id: string;
secret_access_key: string;
region: string;
endpoint: string | null;
force_path_style: boolean;
next_sync_at: Timestamp | null;
last_sync_at: Timestamp | null;
enabled: boolean;
export_frequency: string;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
};
export type Comment = {
id: string;
project_id: string;
@@ -267,6 +297,7 @@ export type JobConfiguration = {
variable_mapping: unknown;
sampling: string;
delay: number;
time_scope: Generated<string[]>;
};
export type JobExecution = {
id: string;
@@ -283,6 +314,77 @@ export type JobExecution = {
job_input_dataset_item_id: string | null;
job_output_score_id: string | null;
};
export type LegacyPrismaObservation = {
id: string;
trace_id: string | null;
project_id: string;
type: LegacyPrismaObservationType;
start_time: Generated<Timestamp>;
end_time: Timestamp | null;
name: string | null;
metadata: unknown | null;
parent_observation_id: string | null;
level: Generated<LegacyPrismaObservationLevel>;
status_message: string | null;
version: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
model: string | null;
internal_model: string | null;
internal_model_id: string | null;
modelParameters: unknown | null;
input: unknown | null;
output: unknown | null;
prompt_tokens: Generated<number>;
completion_tokens: Generated<number>;
total_tokens: Generated<number>;
unit: string | null;
input_cost: string | null;
output_cost: string | null;
total_cost: string | null;
calculated_input_cost: string | null;
calculated_output_cost: string | null;
calculated_total_cost: string | null;
completion_start_time: Timestamp | null;
prompt_id: string | null;
};
export type LegacyPrismaScore = {
id: string;
timestamp: Generated<Timestamp>;
project_id: string;
name: string;
value: number | null;
source: LegacyPrismaScoreSource;
author_user_id: string | null;
comment: string | null;
trace_id: string;
observation_id: string | null;
config_id: string | null;
string_value: string | null;
queue_id: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
data_type: Generated<ScoreDataType>;
};
export type LegacyPrismaTrace = {
id: string;
external_id: string | null;
timestamp: Generated<Timestamp>;
name: string | null;
user_id: string | null;
metadata: unknown | null;
release: string | null;
version: string | null;
project_id: string;
public: Generated<boolean>;
bookmarked: Generated<boolean>;
tags: Generated<string[]>;
input: unknown | null;
output: unknown | null;
session_id: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
};
export type LlmApiKeys = {
id: string;
created_at: Generated<Timestamp>;
@@ -339,40 +441,6 @@ export type Model = {
tokenizer_id: string | null;
tokenizer_config: unknown | null;
};
export type Observation = {
id: string;
trace_id: string | null;
project_id: string;
type: ObservationType;
start_time: Generated<Timestamp>;
end_time: Timestamp | null;
name: string | null;
metadata: unknown | null;
parent_observation_id: string | null;
level: Generated<ObservationLevel>;
status_message: string | null;
version: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
model: string | null;
internal_model: string | null;
internal_model_id: string | null;
modelParameters: unknown | null;
input: unknown | null;
output: unknown | null;
prompt_tokens: Generated<number>;
completion_tokens: Generated<number>;
total_tokens: Generated<number>;
unit: string | null;
input_cost: string | null;
output_cost: string | null;
total_cost: string | null;
calculated_input_cost: string | null;
calculated_output_cost: string | null;
calculated_total_cost: string | null;
completion_start_time: Timestamp | null;
prompt_id: string | null;
};
export type ObservationMedia = {
id: string;
project_id: string;
@@ -383,43 +451,6 @@ export type ObservationMedia = {
observation_id: string;
field: string;
};
export type ObservationView = {
id: string;
trace_id: string | null;
project_id: string;
type: ObservationType;
start_time: Timestamp;
end_time: Timestamp | null;
name: string | null;
metadata: unknown | null;
parent_observation_id: string | null;
level: Generated<ObservationLevel>;
status_message: string | null;
version: string | null;
created_at: Timestamp;
updated_at: Timestamp;
model: string | null;
modelParameters: unknown | null;
input: unknown | null;
output: unknown | null;
prompt_tokens: Generated<number>;
completion_tokens: Generated<number>;
total_tokens: Generated<number>;
unit: string | null;
completion_start_time: Timestamp | null;
prompt_id: string | null;
prompt_name: string | null;
prompt_version: number | null;
model_id: string | null;
input_price: string | null;
output_price: string | null;
total_price: string | null;
calculated_input_cost: string | null;
calculated_output_cost: string | null;
calculated_total_cost: string | null;
latency: number | null;
time_to_first_token: number | null;
};
export type Organization = {
id: string;
name: string;
@@ -484,30 +515,15 @@ export type Prompt = {
labels: Generated<string[]>;
commit_message: string | null;
};
export type QueueBackUp = {
export type PromptDependency = {
id: string;
project_id: string | null;
queue_name: string;
content: unknown;
created_at: Generated<Timestamp>;
};
export type Score = {
id: string;
timestamp: Generated<Timestamp>;
project_id: string;
name: string;
value: number | null;
source: ScoreSource;
author_user_id: string | null;
comment: string | null;
trace_id: string;
observation_id: string | null;
config_id: string | null;
string_value: string | null;
queue_id: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
data_type: Generated<ScoreDataType>;
project_id: string;
parent_id: string;
child_name: string;
child_label: string | null;
child_version: number | null;
};
export type ScoreConfig = {
id: string;
@@ -535,25 +551,6 @@ export type SsoConfig = {
auth_provider: string;
auth_config: unknown | null;
};
export type Trace = {
id: string;
external_id: string | null;
timestamp: Generated<Timestamp>;
name: string | null;
user_id: string | null;
metadata: unknown | null;
release: string | null;
version: string | null;
project_id: string;
public: Generated<boolean>;
bookmarked: Generated<boolean>;
tags: Generated<string[]>;
input: unknown | null;
output: unknown | null;
session_id: string | null;
created_at: Generated<Timestamp>;
updated_at: Generated<Timestamp>;
};
export type TraceMedia = {
id: string;
project_id: string;
@@ -570,26 +567,7 @@ export type TraceSession = {
project_id: string;
bookmarked: Generated<boolean>;
public: Generated<boolean>;
};
export type TraceView = {
id: string;
external_id: string | null;
timestamp: Generated<Timestamp>;
name: string | null;
user_id: string | null;
metadata: unknown | null;
release: string | null;
version: string | null;
project_id: string;
public: Generated<boolean>;
bookmarked: Generated<boolean>;
tags: Generated<string[]>;
input: unknown | null;
output: unknown | null;
session_id: string | null;
created_at: Timestamp;
updated_at: Timestamp;
duration: number | null;
environment: Generated<string>;
};
export type User = {
id: string;
@@ -617,6 +595,7 @@ export type DB = {
background_migrations: BackgroundMigration;
batch_exports: BatchExport;
billing_meter_backups: BillingMeterBackup;
blob_storage_integrations: BlobStorageIntegration;
comments: Comment;
cron_jobs: CronJobs;
dataset_items: DatasetItem;
@@ -631,24 +610,22 @@ export type DB = {
membership_invitations: MembershipInvitation;
models: Model;
observation_media: ObservationMedia;
observations: Observation;
observations_view: ObservationView;
observations: LegacyPrismaObservation;
organization_memberships: OrganizationMembership;
organizations: Organization;
posthog_integrations: PosthogIntegration;
prices: Price;
project_memberships: ProjectMembership;
projects: Project;
prompt_dependencies: PromptDependency;
prompts: Prompt;
queue_backups: QueueBackUp;
score_configs: ScoreConfig;
scores: Score;
scores: LegacyPrismaScore;
Session: Session;
sso_configs: SsoConfig;
trace_media: TraceMedia;
trace_sessions: TraceSession;
traces: Trace;
traces_view: TraceView;
traces: LegacyPrismaTrace;
users: User;
verification_tokens: VerificationToken;
};
@@ -0,0 +1,2 @@
-- AlterTable
ALTER TABLE "job_configurations" ADD COLUMN "time_scope" TEXT[] DEFAULT ARRAY['NEW']::TEXT[];
@@ -0,0 +1,3 @@
-- AlterTable
ALTER TABLE "trace_sessions"
ADD COLUMN "environment" TEXT NOT NULL DEFAULT 'default';
@@ -0,0 +1,2 @@
DROP VIEW IF EXISTS "observations_view";
DROP VIEW IF EXISTS "traces_view";
@@ -0,0 +1,18 @@
CREATE TABLE "prompt_dependencies" (
"id" TEXT NOT NULL,
"created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
"updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
"project_id" TEXT NOT NULL,
"parent_id" TEXT NOT NULL,
"child_name" TEXT NOT NULL,
"child_label" TEXT,
"child_version" INTEGER,
CONSTRAINT "prompt_dependencies_pkey" PRIMARY KEY ("id")
);
ALTER TABLE "prompt_dependencies" ADD CONSTRAINT "prompt_dependencies_project_id_fkey" FOREIGN KEY ("project_id") REFERENCES "projects"("id") ON DELETE CASCADE ON UPDATE CASCADE;
ALTER TABLE "prompt_dependencies" ADD CONSTRAINT "prompt_dependencies_parent_id_fkey" FOREIGN KEY ("parent_id") REFERENCES "prompts"("id") ON DELETE CASCADE ON UPDATE CASCADE;
CREATE INDEX "prompt_dependencies_project_id_parent_id" ON "prompt_dependencies"("project_id", "parent_id");
CREATE INDEX "prompt_dependencies_project_id_child_name" ON "prompt_dependencies"("project_id", "child_name");
@@ -0,0 +1,11 @@
-- CreateEnum
CREATE TYPE "AuditLogRecordType" AS ENUM ('USER', 'API_KEY');
-- AlterTable
ALTER TABLE "audit_logs" ADD COLUMN "api_key_id" TEXT,
ADD COLUMN "type" "AuditLogRecordType" NOT NULL DEFAULT 'USER',
ALTER COLUMN "user_id" DROP NOT NULL,
ALTER COLUMN "user_org_role" DROP NOT NULL;
-- CreateIndex
CREATE INDEX "audit_logs_api_key_id_idx" ON "audit_logs"("api_key_id");
@@ -0,0 +1,2 @@
-- DropTable
DROP TABLE "queue_backups";
@@ -0,0 +1,26 @@
-- CreateEnum
CREATE TYPE "BlobStorageIntegrationType" AS ENUM ('S3', 'S3_COMPATIBLE', 'AZURE_BLOB_STORAGE');
-- CreateTable
CREATE TABLE "blob_storage_integrations" (
"project_id" TEXT NOT NULL,
"type" "BlobStorageIntegrationType" NOT NULL,
"bucket_name" TEXT NOT NULL,
"prefix" TEXT NOT NULL,
"access_key_id" TEXT NOT NULL,
"secret_access_key" TEXT NOT NULL,
"region" TEXT NOT NULL,
"endpoint" TEXT,
"force_path_style" BOOLEAN NOT NULL,
"next_sync_at" TIMESTAMP(3),
"last_sync_at" TIMESTAMP(3),
"enabled" BOOLEAN NOT NULL,
"export_frequency" TEXT NOT NULL,
"created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
"updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
CONSTRAINT "blob_storage_integrations_pkey" PRIMARY KEY ("project_id")
);
-- AddForeignKey
ALTER TABLE "blob_storage_integrations" ADD CONSTRAINT "blob_storage_integrations_project_id_fkey" FOREIGN KEY ("project_id") REFERENCES "projects"("id") ON DELETE CASCADE ON UPDATE CASCADE;
@@ -1,3 +1,3 @@
# Please do not edit this file manually
# It should be added in your version-control system (i.e. Git)
# It should be added in your version-control system (e.g., Git)
provider = "postgresql"
+197 -225
View File
@@ -3,7 +3,7 @@
generator client {
provider = "prisma-client-js"
previewFeatures = ["tracing", "views", "relationJoins", "metrics"]
previewFeatures = ["views", "relationJoins", "metrics"]
}
datasource db {
@@ -109,37 +109,39 @@ model Organization {
}
model Project {
id String @id @default(cuid())
orgId String @map("org_id")
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
deletedAt DateTime? @map("deleted_at")
name String
retentionDays Int? @map("retention_days")
projectMembers ProjectMembership[]
organization Organization @relation(fields: [orgId], references: [id], onUpdate: Cascade, onDelete: Cascade)
traces Trace[]
observations Observation[]
apiKeys ApiKey[]
dataset Dataset[]
invitations MembershipInvitation[]
sessions TraceSession[]
Prompt Prompt[]
Model Model[]
EvalTemplate EvalTemplate[]
JobConfiguration JobConfiguration[]
JobExecution JobExecution[]
LlmApiKeys LlmApiKeys[]
PosthogIntegration PosthogIntegration[]
Score Score[]
scoreConfig ScoreConfig[]
BatchExport BatchExport[]
comment Comment[]
annotationQueue AnnotationQueue[]
annotationQueueItem AnnotationQueueItem[]
TraceMedia TraceMedia[]
Media Media[]
ObservationMedia ObservationMedia[]
id String @id @default(cuid())
orgId String @map("org_id")
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
deletedAt DateTime? @map("deleted_at")
name String
retentionDays Int? @map("retention_days")
projectMembers ProjectMembership[]
organization Organization @relation(fields: [orgId], references: [id], onUpdate: Cascade, onDelete: Cascade)
apiKeys ApiKey[]
dataset Dataset[]
invitations MembershipInvitation[]
sessions TraceSession[]
Prompt Prompt[]
Model Model[]
EvalTemplate EvalTemplate[]
JobConfiguration JobConfiguration[]
JobExecution JobExecution[]
LlmApiKeys LlmApiKeys[]
PosthogIntegration PosthogIntegration[]
BlobStorageIntegration BlobStorageIntegration[]
scoreConfig ScoreConfig[]
BatchExport BatchExport[]
comment Comment[]
annotationQueue AnnotationQueue[]
annotationQueueItem AnnotationQueueItem[]
TraceMedia TraceMedia[]
Media Media[]
ObservationMedia ObservationMedia[]
LegacyTrace LegacyPrismaTrace[]
LegacyObservation LegacyPrismaObservation[]
LegacyScore LegacyPrismaScore[]
PromptDependency PromptDependency[]
@@index([orgId])
@@map("projects")
@@ -267,13 +269,14 @@ enum Role {
}
model TraceSession {
id String @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
bookmarked Boolean @default(false)
public Boolean @default(false)
id String @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
bookmarked Boolean @default(false)
public Boolean @default(false)
environment String @default("default")
@@id([id, projectId])
@@index([projectId])
@@ -282,9 +285,7 @@ model TraceSession {
@@map("trace_sessions")
}
// Update TraceView below when making changes to this model!
model Trace {
model LegacyPrismaTrace {
id String @id @default(cuid())
externalId String? @map("external_id")
timestamp DateTime @default(now())
@@ -315,80 +316,43 @@ model Trace {
@@map("traces")
}
// This view is based on the trace table. Once prisma supports
// inheritance, we should remove code duplication here.
view TraceView {
// trace fields
id String @id @default(cuid())
externalId String? @map("external_id")
timestamp DateTime @default(now())
name String?
userId String? @map("user_id")
metadata Json?
release String?
version String?
projectId String @map("project_id")
public Boolean @default(false)
bookmarked Boolean @default(false)
tags String[] @default([])
input Json?
output Json?
sessionId String? @map("session_id")
createdAt DateTime @map("created_at")
updatedAt DateTime @map("updated_at")
// calculated fields
duration Float? @map("duration") // can be null if no observations in trace
@@map("traces_view")
}
// Update ObservationView below when making changes to this model!
// traceId is optional only due to timing during data injestion
// (traceId is not necessarily known at the time of observation creation)
model Observation {
id String @id @default(cuid())
traceId String? @map("trace_id")
projectId String @map("project_id")
type ObservationType
startTime DateTime @default(now()) @map("start_time")
endTime DateTime? @map("end_time")
name String?
metadata Json?
parentObservationId String? @map("parent_observation_id")
level ObservationLevel @default(DEFAULT)
statusMessage String? @map("status_message")
version String?
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
model LegacyPrismaObservation {
id String @id @default(cuid())
traceId String? @map("trace_id")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
type LegacyPrismaObservationType
startTime DateTime @default(now()) @map("start_time")
endTime DateTime? @map("end_time")
name String?
metadata Json?
parentObservationId String? @map("parent_observation_id")
level LegacyPrismaObservationLevel @default(DEFAULT)
statusMessage String? @map("status_message")
version String?
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
// GENERATION ONLY
model String? // user-provided model attribute
internalModel String? @map("internal_model") // matched model.name that is matched at ingestion time, to be deprecated
internalModelId String? @map("internal_model_id") // matched model.id that is matched at ingestion time
modelParameters Json?
input Json?
output Json?
promptTokens Int @default(0) @map("prompt_tokens")
completionTokens Int @default(0) @map("completion_tokens")
totalTokens Int @default(0) @map("total_tokens")
unit String?
model String? // user-provided model attribute
internalModel String? @map("internal_model") // matched model.name that is matched at ingestion time, to be deprecated
internalModelId String? @map("internal_model_id") // matched model.id that is matched at ingestion time
modelParameters Json?
input Json?
output Json?
promptTokens Int @default(0) @map("prompt_tokens")
completionTokens Int @default(0) @map("completion_tokens")
totalTokens Int @default(0) @map("total_tokens")
unit String?
// User provided cost at ingestion
inputCost Decimal? @map("input_cost")
outputCost Decimal? @map("output_cost")
totalCost Decimal? @map("total_cost")
inputCost Decimal? @map("input_cost")
outputCost Decimal? @map("output_cost")
totalCost Decimal? @map("total_cost")
// Calculated cost
calculatedInputCost Decimal? @map("calculated_input_cost")
calculatedOutputCost Decimal? @map("calculated_output_cost")
calculatedTotalCost Decimal? @map("calculated_total_cost")
completionStartTime DateTime? @map("completion_start_time")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
promptId String? @map("prompt_id") // no fk constraint, prompt can be deleted
calculatedInputCost Decimal? @map("calculated_input_cost")
calculatedOutputCost Decimal? @map("calculated_output_cost")
calculatedTotalCost Decimal? @map("calculated_total_cost")
completionStartTime DateTime? @map("completion_start_time")
promptId String? @map("prompt_id") // no fk constraint, prompt can be deleted
@@unique([id, projectId])
@@index([projectId, internalModel, startTime, unit])
@@ -405,88 +369,42 @@ model Observation {
@@map("observations")
}
// This view is a mix of the observation and model. Once prisma supports
// inheritance, we should remove code duplication here.
view ObservationView {
id String @id @default(cuid())
traceId String? @map("trace_id")
projectId String @map("project_id")
type ObservationType
startTime DateTime @map("start_time")
endTime DateTime? @map("end_time")
name String?
metadata Json?
parentObservationId String? @map("parent_observation_id")
level ObservationLevel @default(DEFAULT)
statusMessage String? @map("status_message")
version String?
createdAt DateTime @map("created_at")
updatedAt DateTime @map("updated_at")
// GENERATION ONLY
model String?
modelParameters Json?
input Json?
output Json?
promptTokens Int @default(0) @map("prompt_tokens")
completionTokens Int @default(0) @map("completion_tokens")
totalTokens Int @default(0) @map("total_tokens")
unit String?
completionStartTime DateTime? @map("completion_start_time")
// prompts
promptId String? @map("prompt_id")
promptName String? @map("prompt_name")
promptVersion Int? @map("prompt_version")
// model fields
modelId String? @map("model_id")
inputPrice Decimal? @map("input_price")
outputPrice Decimal? @map("output_price")
totalPrice Decimal? @map("total_price")
// calculated fields
calculatedInputCost Decimal? @map("calculated_input_cost")
calculatedOutputCost Decimal? @map("calculated_output_cost")
calculatedTotalCost Decimal? @map("calculated_total_cost")
latency Float? @map("latency")
timeToFirstToken Float? @map("time_to_first_token")
@@map("observations_view")
}
enum ObservationType {
enum LegacyPrismaObservationType {
SPAN
EVENT
GENERATION
@@map("ObservationType")
}
enum ObservationLevel {
enum LegacyPrismaObservationLevel {
DEBUG
DEFAULT
WARNING
ERROR
@@map("ObservationLevel")
}
model Score {
id String @id @default(cuid())
timestamp DateTime @default(now())
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
model LegacyPrismaScore {
id String @id @default(cuid())
timestamp DateTime @default(now())
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
name String
value Float? // always defined if data type is NUMERIC or BOOLEAN, optional for CATEGORICAL
source ScoreSource
authorUserId String? @map("author_user_id")
source LegacyPrismaScoreSource
authorUserId String? @map("author_user_id")
comment String?
traceId String @map("trace_id")
observationId String? @map("observation_id")
configId String? @map("config_id")
stringValue String? @map("string_value") // always defined if data type is CATEGORICAL or BOOLEAN, null for NUMERIC
queueId String? @map("queue_id")
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
dataType ScoreDataType @default(NUMERIC) @map("data_type")
scoreConfig ScoreConfig? @relation(fields: [configId], references: [id], onDelete: SetNull)
traceId String @map("trace_id")
observationId String? @map("observation_id")
configId String? @map("config_id")
stringValue String? @map("string_value") // always defined if data type is CATEGORICAL or BOOLEAN, null for NUMERIC
queueId String? @map("queue_id")
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
dataType ScoreDataType @default(NUMERIC) @map("data_type")
scoreConfig ScoreConfig? @relation(fields: [configId], references: [id], onDelete: SetNull)
@@unique([id, projectId]) // used for upserts via prisma
@@index(timestamp)
@@ -501,26 +419,28 @@ model Score {
@@map("scores")
}
enum ScoreSource {
enum LegacyPrismaScoreSource {
ANNOTATION
API
EVAL
@@map("ScoreSource")
}
model ScoreConfig {
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
name String
dataType ScoreDataType @map("data_type")
isArchived Boolean @default(false) @map("is_archived")
minValue Float? @map("min_value")
maxValue Float? @map("max_value")
categories Json? @map("categories")
dataType ScoreDataType @map("data_type")
isArchived Boolean @default(false) @map("is_archived")
minValue Float? @map("min_value")
maxValue Float? @map("max_value")
categories Json? @map("categories")
description String?
score Score[]
legacyScore LegacyPrismaScore[]
@@unique([id, projectId]) // used for upserts via prisma
@@index([dataType])
@@ -722,15 +642,16 @@ model Prompt {
createdBy String @map("created_by")
prompt Json
name String
version Int
type String @default("text")
isActive Boolean? @map("is_active") // Deprecated. To be removed once 'production' labels work as expected.
config Json @default("{}") @db.Json
tags String[] @default([])
labels String[] @default([])
commitMessage String? @map("commit_message")
prompt Json
name String
version Int
type String @default("text")
isActive Boolean? @map("is_active") // Deprecated. To be removed once 'production' labels work as expected.
config Json @default("{}") @db.Json
tags String[] @default([])
labels String[] @default([])
commitMessage String? @map("commit_message")
PromptDependency PromptDependency[]
@@unique([projectId, name, version])
@@index([projectId, id])
@@ -740,6 +661,26 @@ model Prompt {
@@map("prompts")
}
model PromptDependency {
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
projectId String @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
parentId String @map("parent_id")
parent Prompt @relation(fields: [parentId], references: [id], onDelete: Cascade)
childName String @map("child_name")
childLabel String? @map("child_label")
childVersion Int? @map("child_version")
@@index([projectId, parentId], map: "prompt_dependencies_project_id_parent_id")
@@index([projectId, childName], map: "prompt_dependencies_project_id_child_name")
@@map("prompt_dependencies")
}
// Update ObservationView below when making changes to this model!
model Model {
id String @id @default(cuid())
@@ -778,23 +719,31 @@ model Price {
@@map("prices")
}
enum AuditLogRecordType {
USER
API_KEY
}
// No FK constraints to preserve audit logs
model AuditLog {
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
userId String @map("user_id")
orgId String @map("org_id")
userOrgRole String @map("user_org_role")
projectId String? @map("project_id")
userProjectRole String? @map("user_project_role")
resourceType String @map("resource_type")
resourceId String @map("resource_id")
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
type AuditLogRecordType @default(USER)
apiKeyId String? @map("api_key_id")
userId String? @map("user_id")
orgId String @map("org_id")
userOrgRole String? @map("user_org_role")
projectId String? @map("project_id")
userProjectRole String? @map("user_project_role")
resourceType String @map("resource_type")
resourceId String @map("resource_id")
action String
before String? //stringified JSON
before String? // stringified JSON
after String? // stringified JSON
@@index([projectId])
@@index([apiKeyId])
@@index([userId])
@@index([orgId])
@@index([createdAt])
@@ -853,6 +802,7 @@ model JobConfiguration {
variableMapping Json @map("variable_mapping")
sampling Decimal // ratio of jobs that are executed for sampling (0..1)
delay Int // delay in milliseconds
timeScope String[] @default(["NEW"]) @map("time_scope")
JobExecution JobExecution[]
@@index([projectId, id])
@@ -911,7 +861,6 @@ model SsoConfig {
authConfig Json? @map("auth_config")
// e.g. { "clientId": "1234", "clientSecret": "5678" }, null if credentials from env should be used
// secrets like clientSecret are encrypted on the application level
@@map("sso_configs")
}
@@ -927,6 +876,40 @@ model PosthogIntegration {
@@map("posthog_integrations")
}
model BlobStorageIntegration {
projectId String @id @map("project_id")
project Project @relation(fields: [projectId], references: [id], onDelete: Cascade)
// Blob Storage Config
type BlobStorageIntegrationType @map("type")
bucketName String @map("bucket_name")
prefix String @map("prefix")
accessKeyId String @map("access_key_id")
secretAccessKey String @map("secret_access_key")
region String @map("region")
endpoint String? @map("endpoint")
forcePathStyle Boolean @map("force_path_style")
// Integration Config
nextSyncAt DateTime? @map("next_sync_at")
lastSyncAt DateTime? @map("last_sync_at")
enabled Boolean
exportFrequency String @map("export_frequency")
createdAt DateTime @default(now()) @map("created_at")
updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
@@map("blob_storage_integrations")
}
enum BlobStorageIntegrationType {
S3
S3_COMPATIBLE
AZURE_BLOB_STORAGE
@@map("BlobStorageIntegrationType")
}
model BatchExport {
id String @id @default(cuid())
createdAt DateTime @default(now()) @map("created_at")
@@ -1004,17 +987,6 @@ model ObservationMedia {
@@map("observation_media")
}
model QueueBackUp {
id String @id @default(cuid())
projectId String? @map("project_id")
queueName String @map("queue_name")
content Json
createdAt DateTime @default(now()) @map("created_at")
@@map("queue_backups")
}
model BillingMeterBackup {
// unique
stripeCustomerId String @map("stripe_customer_id")
+28 -113
View File
@@ -2,8 +2,6 @@ import {
PrismaClient,
type Project,
type Prisma,
ObservationType,
ScoreSource,
ScoreDataType,
AnnotationQueueObjectType,
} from "../src/index";
@@ -74,6 +72,9 @@ async function main() {
where: { id: seedOrgId },
update: {
name: "Seed Org",
cloudConfig: {
plan: "Team",
},
},
create: {
id: seedOrgId,
@@ -266,38 +267,23 @@ async function main() {
const traceVolume = environment === "load" ? LOAD_TRACE_VOLUME : 100;
const {
traces,
observations,
scores,
sessions,
events,
comments,
queueItems,
} = createObjects(
traceVolume,
envTags,
colorTags,
project1,
project2,
promptIds,
queueIds,
configIdsAndNames,
);
const { traces, observations, scores, sessions, comments, queueItems } =
createObjects(
traceVolume,
envTags,
colorTags,
project1,
project2,
promptIds,
queueIds,
configIdsAndNames,
);
logger.info(
`Seeding ${traces.length} traces, ${observations.length} observations, and ${scores.length} scores`,
);
await uploadObjects(
traces,
observations,
scores,
sessions,
events,
comments,
queueItems,
);
await uploadObjects(sessions, comments, queueItems);
// If openai key is in environment, add it to the projects LLM API keys
const OPENAI_API_KEY = process.env.OPENAI_API_KEY;
@@ -420,7 +406,7 @@ export async function createDatasets(
updatedAt: Date;
name: string;
},
observations: Prisma.ObservationCreateManyInput[],
observations: { id?: string; projectId?: string; traceId?: string | null }[],
) {
for (let datasetNumber = 0; datasetNumber < 2; datasetNumber++) {
for (const projectId of [project1.id, project2.id]) {
@@ -532,11 +518,7 @@ export async function createDatasets(
}
async function uploadObjects(
traces: Prisma.TraceCreateManyInput[],
observations: Prisma.ObservationCreateManyInput[],
scores: Prisma.ScoreCreateManyInput[],
sessions: Prisma.TraceSessionCreateManyInput[],
events: Prisma.ObservationCreateManyInput[],
comments: Prisma.CommentCreateManyInput[],
queueItems: Prisma.AnnotationQueueItemCreateManyInput[],
) {
@@ -564,73 +546,6 @@ async function uploadObjects(
await promises[i];
}
promises = [];
chunk(traces, chunkSize).forEach((chunk) => {
promises.push(
prisma.trace.createMany({
data: chunk,
}),
);
});
for (let i = 0; i < promises.length; i++) {
if (i + 1 >= promises.length || i % Math.ceil(promises.length / 10) === 0)
logger.info(
`Seeding of Traces ${((i + 1) / promises.length) * 100}% complete`,
);
await promises[i];
}
promises = [];
chunk(observations, chunkSize).forEach((chunk) => {
promises.push(
prisma.observation.createMany({
data: chunk,
}),
);
});
for (let i = 0; i < promises.length; i++) {
if (i + 1 >= promises.length || i % Math.ceil(promises.length / 10) === 0)
logger.info(
`Seeding of Observations ${((i + 1) / promises.length) * 100}% complete`,
);
await promises[i];
}
promises = [];
chunk(events, chunkSize).forEach((chunk) => {
promises.push(
prisma.observation.createMany({
data: chunk,
}),
);
});
for (let i = 0; i < promises.length; i++) {
if (i + 1 >= promises.length || i % Math.ceil(promises.length / 10) === 0)
logger.info(
`Seeding of Events ${((i + 1) / promises.length) * 100}% complete`,
);
await promises[i];
}
promises = [];
chunk(scores, chunkSize).forEach((chunk) => {
promises.push(
prisma.score.createMany({
data: chunk,
}),
);
});
for (let i = 0; i < promises.length; i++) {
if (i + 1 >= promises.length || i % Math.ceil(promises.length / 10) === 0)
logger.info(
`Seeding of Scores ${((i + 1) / promises.length) * 100}% complete`,
);
await promises[i];
}
promises = [];
chunk(comments, chunkSize).forEach((chunk) => {
promises.push(
@@ -682,11 +597,11 @@ function createObjects(
}[]
>,
) {
const traces: Prisma.TraceCreateManyInput[] = [];
const observations: Prisma.ObservationCreateManyInput[] = [];
const scores: Prisma.ScoreCreateManyInput[] = [];
const traces: any[] = [];
const observations: any[] = [];
const scores: any[] = [];
const sessions: Prisma.TraceSessionCreateManyInput[] = [];
const events: Prisma.ObservationCreateManyInput[] = [];
const events: any[] = [];
const configs: Prisma.ScoreConfigCreateManyInput[] = [];
const comments: Prisma.CommentCreateManyInput[] = [];
const queueItems: Prisma.AnnotationQueueItemCreateManyInput[] = [];
@@ -795,7 +710,7 @@ function createObjects(
name: annotationScoreName,
timestamp: traceTs,
createdAt: traceTs,
source: ScoreSource.ANNOTATION,
source: "ANNOTATION",
projectId,
authorUserId: `user-${i}`,
dataType,
@@ -812,7 +727,7 @@ function createObjects(
value: Math.floor(Math.random() * 10) - 5,
timestamp: traceTs,
createdAt: traceTs,
source: ScoreSource.API,
source: "API",
projectId,
dataType: ScoreDataType.NUMERIC,
},
@@ -825,7 +740,7 @@ function createObjects(
name: "Completeness",
timestamp: traceTs,
createdAt: traceTs,
source: ScoreSource.API,
source: "API",
projectId,
dataType: ScoreDataType.CATEGORICAL,
stringValue:
@@ -859,7 +774,7 @@ function createObjects(
);
const span = {
type: ObservationType.SPAN,
type: "SPAN",
id: `span-${v4()}`,
startTime: spanTsStart,
createdAt: spanTsStart,
@@ -933,7 +848,7 @@ function createObjects(
const { input, output } = getGenerationInputOutput();
const generation = {
type: ObservationType.GENERATION,
type: "GENERATION",
id: `generation-${v4()}`,
startTime: generationTsStart,
createdAt: generationTsStart,
@@ -978,7 +893,7 @@ function createObjects(
value: Math.random() * 2 - 1,
observationId: generation.id,
traceId: trace.id,
source: ScoreSource.API,
source: "API",
projectId: trace.projectId,
timestamp: generationTsEnd,
createdAt: traceTs,
@@ -989,7 +904,7 @@ function createObjects(
value: Math.random() * 2 - 1,
observationId: generation.id,
traceId: trace.id,
source: ScoreSource.API,
source: "API",
projectId: trace.projectId,
timestamp: generationTsEnd,
createdAt: traceTs,
@@ -1013,7 +928,7 @@ function createObjects(
);
events.push({
type: ObservationType.EVENT,
type: "EVENT",
id: `event-${v4()}`,
startTime: eventTs,
createdAt: eventTs,
+3
View File
@@ -27,6 +27,9 @@ const prepareProjectsAndApiKeys = async (
create: {
id: orgId,
name: `Organization for ${projectId}`,
cloudConfig: {
plan: "Team",
},
},
});
+4 -1
View File
@@ -67,6 +67,7 @@ export const prepareClickhouse = async (
concat('release_', toString(randUniform(0, 100))) AS release,
concat('version_', toString(randUniform(0, 100))) AS version,
'${projectId}' AS project_id,
'default' AS environment,
if(rand() < 0.8, true, false) as public,
if(rand() < 0.8, true, false) as bookmarked,
array('tag1', 'tag2') as tags,
@@ -85,6 +86,7 @@ export const prepareClickhouse = async (
SELECT toString(number) AS id,
toString(floor(randUniform(0, ${tracesPerProject}))) AS trace_id,
'${projectId}' AS project_id,
'default' AS environment,
if(randUniform(0, 1) < 0.47, 'GENERATION', if(randUniform(0, 1) < 0.94, 'SPAN', 'EVENT')) AS type,
toString(rand()) AS parent_observation_id,
toDateTime(now() - randUniform(0, ${opts.numberOfDays} * 24 * 60 * 60)) AS start_time,
@@ -143,9 +145,10 @@ export const prepareClickhouse = async (
const scoresQuery = `
INSERT INTO scores
SELECT toString(floor(randUniform(0, 100))) AS id,
SELECT toString(number) AS id,
toDateTime(now() - randUniform(0, ${opts.numberOfDays} * 24 * 60 * 60)) AS timestamp,
'${projectId}' AS project_id,
'default' AS environment,
toString(floor(randUniform(0, ${tracesPerProject}))) AS trace_id,
if(
rand() > 0.9,
+3 -3
View File
@@ -2,7 +2,7 @@ import crypto from "crypto";
import { env } from "../env";
const ENCRYPTION_KEY: string | undefined = env.ENCRYPTION_KEY; // Must be 256 bits (32 bytes, 64 hex characters)
const IV_LENGTH: number = 16; // For AES, this is always 16
const IV_LENGTH: number = 12; // For AES-GCM, this is always 12
// Alternatively: openssl rand -hex 32
export function keyGen() {
@@ -23,7 +23,7 @@ export function encrypt(plainText: string): string {
const cipher = crypto.createCipheriv(
"aes-256-gcm",
Buffer.from(ENCRYPTION_KEY, "hex"),
iv
iv,
);
let encrypted = cipher.update(plainText, "utf8", "hex");
encrypted += cipher.final("hex");
@@ -49,7 +49,7 @@ export function decrypt(text: string): string {
const decipher = crypto.createDecipheriv(
"aes-256-gcm",
Buffer.from(ENCRYPTION_KEY, "hex"),
iv
iv,
);
decipher.setAuthTag(authTag);
+6 -1
View File
@@ -10,7 +10,7 @@ const EnvSchema = z.object({
REDIS_PORT: z.coerce
.number({
description:
".env files convert numbers to strings, therefoore we have to enforce them to be numbers",
".env files convert numbers to strings, therefore we have to enforce them to be numbers",
})
.positive()
.max(65536, `options.port should be >= 0 and < 65536`)
@@ -50,6 +50,7 @@ const EnvSchema = z.object({
ENABLE_AWS_CLOUDWATCH_METRIC_PUBLISHING: z
.enum(["true", "false"])
.default("false"),
LANGFUSE_S3_CONCURRENT_WRITES: z.coerce.number().positive().default(50),
LANGFUSE_S3_EVENT_UPLOAD_BUCKET: z.string({
required_error: "Langfuse requires a bucket name for S3 Event Uploads.",
}),
@@ -70,6 +71,10 @@ const EnvSchema = z.object({
LANGFUSE_POSTGRES_METERING_DATA_EXPORT_IS_ENABLED: z
.enum(["true", "false"])
.default("false"),
LANGFUSE_CUSTOM_SSO_EMAIL_CLAIM: z.string().default("email"),
LANGFUSE_CUSTOM_SSO_NAME_CLAIM: z.string().default("name"),
LANGFUSE_CUSTOM_SSO_SUB_CLAIM: z.string().default("sub"),
});
export const env: z.infer<typeof EnvSchema> =
@@ -28,6 +28,7 @@ const BooleanData = z.object({
const CreateAnnotationScoreBase = z.object({
name: z.string(),
projectId: z.string(),
environment: z.string().default("default"),
traceId: z.string(),
configId: z.string().optional(),
observationId: z.string().optional(),
@@ -10,6 +10,7 @@ export enum BatchActionType {
}
const ActionIdSchema = z.enum([
"score-delete",
"trace-delete",
"trace-add-to-annotation-queue",
]);
@@ -18,9 +18,11 @@ export enum BatchExportFileFormat {
}
export enum BatchExportTableName {
Scores = "scores",
Sessions = "sessions",
Traces = "traces",
Generations = "generations",
Observations = "observations",
DatasetRunItems = "dataset_run_items",
}
export const exportOptions: Record<
@@ -1,8 +1,19 @@
// used on organization.cloudConfig.plan
export const cloudConfigPlans = [
"Hobby",
"Core",
"Pro",
"Team",
"Enterprise",
] as const;
export const planLabels = {
oss: "OSS",
"cloud:hobby": "Hobby",
"cloud:core": "Core",
"cloud:pro": "Pro",
"cloud:team": "Team",
"cloud:enterprise": "Enterprise",
"self-hosted:pro": "Pro (self-hosted)",
"self-hosted:enterprise": "Enterprise (self-hosted)",
} as const;
@@ -12,8 +23,9 @@ export type Plan = keyof typeof planLabels;
export const plans = Object.keys(planLabels) as Plan[];
// These functions are kept here to ensure consistency when updating plan names in the future.
export const isCloudPlan = (plan: Plan) => plan.startsWith("cloud");
export const isSelfHostedPlan = (plan: Plan) => plan.startsWith("self-hosted");
export const isCloudPlan = (plan?: Plan) => plan?.startsWith("cloud");
export const isSelfHostedPlan = (plan?: Plan) =>
plan?.startsWith("self-hosted");
export const isPlan = (value: string): value is Plan =>
plans.includes(value as Plan);
@@ -107,3 +107,8 @@ export const OutputSchema = z.object({
});
export const DEFAULT_TRACE_JOB_DELAY = 10_000;
export const JobTimeScopeZod = z.enum(["NEW", "EXISTING"]);
export type JobTimeScope = z.infer<typeof JobTimeScopeZod>;
export const TimeScopeSchema = z.array(JobTimeScopeZod).default(["NEW"]);
@@ -0,0 +1,61 @@
import { z } from "zod";
export const PromptDependencyRegex = /@@@langfusePrompt:(.*?)@@@/g;
export const ParsedPromptDependencySchema = z.union([
z.object({
name: z.string(),
type: z.literal("version"),
version: z.coerce.number(),
}),
z.object({ name: z.string(), type: z.literal("label"), label: z.string() }),
]);
export type ParsedPromptDependencyTag = z.infer<
typeof ParsedPromptDependencySchema
>;
export function parsePromptDependencyTags(
content: string | object,
): ParsedPromptDependencyTag[] {
const matchedTags = JSON.stringify(content).match(PromptDependencyRegex);
const validTags: ParsedPromptDependencyTag[] = [];
for (const match of new Set(matchedTags ?? [])) {
const innerContent = match.replace(/^@@@langfusePrompt:|@@@$/g, "");
const parts = innerContent.split("|");
const params: Record<string, string> = {};
// Check if the first parameter is name
const firstPart = parts[0];
if (!firstPart || !firstPart.startsWith("name=")) {
continue; // Skip this tag if name is not the first parameter. This makes it easier to replace the tag with the resolved prompt.
}
// There can be only 2 parts
if (parts.length !== 2) {
continue;
}
parts.forEach((part) => {
const [key, value] = part.split("=");
params[key] = value;
});
if (params.name) {
const parsed = ParsedPromptDependencySchema.safeParse({
name: params.name,
...(params.version
? { version: params.version, type: "version" }
: { label: params.label, type: "label" }),
});
if (parsed.success) {
validTags.push(parsed.data);
}
}
}
return validTags;
}
@@ -1,8 +1,7 @@
import { z } from "zod";
import { Score } from "@prisma/client";
import { isPresent, stringDateTime } from "../../utils/typeChecks";
import { Score } from "../../server";
import {
NonEmptyString,
paginationMetaResponseZod,
@@ -46,6 +45,7 @@ const ScoreBase = z.object({
id: z.string(),
timestamp: z.coerce.date(),
projectId: z.string(),
environment: z.string().default("default"),
name: z.string(),
source: z.enum(ScoreSource),
authorUserId: z.string().nullish(),
@@ -64,6 +64,7 @@ const BaseScoreBody = z.object({
traceId: z.string(),
observationId: z.string().nullish(),
comment: z.string().nullish(),
environment: z.string().default("default"),
});
/**
@@ -237,6 +238,7 @@ export const GetScoresQuery = z.object({
configId: z.string().nullish(),
queueId: z.string().nullish(),
traceTags: z.union([z.array(z.string()), z.string()]).nullish(),
environment: z.union([z.array(z.string()), z.string()]).nullish(),
name: z.string().nullish(),
fromTimestamp: stringDateTime,
toTimestamp: stringDateTime,
@@ -259,6 +261,7 @@ const LegacyGetScoreResponseDataV1 = z.intersection(
trace: z.object({
userId: z.string().nullish(),
tags: z.array(z.string()).nullish(),
environment: z.string().nullish(),
}),
}),
);
+3 -2
View File
@@ -1,4 +1,5 @@
import { ScoreDataType, ScoreSource } from "@prisma/client";
import { ScoreDataType } from "@prisma/client";
import { ScoreSourceType } from "../../server";
export type CategoricalAggregate = {
type: "CATEGORICAL";
@@ -22,7 +23,7 @@ export type ScoreAggregate = Record<
export type ScoreSimplified = {
name: string;
dataType: ScoreDataType;
source: ScoreSource;
source: ScoreSourceType;
value?: number | null;
comment?: string | null;
stringValue?: string | null;
+4
View File
@@ -38,9 +38,13 @@ export * from "./features/comments/types";
// experiments
export * from "./features/experiments/utils";
// prompts
export * from "./features/prompts/parsePromptDependencyTags";
// export db types only
export * from "@prisma/client";
export { type DB } from "../prisma/generated/types";
export * from "./server/repositories/types";
// errors
export * from "./errors/index";
@@ -1,8 +1,9 @@
import { z } from "zod";
import { CloudConfigRateLimit } from "./rate-limits";
import { cloudConfigPlans } from "../features/entitlements/plans";
export const CloudConfigSchema = z.object({
plan: z.enum(["Hobby", "Pro", "Team", "Enterprise"]).optional(),
plan: z.enum(cloudConfigPlans).optional(),
monthlyObservationLimit: z.number().int().positive().optional(),
// used for table and dashboard queries
defaultLookBackDays: z.number().int().positive().optional(),
+8 -3
View File
@@ -1,8 +1,8 @@
import { ObservationLevel } from "@prisma/client";
import {
type OptionsDefinition,
type ColumnDefinition,
} from "./tableDefinitions/types";
} from "./tableDefinitions";
import { ObservationLevelType } from "./server";
// to be used server side
export const observationsTableCols: ColumnDefinition[] = [
@@ -102,7 +102,12 @@ export const observationsTableCols: ColumnDefinition[] = [
id: "level",
type: "stringOptions",
internal: 'o."level"::text',
options: Object.values(ObservationLevel).map((value) => ({ value })),
options: [
{ value: "DEBUG" },
{ value: "DEFAULT" },
{ value: "WARNING" },
{ value: "ERROR" },
] as { value: ObservationLevelType }[],
},
{
name: "Status Message",
@@ -1,4 +1,9 @@
import type { OAuthConfig, OAuthUserConfig } from "next-auth/providers/oauth";
import { env } from "../../env";
const CUSTOM_EMAIL_CLAIM = env.LANGFUSE_CUSTOM_SSO_EMAIL_CLAIM;
const CUSTOM_NAME_CLAIM = env.LANGFUSE_CUSTOM_SSO_NAME_CLAIM;
const CUSTOM_SUB_CLAIM = env.LANGFUSE_CUSTOM_SSO_SUB_CLAIM;
interface CustomSSOUser extends Record<string, any> {
email: string;
@@ -7,6 +12,7 @@ interface CustomSSOUser extends Record<string, any> {
verified: boolean;
}
export function CustomSSOProvider<P extends CustomSSOUser>(
options: OAuthUserConfig<P>
): OAuthConfig<P> {
@@ -19,9 +25,9 @@ export function CustomSSOProvider<P extends CustomSSOUser>(
checks: ["pkce", "state"],
profile(profile) {
return {
id: profile.sub,
name: profile.name,
email: profile.email,
id: profile[CUSTOM_SUB_CLAIM],
name: profile[CUSTOM_NAME_CLAIM],
email: profile[CUSTOM_EMAIL_CLAIM],
image: null,
};
},
+1
View File
@@ -43,4 +43,5 @@ export type ApiAccessScope = {
orgId: string;
plan: Plan;
rateLimitOverrides: z.infer<typeof CloudConfigRateLimit>;
apiKeyId: string;
};
@@ -29,6 +29,7 @@ export const clickhouseClient = (
log_comment: JSON.stringify(params.tags ?? {}),
async_insert: 1,
wait_for_async_insert: 1, // if disabled, we won't get errors from clickhouse
input_format_json_throw_on_bad_escape_sequence: 0,
},
});
};
@@ -2,6 +2,11 @@ export const ClickhouseTableNames = {
traces: "traces",
observations: "observations",
scores: "scores",
// Virtual tables for dashboards
// TODO: Check if we can do this more elegantly
scores_numeric: "scores_numeric",
scores_categorical: "scores_categorical",
} as const;
export type ClickhouseTableName = keyof typeof ClickhouseTableNames;
+5 -1
View File
@@ -17,22 +17,25 @@ export * from "./clickhouse/schemaUtils";
export * from "./clickhouse/schema";
export * from "./repositories/definitions";
export * from "../server/ingestion/types";
export * from "./ingestion/modelMatch";
export * from "./ingestion/processEventBatch";
export * from "../server/ingestion/types";
export * from "../server/ingestion/validateAndInflateScore";
export * from "./redis/redis";
export * from "./redis/traceUpsert";
export * from "./redis/createEvalQueue";
export * from "./redis/cloudUsageMeteringQueue";
export * from "./redis/getQueue";
export * from "./redis/traceDelete";
export * from "./redis/projectDelete";
export * from "./redis/scoreDelete";
export * from "./redis/datasetRunItemUpsert";
export * from "./redis/batchExport";
export * from "./redis/batchActionQueue";
export * from "./redis/ingestionQueue";
export * from "./redis/postHogIntegrationQueue";
export * from "./redis/postHogIntegrationProcessingQueue";
export * from "./redis/blobStorageIntegrationQueue";
export * from "./redis/blobStorageIntegrationProcessingQueue";
export * from "./redis/dataRetentionQueue";
export * from "./redis/dataRetentionProcessingQueue";
export * from "./redis/coreDataS3ExportQueue";
@@ -48,6 +51,7 @@ export * from "./queries";
export * from "./repositories";
export * from "./redis/evalExecutionQueue";
export * from "./services/sessions-ui-table-service";
export * from "./services/datasets-ui-table-service";
// test utils
export * from "./test-utils";
@@ -1,63 +0,0 @@
import { Model, Observation, Prisma } from "@prisma/client";
import { prisma } from "../../db";
export async function findModel(p: {
event: {
projectId: string;
model?: string;
unit?: string;
startTime?: Date;
};
existingDbObservation?: Omit<Observation, "input" | "output">;
}): Promise<Model | null> {
const { event, existingDbObservation } = p;
// either get the model from the existing observation
// or match pattern on the user provided model name
const modelCondition = event.model
? Prisma.sql`AND ${event.model} ~ match_pattern`
: existingDbObservation?.internalModel
? Prisma.sql`AND model_name = ${existingDbObservation.internalModel}`
: undefined;
if (!modelCondition) return null;
// unit based on the current event or the existing observation, both can be undefined
const mergedUnit = event.unit ?? existingDbObservation?.unit;
const unitCondition = mergedUnit
? Prisma.sql`AND unit = ${mergedUnit}`
: Prisma.empty;
const sql = Prisma.sql`
SELECT
id,
created_at AS "createdAt",
updated_at AS "updatedAt",
project_id AS "projectId",
model_name AS "modelName",
match_pattern AS "matchPattern",
start_date AS "startDate",
input_price AS "inputPrice",
output_price AS "outputPrice",
total_price AS "totalPrice",
unit,
tokenizer_id AS "tokenizerId",
tokenizer_config AS "tokenizerConfig"
FROM
models
WHERE (project_id = ${event.projectId}
OR project_id IS NULL)
${modelCondition}
${unitCondition}
AND (start_date IS NULL OR start_date <= ${
event.startTime ? new Date(event.startTime) : new Date()
}::timestamp with time zone at time zone 'UTC')
ORDER BY
project_id ASC,
start_date DESC NULLS LAST
LIMIT 1
`;
const foundModels = await prisma.$queryRaw<Array<Model>>(sql);
return foundModels[0] ?? null;
}
@@ -14,6 +14,7 @@ import {
getCurrentSpan,
instrumentAsync,
instrumentSync,
recordDistribution,
recordIncrement,
traceException,
} from "../instrumentation";
@@ -94,6 +95,8 @@ export const processEventBatch = async (
// add context of api call to the span
const currentSpan = getCurrentSpan();
recordIncrement("langfuse.ingestion.event", input.length);
recordDistribution("langfuse.ingestion.event_distribution", input.length);
currentSpan?.setAttribute("langfuse.ingestion.batch_size", input.length);
currentSpan?.setAttribute("langfuse.project.id", authCheck.scope.projectId);
currentSpan?.setAttribute("langfuse.org.id", authCheck.scope.orgId);
+138 -80
View File
@@ -3,7 +3,17 @@ import { z } from "zod";
import { NonEmptyString, jsonSchema } from "../../utils/zod";
import { ModelUsageUnit } from "../../constants";
import { ObservationLevel, ScoreSource } from "@prisma/client";
import { type ScoreSourceType } from "../repositories";
export const idSchema = z
.string()
.min(1)
.max(800) // AWS S3 allows for 1024 bytes for object keys and we need enough room to construct the entire key/path. https://docs.aws.amazon.com/AmazonS3/latest/userguide/object-keys.html
.refine((id) => !id.includes("\r"), {
message: "ID cannot contain carriage return characters",
});
const ObservationLevel = z.enum(["DEBUG", "DEFAULT", "WARNING", "ERROR"]);
export const Usage = z.object({
input: z.number().int().nullish(),
@@ -56,21 +66,25 @@ export const usage = MixedUsage.nullish()
// ensure output is always of new usage model
.pipe(Usage.nullish());
const RawUsageOrCostDetails = z.record(
const CostDetails = z
.record(z.string(), z.number().nonnegative().nullish())
.nullish();
const RawUsageDetails = z.record(
z.string(),
z.number().nonnegative().nullish(),
z.number().int().nonnegative().nullish(),
);
const OpenAIUsageSchema = z
const OpenAICompletionUsageSchema = z
.object({
prompt_tokens: z.number().nonnegative(),
completion_tokens: z.number().nonnegative(),
total_tokens: z.number().nonnegative(),
prompt_tokens: z.number().int().nonnegative(),
completion_tokens: z.number().int().nonnegative(),
total_tokens: z.number().int().nonnegative(),
prompt_tokens_details: z
.record(z.string(), z.number().nonnegative())
.record(z.string(), z.number().int().nonnegative().nullish())
.nullish(),
completion_tokens_details: z
.record(z.string(), z.number().nonnegative())
.record(z.string(), z.number().int().nonnegative().nullish())
.nullish(),
})
.strict()
@@ -84,7 +98,7 @@ const OpenAIUsageSchema = z
prompt_tokens_details,
completion_tokens_details,
} = v;
const result: z.infer<typeof RawUsageOrCostDetails> & {
const result: z.infer<typeof RawUsageDetails> & {
input: number;
output: number;
total: number;
@@ -96,31 +110,104 @@ const OpenAIUsageSchema = z
if (prompt_tokens_details) {
for (const [key, value] of Object.entries(prompt_tokens_details)) {
result[`input_${key}`] = value;
result.input = Math.max(result.input - (value ?? 0), 0);
if (value !== null && value !== undefined) {
result[`input_${key}`] = value;
result.input = Math.max(result.input - (value ?? 0), 0);
}
}
}
if (completion_tokens_details) {
for (const [key, value] of Object.entries(completion_tokens_details)) {
result[`output_${key}`] = value;
result.output = Math.max(result.output - (value ?? 0), 0);
if (value !== null && value !== undefined) {
result[`output_${key}`] = value;
result.output = Math.max(result.output - (value ?? 0), 0);
}
}
}
return result;
})
.pipe(RawUsageOrCostDetails);
.pipe(RawUsageDetails);
export const UsageOrCostDetails = z
.union([OpenAIUsageSchema, RawUsageOrCostDetails])
// The new OpenAI Response API uses a new Usage schema that departs from the Completion API Usage schema
const OpenAIResponseUsageSchema = z
.object({
input_tokens: z.number().int().nonnegative(),
output_tokens: z.number().int().nonnegative(),
total_tokens: z.number().int().nonnegative(),
input_tokens_details: z
.record(z.string(), z.number().int().nonnegative().nullish())
.nullish(),
output_tokens_details: z
.record(z.string(), z.number().int().nonnegative().nullish())
.nullish(),
})
.strict()
.transform((v) => {
if (!v) return;
const {
input_tokens,
output_tokens,
total_tokens,
input_tokens_details,
output_tokens_details,
} = v;
const result: z.infer<typeof RawUsageDetails> & {
input: number;
output: number;
total: number;
} = {
input: input_tokens,
output: output_tokens,
total: total_tokens,
};
if (input_tokens_details) {
for (const [key, value] of Object.entries(input_tokens_details)) {
if (value !== null && value !== undefined) {
result[`input_${key}`] = value;
result.input = Math.max(result.input - (value ?? 0), 0);
}
}
}
if (output_tokens_details) {
for (const [key, value] of Object.entries(output_tokens_details)) {
if (value !== null && value !== undefined) {
result[`output_${key}`] = value;
result.output = Math.max(result.output - (value ?? 0), 0);
}
}
}
return result;
})
.pipe(RawUsageDetails);
export const UsageDetails = z
.union([
OpenAICompletionUsageSchema,
OpenAIResponseUsageSchema,
RawUsageDetails,
])
.nullish();
export const EnvironmentName = z
.string()
.max(40, "Maximum length is 40 characters")
.regex(
/^(?!langfuse)[a-z0-9-_]+$/,
"Only alphanumeric lower case characters, hyphens, and underscores are allowed, and it must not start with 'langfuse'",
)
.default("default");
// Using z.any instead of jsonSchema for input/output as we saw huge CPU overhead for large numeric arrays.
// With this setup parsing should be more lightweight and doesn't block other requests.
// As we allow plain values, arrays, and objects the JSON parse via bodyParser should suffice.
export const TraceBody = z.object({
id: z.string().nullish(),
id: idSchema.nullish(),
timestamp: stringDateTime,
name: z.string().max(1000).nullish(),
externalId: z.string().nullish(),
@@ -128,6 +215,7 @@ export const TraceBody = z.object({
output: z.any().nullish(),
sessionId: z.string().nullish(),
userId: z.string().nullish(),
environment: EnvironmentName,
metadata: jsonSchema.nullish(),
release: z.string().nullish(),
version: z.string().nullish(),
@@ -136,24 +224,25 @@ export const TraceBody = z.object({
});
export const OptionalObservationBody = z.object({
traceId: z.string().nullish(),
traceId: idSchema.nullish(),
environment: EnvironmentName,
name: z.string().nullish(),
startTime: stringDateTime,
metadata: jsonSchema.nullish(),
input: z.any().nullish(),
output: z.any().nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
parentObservationId: z.string().nullish(),
version: z.string().nullish(),
});
export const CreateEventEvent = OptionalObservationBody.extend({
id: NonEmptyString,
id: idSchema,
});
export const UpdateEventEvent = OptionalObservationBody.extend({
id: NonEmptyString,
id: idSchema,
});
export const CreateSpanBody = CreateEventEvent.extend({
@@ -182,8 +271,8 @@ export const CreateGenerationBody = CreateSpanBody.extend({
)
.nullish(),
usage: usage,
usageDetails: UsageOrCostDetails,
costDetails: UsageOrCostDetails,
usageDetails: UsageDetails,
costDetails: CostDetails,
promptName: z.string().nullish(),
promptVersion: z.number().int().nullish(),
}).refine((value) => {
@@ -212,8 +301,8 @@ export const UpdateGenerationBody = UpdateSpanBody.extend({
)
.nullish(),
usage: usage,
usageDetails: UsageOrCostDetails,
costDetails: UsageOrCostDetails,
usageDetails: UsageDetails,
costDetails: CostDetails,
promptName: z.string().nullish(),
promptVersion: z.number().int().nullish(),
}).refine((value) => {
@@ -225,12 +314,15 @@ export const UpdateGenerationBody = UpdateSpanBody.extend({
});
const BaseScoreBody = z.object({
id: z.string().nullish(),
id: idSchema.nullish(),
name: NonEmptyString,
traceId: z.string(),
environment: EnvironmentName,
observationId: z.string().nullish(),
comment: z.string().nullish(),
source: z.nativeEnum(ScoreSource).default(ScoreSource.API),
source: z
.enum(["API", "EVAL", "ANNOTATION"])
.default("API" as ScoreSourceType),
});
/**
@@ -272,8 +364,8 @@ export const ScoreBody = z.discriminatedUnion("dataType", [
// LEGACY, only required for backwards compatibility
export const LegacySpanPostSchema = z.object({
id: z.string().nullish(),
traceId: z.string().nullish(),
id: idSchema.nullish(),
traceId: idSchema.nullish(),
name: z.string().nullish(),
startTime: stringDateTime,
endTime: stringDateTime,
@@ -281,28 +373,28 @@ export const LegacySpanPostSchema = z.object({
input: jsonSchema.nullish(),
output: jsonSchema.nullish(),
parentObservationId: z.string().nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
version: z.string().nullish(),
});
export const LegacySpanPatchSchema = z.object({
spanId: z.string(),
traceId: z.string().nullish(),
spanId: idSchema,
traceId: idSchema.nullish(),
name: z.string().nullish(),
startTime: stringDateTime,
endTime: stringDateTime,
metadata: jsonSchema.nullish(),
input: jsonSchema.nullish(),
output: jsonSchema.nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
version: z.string().nullish(),
});
export const LegacyGenerationsCreateSchema = z.object({
id: z.string().nullish(),
traceId: z.string().nullish(),
id: idSchema.nullish(),
traceId: idSchema.nullish(),
name: z.string().nullish(),
startTime: stringDateTime,
endTime: stringDateTime,
@@ -319,14 +411,14 @@ export const LegacyGenerationsCreateSchema = z.object({
usage: usage,
metadata: jsonSchema.nullish(),
parentObservationId: z.string().nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
version: z.string().nullish(),
});
export const LegacyGenerationPatchSchema = z.object({
generationId: z.string(),
traceId: z.string().nullish(),
generationId: idSchema,
traceId: idSchema.nullish(),
name: z.string().nullish(),
startTime: stringDateTime,
endTime: stringDateTime,
@@ -342,14 +434,14 @@ export const LegacyGenerationPatchSchema = z.object({
completion: jsonSchema.nullish(),
usage: usage,
metadata: jsonSchema.nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
version: z.string().nullish(),
});
export const LegacyObservationBody = z.object({
id: z.string().nullish(),
traceId: z.string().nullish(),
id: idSchema.nullish(),
traceId: idSchema.nullish(),
type: z.enum(["GENERATION", "SPAN", "EVENT"]),
name: z.string().nullish(),
startTime: stringDateTime,
@@ -365,11 +457,11 @@ export const LegacyObservationBody = z.object({
input: jsonSchema.nullish(),
output: jsonSchema.nullish(),
usage: usage,
usageDetails: UsageOrCostDetails,
costDetails: UsageOrCostDetails,
usageDetails: UsageDetails,
costDetails: CostDetails,
metadata: jsonSchema.nullish(),
parentObservationId: z.string().nullish(),
level: z.nativeEnum(ObservationLevel).nullish(),
level: ObservationLevel.nullish(),
statusMessage: z.string().nullish(),
version: z.string().nullish(),
});
@@ -379,18 +471,6 @@ export const SdkLogEvent = z.object({
id: z.string().nullish(), // Not used, but makes downstream processing easier.
});
// definitions for the ingestion API
export const observationTypes = [
"observation-create",
"observation-update",
"generation-create",
"generation-update",
"span-create",
"span-update",
"event-create",
];
export const eventTypes = {
TRACE_CREATE: "trace-create",
SCORE_CREATE: "score-create",
@@ -406,7 +486,7 @@ export const eventTypes = {
} as const;
const base = z.object({
id: z.string(),
id: idSchema,
timestamp: z.string().datetime({ offset: true }),
metadata: jsonSchema.nullish(),
});
@@ -469,28 +549,6 @@ export const ingestionEvent = z.discriminatedUnion("type", [
]);
export type IngestionEventType = z.infer<typeof ingestionEvent>;
export const ingestionBatchEvent = z.array(ingestionEvent);
export type IngestionBatchEventType = z.infer<typeof ingestionBatchEvent>;
export const ingestionEventWithProjectId = ingestionEvent.and(
z.object({ projectId: z.string() }),
);
export type IngestionEventWithProjectIdType = z.infer<
typeof ingestionEventWithProjectId
>;
export const ingestionApiSchema = z.object({
batch: ingestionBatchEvent,
metadata: jsonSchema.nullish(),
});
export const ingestionApiSchemaWithProjectId = ingestionApiSchema.extend({
projectId: z.string(),
});
export type IngestionApiSchemaWithProjectId = z.infer<
typeof ingestionApiSchemaWithProjectId
>;
export type ObservationEvent =
| z.infer<typeof legacyObservationCreateEvent>
| z.infer<typeof legacyObservationUpdateEvent>
@@ -4,7 +4,8 @@ import {
validateDbScoreConfigSafe,
ValidatedScoreConfig,
} from "../../../src";
import { prisma, Score, ScoreDataType } from "../../db";
import { prisma, ScoreDataType } from "../../db";
import { Score } from "../repositories";
import { InvalidRequestError, LangfuseNotFoundError } from "../../errors";
@@ -160,35 +160,49 @@ export const addUserToSpan = (
export const getTracer = (name: string) => opentelemetry.trace.getTracer(name);
const cloudWatchClient = new CloudWatchClient();
const cloudWatchLastSubmitted: Record<string, number> = {};
const sendCloudWatchMetric = (key: string, value: number | undefined) => {
const currentTime = Date.now();
const interval = 30 * 1000;
let lastFlushTime = 0;
let metricCache: Record<string, number> = {};
// Check if the function has been executed in the last 30s for this key
if (
!cloudWatchLastSubmitted[key] ||
currentTime - cloudWatchLastSubmitted[key] >= interval
) {
cloudWatchLastSubmitted[key] = currentTime;
cloudWatchClient
.send(
new PutMetricDataCommand({
Namespace: "Langfuse",
MetricData: [
{
MetricName: key,
Value: value ?? 0,
},
],
}),
)
.catch((error) => {
logger.warn("Failed to send metric to CloudWatch", error);
});
// Caches metrics and flushes them on schedule
const sendCloudWatchMetric = (key: string, value: number, replace: boolean) => {
// Store the latest value for each metric key. If replace is false (e.g. for increments) we add the value to the existing value.
metricCache[key] = replace ? value : (metricCache[key] || 0) + value;
const currentTime = Date.now();
const flushInterval = 30 * 1000; // 30 seconds
// Check if it's time to flush the metrics
if (currentTime - lastFlushTime >= flushInterval) {
flushMetricsToCloudWatch();
}
};
// Flush all cached metrics in a single API call
const flushMetricsToCloudWatch = () => {
if (Object.keys(metricCache).length === 0) return;
lastFlushTime = Date.now();
const metricData = Object.entries(metricCache).map(([key, value]) => ({
MetricName: key,
Value: value,
}));
// Clear the cache after preparing the metrics
metricCache = {};
cloudWatchClient
.send(
new PutMetricDataCommand({
Namespace: "Langfuse",
MetricData: metricData,
}),
)
.catch((error) => {
logger.warn("Failed to send metrics to CloudWatch", error);
});
};
export const recordGauge = (
stat: string,
value?: number | undefined,
@@ -199,7 +213,7 @@ export const recordGauge = (
| undefined,
) => {
if (env.ENABLE_AWS_CLOUDWATCH_METRIC_PUBLISHING === "true") {
sendCloudWatchMetric(stat, value);
sendCloudWatchMetric(stat, value ?? 0, true);
}
dd.dogstatsd.gauge(stat, value, tags);
};
@@ -210,7 +224,7 @@ export const recordIncrement = (
tags?: { [tag: string]: string | number } | undefined,
) => {
if (env.ENABLE_AWS_CLOUDWATCH_METRIC_PUBLISHING === "true") {
sendCloudWatchMetric(stat, value);
sendCloudWatchMetric(stat, value ?? 1, false);
}
dd.dogstatsd.increment(stat, value, tags);
};
@@ -220,9 +234,6 @@ export const recordHistogram = (
value?: number | undefined,
tags?: { [tag: string]: string | number } | undefined,
) => {
if (env.ENABLE_AWS_CLOUDWATCH_METRIC_PUBLISHING === "true") {
sendCloudWatchMetric(stat, value);
}
dd.dogstatsd.histogram(stat, value, tags);
};
@@ -3,6 +3,7 @@ import type { ZodSchema } from "zod";
import { ChatAnthropic } from "@langchain/anthropic";
import { ChatVertexAI } from "@langchain/google-vertexai";
import { ChatBedrockConverse } from "@langchain/aws";
import { ChatGoogleGenerativeAI } from "@langchain/google-genai";
import {
AIMessage,
BaseMessage,
@@ -133,7 +134,10 @@ export async function fetchLLMCompletion(
finalMessages = messages.map((message) => {
if (message.role === ChatMessageRole.User)
return new HumanMessage(message.content);
if (message.role === ChatMessageRole.System)
if (
message.role === ChatMessageRole.System ||
message.role === ChatMessageRole.Developer
)
return new SystemMessage(message.content);
return new AIMessage(message.content);
@@ -146,7 +150,8 @@ export async function fetchLLMCompletion(
| ChatOpenAI
| ChatAnthropic
| ChatBedrockConverse
| ChatVertexAI;
| ChatVertexAI
| ChatGoogleGenerativeAI;
if (modelParams.adapter === LLMAdapter.Anthropic) {
chatModel = new ChatAnthropic({
anthropicApiKey: apiKey,
@@ -179,13 +184,16 @@ export async function fetchLLMCompletion(
azureOpenAIApiKey: apiKey,
azureOpenAIBasePath: baseURL,
azureOpenAIApiDeploymentName: modelParams.model,
azureOpenAIApiVersion: "2024-02-01",
azureOpenAIApiVersion: "2025-02-01-preview",
temperature: modelParams.temperature,
maxTokens: modelParams.max_tokens,
topP: modelParams.top_p,
callbacks: finalCallbacks,
maxRetries,
timeout: 1000 * 60 * 2, // 2 minutes timeout
configuration: {
defaultHeaders: extraHeaders,
},
});
} else if (modelParams.adapter === LLMAdapter.Bedrock) {
const { region } = BedrockConfigSchema.parse(config);
@@ -219,6 +227,32 @@ export async function fetchLLMCompletion(
credentials,
},
});
} else if (modelParams.adapter === LLMAdapter.GoogleAIStudio) {
chatModel = new ChatGoogleGenerativeAI({
model: modelParams.model,
temperature: modelParams.temperature,
maxOutputTokens: modelParams.max_tokens,
topP: modelParams.top_p,
callbacks: finalCallbacks,
maxRetries,
apiKey,
});
} else if (modelParams.adapter === LLMAdapter.Atla) {
// Atla models do not support:
// - temperature
// - max_tokens
// - top_p
chatModel = new ChatOpenAI({
openAIApiKey: apiKey,
modelName: modelParams.model,
callbacks: finalCallbacks,
maxRetries,
configuration: {
baseURL: baseURL,
defaultHeaders: extraHeaders,
},
timeout: 1000 * 60, // 1 minute timeout
});
} else {
// eslint-disable-next-line no-unused-vars
const _exhaustiveCheck: never = modelParams.adapter;
@@ -244,8 +278,8 @@ export async function fetchLLMCompletion(
/*
Workaround OpenAI reasoning models:
This is a temporary workaround to avoid sending system messages to OpenAI's O1 models.
O1 models do not support in beta:
This is a temporary workaround to avoid sending unsupported parameters to OpenAI's O1 models.
O1 models do not support:
- system messages
- top_p
- max_tokens at all, one has to use max_completion_tokens instead
@@ -257,6 +291,12 @@ export async function fetchLLMCompletion(
modelParams.model.startsWith("o1-") ||
modelParams.model.startsWith("o3-")
) {
const filteredMessages = finalMessages.filter((message) => {
return (
modelParams.model.startsWith("o3-") || message._getType() !== "system"
);
});
return {
completion: await new ChatOpenAI({
openAIApiKey: apiKey,
@@ -275,10 +315,7 @@ export async function fetchLLMCompletion(
timeout: 1000 * 60 * 2, // 2 minutes timeout
})
.pipe(new StringOutputParser())
.invoke(
finalMessages.filter((message) => message._getType() !== "system"),
runConfig,
),
.invoke(filteredMessages, runConfig),
processTracedEvents,
};
}
+27
View File
@@ -4,6 +4,9 @@ import { BedrockConfigSchema } from "../../interfaces/customLLMProviderConfigSch
import { TokenCountDelegate } from "../ingestion/processEventBatch";
import { AuthHeaderValidVerificationResult } from "../auth/types";
/* eslint-disable no-unused-vars */
// disable lint as this is exported and used in web/worker
export type PromptVariable = { name: string; value: string; isUsed: boolean };
export type ChatMessage = {
@@ -16,17 +19,25 @@ export type ChatMessageWithId = ChatMessage & { id: string };
export enum LLMAdapter {
Anthropic = "anthropic",
OpenAI = "openai",
Atla = "atla",
Azure = "azure",
Bedrock = "bedrock",
VertexAI = "google-vertex-ai",
GoogleAIStudio = "google-ai-studio",
}
export enum ChatMessageRole {
System = "system",
User = "user",
Assistant = "assistant",
Developer = "developer",
}
export const SYSTEM_ROLES: string[] = [
ChatMessageRole.System,
ChatMessageRole.Developer,
];
export const ChatMessageDefaultRoleSchema = z.nativeEnum(ChatMessageRole);
const ChatMessageSchema = z.object({
@@ -91,6 +102,8 @@ export const openAIModels = [
"o1-preview-2024-09-12",
"o1-mini",
"o1-mini-2024-09-12",
"gpt-4.5-preview",
"gpt-4.5-preview-2025-02-27",
"gpt-4-turbo-preview",
"gpt-4-1106-preview",
"gpt-4-0613",
@@ -109,6 +122,7 @@ export type OpenAIModel = (typeof openAIModels)[number];
// NOTE: Update docs page when changing this! https://langfuse.com/docs/playground#openai-playground--anthropic-playground
export const anthropicModels = [
"claude-3-7-sonnet-20250219",
"claude-3-5-sonnet-20241022",
"claude-3-5-sonnet-20240620",
"claude-3-opus-20240229",
@@ -130,14 +144,27 @@ export const vertexAIModels = [
"gemini-1.0-pro",
] as const;
export const googleAIStudioModels = [
"gemini-2.0-flash",
"gemini-2.0-flash-lite-preview-02-05",
"gemini-2.0-flash-thinking-exp-01-21",
"gemini-1.5-pro",
"gemini-1.5-flash",
"gemini-1.5-flash-8b",
] as const;
export const atlaModels = ["atla-selene", "atla-selene-20250214"] as const;
export type AnthropicModel = (typeof anthropicModels)[number];
export type VertexAIModel = (typeof vertexAIModels)[number];
export const supportedModels = {
[LLMAdapter.Anthropic]: anthropicModels,
[LLMAdapter.OpenAI]: openAIModels,
[LLMAdapter.VertexAI]: vertexAIModels,
[LLMAdapter.GoogleAIStudio]: googleAIStudioModels,
[LLMAdapter.Azure]: [],
[LLMAdapter.Bedrock]: [],
[LLMAdapter.Atla]: atlaModels,
} as const;
export type LLMFunctionCall = {
@@ -383,6 +383,10 @@ export class FilterList {
return this.filters.find(predicate);
}
filter(predicate: (filter: Filter) => boolean) {
return new FilterList(this.filters.filter(predicate));
}
some(predicate: (filter: Filter) => boolean) {
return this.filters.some(predicate);
}
@@ -3,7 +3,7 @@ import { singleFilter } from "../../../interfaces/filters";
import { FilterCondition } from "../../../types";
import { isValidTableName } from "../../clickhouse/schemaUtils";
import { logger } from "../../logger";
import { UiColumnMapping } from "../../../tableDefinitions";
import { UiColumnMappings } from "../../../tableDefinitions";
import {
StringFilter,
DateTimeFilter,
@@ -29,7 +29,7 @@ export class QueryBuilderError extends Error {
// User input for values (e.g. project_id = <value>) are sent to Clickhouse as parameters to prevent SQL injection
export const createFilterFromFilterState = (
filter: FilterCondition[],
columnMapping: UiColumnMapping[],
columnMapping: UiColumnMappings,
) => {
return filter.map((frontEndFilter) => {
// checks if the column exists in the clickhouse schema
@@ -120,7 +120,7 @@ export const createFilterFromFilterState = (
const matchAndVerifyTracesUiColumn = (
filter: z.infer<typeof singleFilter>,
uiTableDefinitions: UiColumnMapping[],
uiTableDefinitions: UiColumnMappings,
) => {
// tries to match the column name to the clickhouse table name
logger.debug(`Filter to match: ${JSON.stringify(filter)}`);
@@ -1,13 +1,13 @@
import z from "zod";
import { OrderByState } from "../../../interfaces/orderBy";
import { UiColumnMapping } from "../../../tableDefinitions";
import { UiColumnMappings } from "../../../tableDefinitions";
import { logger } from "../../logger";
type OrderByStateNotNull = Exclude<OrderByState, null>;
export function orderByToClickhouseSql(
orderBy: OrderByState | OrderByState[] = [],
tableColumns: UiColumnMapping[],
tableColumns: UiColumnMappings,
): string {
if (
!orderBy ||
@@ -1,4 +1,4 @@
import { ObservationView } from "@prisma/client";
import { ObservationView } from "../repositories";
type AdditionalObservationFields = {
traceName: string | null;
@@ -18,3 +18,4 @@ export {
type ClickhouseOperator,
} from "./clickhouse-sql/clickhouse-filter";
export { orderByToClickhouseSql } from "./clickhouse-sql/orderby-factory";
export { createFilterFromFilterState } from "./clickhouse-sql/factory";
+95 -9
View File
@@ -1,3 +1,4 @@
/* eslint-disable no-unused-vars */
import { z } from "zod";
import { eventTypes } from ".";
import {
@@ -33,6 +34,10 @@ export const TracesQueueEventSchema = z.object({
projectId: z.string(),
traceIds: z.array(z.string()),
});
export const ScoresQueueEventSchema = z.object({
projectId: z.string(),
scoreIds: z.array(z.string()),
});
export const ProjectQueueEventSchema = z.object({
projectId: z.string(),
orgId: z.string(),
@@ -51,6 +56,9 @@ export const EvalExecutionEvent = z.object({
export const PostHogIntegrationProcessingEventSchema = z.object({
projectId: z.string(),
});
export const BlobStorageIntegrationProcessingEventSchema = z.object({
projectId: z.string(),
});
export const ExperimentCreateEventSchema = z.object({
projectId: z.string(),
datasetId: z.string(),
@@ -61,19 +69,68 @@ export const DataRetentionProcessingEventSchema = z.object({
projectId: z.string(),
retention: z.number(),
});
export const BatchActionProcessingEventSchema = z.object({
projectId: z.string(),
actionId: z.string(),
query: BatchActionQuerySchema,
tableName: z.nativeEnum(BatchExportTableName),
cutoffCreatedAt: z.date(),
targetId: z.string().optional(),
type: z.nativeEnum(BatchActionType),
});
export const BatchActionProcessingEventSchema = z.discriminatedUnion(
"actionId",
[
z.object({
actionId: z.literal("score-delete"),
projectId: z.string(),
query: BatchActionQuerySchema,
tableName: z.nativeEnum(BatchExportTableName),
cutoffCreatedAt: z.date(),
targetId: z.string().optional(),
type: z.nativeEnum(BatchActionType),
}),
z.object({
actionId: z.literal("trace-delete"),
projectId: z.string(),
query: BatchActionQuerySchema,
tableName: z.nativeEnum(BatchExportTableName),
cutoffCreatedAt: z.date(),
targetId: z.string().optional(),
type: z.nativeEnum(BatchActionType),
}),
z.object({
actionId: z.literal("trace-add-to-annotation-queue"),
projectId: z.string(),
query: BatchActionQuerySchema,
tableName: z.nativeEnum(BatchExportTableName),
cutoffCreatedAt: z.date(),
targetId: z.string().optional(),
type: z.nativeEnum(BatchActionType),
}),
z.object({
actionId: z.literal("eval-create"),
targetObject: z.enum(["trace", "dataset"]),
configId: z.string(),
projectId: z.string(),
cutoffCreatedAt: z.date(),
query: BatchActionQuerySchema,
}),
],
);
export const CreateEvalQueueEventSchema = DatasetRunItemUpsertEventSchema.and(
z.object({
configId: z.string(),
timestamp: z.date(),
}),
).or(
TraceQueueEventSchema.and(
z.object({
timestamp: z.date(),
configId: z.string(),
}),
),
);
export type CreateEvalQueueEventType = z.infer<
typeof CreateEvalQueueEventSchema
>;
export type BatchExportJobType = z.infer<typeof BatchExportJobSchema>;
export type TraceQueueEventType = z.infer<typeof TraceQueueEventSchema>;
export type TracesQueueEventType = z.infer<typeof TracesQueueEventSchema>;
export type ScoresQueueEventType = z.infer<typeof ScoresQueueEventSchema>;
export type ProjectQueueEventType = z.infer<typeof ProjectQueueEventSchema>;
export type DatasetRunItemUpsertEventType = z.infer<
typeof DatasetRunItemUpsertEventSchema
@@ -92,6 +149,9 @@ export type DataRetentionProcessingEventType = z.infer<
export type BatchActionProcessingEventType = z.infer<
typeof BatchActionProcessingEventSchema
>;
export type BlobStorageIntegrationProcessingEventType = z.infer<
typeof BlobStorageIntegrationProcessingEventSchema
>;
export enum QueueName {
TraceUpsert = "trace-upsert", // Ingestion pipeline adds events on each Trace upsert
@@ -106,11 +166,15 @@ export enum QueueName {
ExperimentCreate = "experiment-create-queue",
PostHogIntegrationQueue = "posthog-integration-queue",
PostHogIntegrationProcessingQueue = "posthog-integration-processing-queue",
BlobStorageIntegrationQueue = "blobstorage-integration-queue",
BlobStorageIntegrationProcessingQueue = "blobstorage-integration-processing-queue",
CoreDataS3ExportQueue = "core-data-s3-export-queue",
MeteringDataPostgresExportQueue = "metering-data-postgres-export-queue",
DataRetentionQueue = "data-retention-queue",
DataRetentionProcessingQueue = "data-retention-processing-queue",
BatchActionQueue = "batch-action-queue",
CreateEvalQueue = "create-eval-queue",
ScoreDelete = "score-delete",
}
export enum QueueJobs {
@@ -126,11 +190,15 @@ export enum QueueJobs {
ExperimentCreateJob = "experiment-create-job",
PostHogIntegrationJob = "posthog-integration-job",
PostHogIntegrationProcessingJob = "posthog-integration-processing-job",
BlobStorageIntegrationJob = "blobstorage-integration-job",
BlobStorageIntegrationProcessingJob = "blobstorage-integration-processing-job",
CoreDataS3ExportJob = "core-data-s3-export-job",
MeteringDataPostgresExportJob = "metering-data-postgres-export-job",
DataRetentionJob = "data-retention-job",
DataRetentionProcessingJob = "data-retention-processing-job",
BatchActionProcessingJob = "batch-action-processing-job",
CreateEvalJob = "create-eval-job",
ScoreDelete = "score-delete",
}
export type TQueueJobTypes = {
@@ -146,6 +214,12 @@ export type TQueueJobTypes = {
payload: TracesQueueEventType | TraceQueueEventType;
name: QueueJobs.TraceDelete;
};
[QueueName.ScoreDelete]: {
timestamp: Date;
id: string;
payload: ScoresQueueEventType;
name: QueueJobs.ScoreDelete;
};
[QueueName.ProjectDelete]: {
timestamp: Date;
id: string;
@@ -206,4 +280,16 @@ export type TQueueJobTypes = {
payload: BatchActionProcessingEventType;
name: QueueJobs.BatchActionProcessingJob;
};
[QueueName.CreateEvalQueue]: {
timestamp: Date;
id: string;
payload: CreateEvalQueueEventType;
name: QueueJobs.CreateEvalJob;
};
[QueueName.BlobStorageIntegrationProcessingQueue]: {
timestamp: Date;
id: string;
payload: BlobStorageIntegrationProcessingEventType;
name: QueueJobs.BlobStorageIntegrationProcessingJob;
};
};
@@ -0,0 +1,40 @@
import { Queue } from "bullmq";
import { QueueName } from "../queues";
import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
import { logger } from "../logger";
export class BlobStorageIntegrationProcessingQueue {
private static instance: Queue | null = null;
public static getInstance(): Queue | null {
if (BlobStorageIntegrationProcessingQueue.instance) {
return BlobStorageIntegrationProcessingQueue.instance;
}
const newRedis = createNewRedisInstance({
enableOfflineQueue: false,
...redisQueueRetryOptions,
});
BlobStorageIntegrationProcessingQueue.instance = newRedis
? new Queue(QueueName.BlobStorageIntegrationProcessingQueue, {
connection: newRedis,
defaultJobOptions: {
removeOnComplete: true,
removeOnFail: 100_000,
attempts: 5,
backoff: {
type: "exponential",
delay: 5000,
},
},
})
: null;
BlobStorageIntegrationProcessingQueue.instance?.on("error", (err) => {
logger.error("BlobStorageIntegrationProcessingQueue error", err);
});
return BlobStorageIntegrationProcessingQueue.instance;
}
}
@@ -0,0 +1,55 @@
import { Queue } from "bullmq";
import { QueueName, QueueJobs } from "../queues";
import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
import { logger } from "../logger";
export class BlobStorageIntegrationQueue {
private static instance: Queue | null = null;
public static getInstance(): Queue | null {
if (BlobStorageIntegrationQueue.instance) {
return BlobStorageIntegrationQueue.instance;
}
const newRedis = createNewRedisInstance({
enableOfflineQueue: false,
...redisQueueRetryOptions,
});
BlobStorageIntegrationQueue.instance = newRedis
? new Queue(QueueName.BlobStorageIntegrationQueue, {
connection: newRedis,
defaultJobOptions: {
removeOnComplete: true,
removeOnFail: 100,
attempts: 5,
backoff: {
type: "exponential",
delay: 5000,
},
},
})
: null;
BlobStorageIntegrationQueue.instance?.on("error", (err) => {
logger.error("BlobStorageIntegrationQueue error", err);
});
if (BlobStorageIntegrationQueue.instance) {
logger.debug("Scheduling jobs for BlobStorageIntegrationQueue");
BlobStorageIntegrationQueue.instance
.add(
QueueJobs.BlobStorageIntegrationJob,
{},
{
repeat: { pattern: "20 * * * *" }, // every hour at 20 minutes past
},
)
.catch((err) => {
logger.error("Error adding BlobStorageIntegrationJob schedule", err);
});
}
return BlobStorageIntegrationQueue.instance;
}
}
@@ -0,0 +1,45 @@
import { QueueName, TQueueJobTypes } from "../queues";
import { Queue } from "bullmq";
import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
import { logger } from "../logger";
export class CreateEvalQueue {
private static instance: Queue<
TQueueJobTypes[QueueName.CreateEvalQueue]
> | null = null;
public static getInstance(): Queue<
TQueueJobTypes[QueueName.CreateEvalQueue]
> | null {
if (CreateEvalQueue.instance) return CreateEvalQueue.instance;
const newRedis = createNewRedisInstance({
enableOfflineQueue: false,
...redisQueueRetryOptions,
});
CreateEvalQueue.instance = newRedis
? new Queue<TQueueJobTypes[QueueName.CreateEvalQueue]>(
QueueName.CreateEvalQueue,
{
connection: newRedis,
defaultJobOptions: {
removeOnComplete: 100, // Important: If not true, new jobs for that ID would be ignored as jobs in the complete set are still considered as part of the queue
removeOnFail: 100_000,
attempts: 5,
backoff: {
type: "exponential",
delay: 5000,
},
},
},
)
: null;
CreateEvalQueue.instance?.on("error", (err) => {
logger.error("CreateEvalQueue error", err);
});
return CreateEvalQueue.instance;
}
}
@@ -11,11 +11,15 @@ import { TraceDeleteQueue } from "./traceDelete";
import { ProjectDeleteQueue } from "./projectDelete";
import { PostHogIntegrationQueue } from "./postHogIntegrationQueue";
import { PostHogIntegrationProcessingQueue } from "./postHogIntegrationProcessingQueue";
import { BlobStorageIntegrationQueue } from "./blobStorageIntegrationQueue";
import { BlobStorageIntegrationProcessingQueue } from "./blobStorageIntegrationProcessingQueue";
import { CoreDataS3ExportQueue } from "./coreDataS3ExportQueue";
import { MeteringDataPostgresExportQueue } from "./meteringDataPostgresExportQueue";
import { DataRetentionQueue } from "./dataRetentionQueue";
import { DataRetentionProcessingQueue } from "./dataRetentionProcessingQueue";
import { BatchActionQueue } from "./batchActionQueue";
import { CreateEvalQueue } from "./createEvalQueue";
import { ScoreDeleteQueue } from "./scoreDelete";
export function getQueue(queueName: QueueName): Queue | null {
switch (queueName) {
@@ -41,6 +45,10 @@ export function getQueue(queueName: QueueName): Queue | null {
return PostHogIntegrationQueue.getInstance();
case QueueName.PostHogIntegrationProcessingQueue:
return PostHogIntegrationProcessingQueue.getInstance();
case QueueName.BlobStorageIntegrationQueue:
return BlobStorageIntegrationQueue.getInstance();
case QueueName.BlobStorageIntegrationProcessingQueue:
return BlobStorageIntegrationProcessingQueue.getInstance();
case QueueName.IngestionSecondaryQueue:
return SecondaryIngestionQueue.getInstance();
case QueueName.CoreDataS3ExportQueue:
@@ -53,6 +61,10 @@ export function getQueue(queueName: QueueName): Queue | null {
return DataRetentionProcessingQueue.getInstance();
case QueueName.BatchActionQueue:
return BatchActionQueue.getInstance();
case QueueName.CreateEvalQueue:
return CreateEvalQueue.getInstance();
case QueueName.ScoreDelete:
return ScoreDeleteQueue.getInstance();
default:
const exhaustiveCheckDefault: never = queueName;
throw new Error(`Queue ${queueName} not found`);
@@ -0,0 +1,38 @@
import { QueueName, TQueueJobTypes } from "../queues";
import { Queue } from "bullmq";
import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
import { logger } from "../logger";
export class ScoreDeleteQueue {
private static instance: Queue<TQueueJobTypes[QueueName.ScoreDelete]> | null = null;
public static getInstance(): Queue<TQueueJobTypes[QueueName.ScoreDelete]> | null {
if (ScoreDeleteQueue.instance) return ScoreDeleteQueue.instance;
const newRedis = createNewRedisInstance({
enableOfflineQueue: false,
...redisQueueRetryOptions,
});
ScoreDeleteQueue.instance = newRedis
? new Queue<TQueueJobTypes[QueueName.ScoreDelete]>(QueueName.ScoreDelete, {
connection: newRedis,
defaultJobOptions: {
removeOnComplete: true,
removeOnFail: 100_000,
attempts: 2,
backoff: {
type: "exponential",
delay: 30_000,
},
},
})
: null;
ScoreDeleteQueue.instance?.on("error", (err) => {
logger.error("ScoreDeleteQueue error", err);
});
return ScoreDeleteQueue.instance;
}
}
@@ -25,10 +25,10 @@ export class TraceDeleteQueue {
defaultJobOptions: {
removeOnComplete: true,
removeOnFail: 100_000,
attempts: 5,
attempts: 2,
backoff: {
type: "exponential",
delay: 5000,
delay: 30_000,
},
},
},
@@ -4,11 +4,8 @@ import {
} from "./clickhouse";
import { createFilterFromFilterState } from "../queries/clickhouse-sql/factory";
import { FilterState } from "../../types";
import {
DateTimeFilter,
FilterList,
} from "../queries/clickhouse-sql/clickhouse-filter";
import { dashboardColumnDefinitions } from "../../tableDefinitions/mapDashboards";
import { DateTimeFilter, FilterList } from "../queries";
import { dashboardColumnDefinitions } from "../../tableDefinitions";
import { convertDateToClickhouseDateTime } from "../clickhouse/client";
import {
OBSERVATIONS_TO_TRACE_INTERVAL,
@@ -18,12 +15,37 @@ import {
export type DateTrunc = "year" | "month" | "week" | "day" | "hour" | "minute";
const extractEnvironmentFilterFromFilters = (
filter: FilterState,
): { envFilter: FilterState; remainingFilters: FilterState } => {
return {
envFilter: filter.filter((f) => f.column === "environment"),
remainingFilters: filter.filter((f) => f.column !== "environment"),
};
};
const convertEnvFilterToClickhouseFilter = (filter: FilterState) => {
return createFilterFromFilterState(filter, [
{
clickhouseSelect: "environment",
clickhouseTableName: "traces",
uiTableId: "environment",
uiTableName: "Environment",
},
]);
};
export const getTotalTraces = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
).apply();
const query = `
@@ -31,13 +53,16 @@ export const getTotalTraces = async (
count(id) as count
FROM traces t FINAL
WHERE project_id = {projectId: String}
AND ${chFilter.query}`;
AND ${chFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
`;
const result = await queryClickhouse<{ count: number }>({
query,
params: {
projectId,
...chFilter.params,
...environmentFilter.params,
},
tags: {
feature: "dashboard",
@@ -58,8 +83,13 @@ export const getObservationsCostGroupedByName = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -74,6 +104,7 @@ export const getObservationsCostGroupedByName = async (
FROM observations o FINAL ${hasTraceFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
${appliedFilter.query ? `AND ${appliedFilter.query}` : ""}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
GROUP BY provided_model_name
ORDER BY sumMap(cost_details)['total'] DESC
LIMIT 50
@@ -88,6 +119,7 @@ export const getObservationsCostGroupedByName = async (
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
},
tags: {
feature: "dashboard",
@@ -104,8 +136,13 @@ export const getScoreAggregate = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const timeFilter = chFilter.find(
@@ -129,6 +166,7 @@ export const getScoreAggregate = async (
${hasTraceFilter ? "JOIN traces t FINAL ON t.id = s.trace_id AND t.project_id = s.project_id" : ""}
WHERE s.project_id = {projectId: String}
AND ${chFilterApplied.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter && hasTraceFilter ? `AND t.timestamp >= {tracesTimestamp: DateTime64(3)} - ${SCORE_TO_TRACE_OBSERVATIONS_INTERVAL}` : ""}
GROUP BY s.name, s.source, s.data_type
ORDER BY count(*) DESC
@@ -145,6 +183,7 @@ export const getScoreAggregate = async (
params: {
projectId,
...chFilterApplied.params,
...environmentFilter.params,
...(timeFilter
? { tracesTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
: {}),
@@ -164,8 +203,13 @@ export const groupTracesByTime = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
).apply();
const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
@@ -180,6 +224,7 @@ export const groupTracesByTime = async (
FROM traces t FINAL
WHERE project_id = {projectId: String}
AND ${chFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
GROUP BY timestamp
${orderByQuery}
`;
@@ -191,6 +236,7 @@ export const groupTracesByTime = async (
params: {
projectId,
...chFilter.params,
...environmentFilter.params,
...orderByParams,
},
tags: {
@@ -207,12 +253,17 @@ export const groupTracesByTime = async (
}));
};
export const getObservationUsageByTime = async (
export const getTotalObservationUsageByTimeByModel = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -235,13 +286,14 @@ export const getObservationUsageByTime = async (
const query = `
SELECT
${selectTimeseriesColumn(bucketSizeInSeconds, "start_time", "start_time")},
sumMap(usage_details) as units,
sumMap(cost_details) as cost,
sumMap(usage_details)['total'] as units,
sumMap(cost_details)['total'] as cost,
provided_model_name
FROM observations o FINAL
${tracesFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
GROUP BY start_time, provided_model_name
${orderByQuery}
@@ -249,14 +301,15 @@ export const getObservationUsageByTime = async (
const result = await queryClickhouse<{
start_time: string;
units: Record<string, number>;
cost: Record<string, number>;
units: string;
cost: string;
provided_model_name: string;
}>({
query,
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...orderByParams,
...(timeFilter
? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
@@ -271,29 +324,220 @@ export const getObservationUsageByTime = async (
});
return result.map((row) => ({
start_time: parseClickhouseUTCDateTimeFormat(row.start_time),
units: Object.fromEntries(
Object.entries(row.units ?? {}).map(([key, value]) => [
key,
Number(value),
]),
),
cost: Object.fromEntries(
Object.entries(row.cost ?? {}).map(([key, value]) => [
key,
Number(value),
]),
),
provided_model_name: row.provided_model_name,
startTime: parseClickhouseUTCDateTimeFormat(row.start_time),
units: Number(row.units),
cost: Number(row.cost),
model: row.provided_model_name,
}));
};
export const getObservationCostByTypeByTime = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
const tracesFilter = chFilter.find((f) => f.clickhouseTable === "traces");
const timeFilter = tracesFilter
? (chFilter.find(
(f) =>
f.clickhouseTable === "observations" &&
f.field.includes("start_time") &&
(f.operator === ">=" || f.operator === ">"),
) as DateTimeFilter | undefined)
: undefined;
const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
filter,
"start_time",
);
const query = `
SELECT
start_time,
groupArray((cost_key, cost_sum)) AS costs
FROM (
SELECT
${selectTimeseriesColumn(bucketSizeInSeconds, "start_time", "start_time")},
cost_key,
SUM(cost) AS cost_sum
FROM
observations o FINAL
${tracesFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
ARRAY JOIN
mapKeys(cost_details) AS cost_key,
mapValues(cost_details) AS cost
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
GROUP BY
start_time,
cost_key
)
GROUP BY
start_time
${orderByQuery}
`;
const result = await queryClickhouse<{
start_time: string;
costs: Array<[string, number | null]>;
}>({
query,
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...orderByParams,
...(timeFilter
? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
: {}),
},
tags: {
feature: "dashboard",
type: "observationCostByTypeByTime",
kind: "analytic",
projectId,
},
});
const types = result.flatMap((row) => {
return row.costs.map((cost) => cost[0]);
});
const uniqueTypes = [...new Set(types)];
return result.flatMap((row) => {
const intervalStart = parseClickhouseUTCDateTimeFormat(row.start_time);
return uniqueTypes.map((type) => ({
intervalStart: intervalStart,
key: type,
sum: row.costs.find((cost) => cost[0] === type)?.[1]
? Number(row.costs.find((cost) => cost[0] === type)?.[1])
: 0,
}));
});
};
export const getObservationUsageByTypeByTime = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
const tracesFilter = chFilter.find((f) => f.clickhouseTable === "traces");
const timeFilter = tracesFilter
? (chFilter.find(
(f) =>
f.clickhouseTable === "observations" &&
f.field.includes("start_time") &&
(f.operator === ">=" || f.operator === ">"),
) as DateTimeFilter | undefined)
: undefined;
const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
filter,
"start_time",
);
const query = `
SELECT
start_time,
groupArray((usage_key, usage_sum)) AS usages
FROM (
SELECT
${selectTimeseriesColumn(bucketSizeInSeconds, "start_time", "start_time")} ,
usage_key,
SUM(usage) AS usage_sum
FROM
observations o FINAL
${tracesFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
ARRAY JOIN
mapKeys(usage_details) AS usage_key,
mapValues(usage_details) AS usage
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
GROUP BY
start_time,
usage_key
)
GROUP BY
start_time
${orderByQuery}
`;
const result = await queryClickhouse<{
start_time: string;
usages: Array<[string, number | null]>;
}>({
query,
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...orderByParams,
...(timeFilter
? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
: {}),
},
tags: {
feature: "dashboard",
type: "observationUsageByTime",
kind: "analytic",
projectId,
},
});
const types = result.flatMap((row) => {
return row.usages.map((usage) => usage[0]);
});
const uniqueTypes = [...new Set(types)];
return result.flatMap((row) => {
const intervalStart = parseClickhouseUTCDateTimeFormat(row.start_time);
return uniqueTypes.map((type) => ({
intervalStart: intervalStart,
key: type,
sum: row.usages.find((usage) => usage[0] === type)?.[1]
? Number(row.usages.find((usage) => usage[0] === type)?.[1])
: 0,
}));
});
};
export const getDistinctModels = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -315,6 +559,7 @@ export const getDistinctModels = async (
${tracesFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
GROUP BY provided_model_name
ORDER BY count(*) DESC
@@ -326,6 +571,7 @@ export const getDistinctModels = async (
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...(timeFilter
? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
: {}),
@@ -345,8 +591,13 @@ export const getScoresAggregateOverTime = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -369,6 +620,7 @@ export const getScoresAggregateOverTime = async (
FROM scores FINAL
${traceFilter ? "JOIN traces t ON scores.trace_id = t.id AND scores.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
AND ${appliedFilter.query}
AND data_type IN ('NUMERIC', 'BOOLEAN')
GROUP BY
@@ -390,6 +642,7 @@ export const getScoresAggregateOverTime = async (
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...orderByParams,
},
tags: {
@@ -413,8 +666,13 @@ export const getModelUsageByUser = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -437,6 +695,7 @@ export const getModelUsageByUser = async (
WHERE project_id = {projectId: String}
AND t.user_id IS NOT NULL
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
GROUP BY user_id
ORDER BY sum_cost_details DESC
@@ -451,6 +710,7 @@ export const getModelUsageByUser = async (
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...(timeFilter
? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
: {}),
@@ -474,8 +734,13 @@ export const getObservationLatencies = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -489,13 +754,18 @@ export const getObservationLatencies = async (
${chFilter.find((f) => f.clickhouseTable === "traces") ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
GROUP BY name
ORDER BY quantiles[2] DESC
`;
const result = await queryClickhouse<{ quantiles: string[]; name: string }>({
query,
params: { projectId, ...appliedFilter.params },
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
},
tags: {
feature: "dashboard",
type: "observationLatencies",
@@ -517,8 +787,13 @@ export const getTracesLatencies = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -542,6 +817,7 @@ export const getTracesLatencies = async (
ON o.trace_id = t.id AND o.project_id = t.project_id
WHERE project_id = {projectId: String}
AND ${appliedFilter.query}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${timestampFilter ? `AND o.start_time > {dateTimeFilterObservations: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}` : ""}
GROUP BY o.project_id, o.trace_id, t.name
)
@@ -559,6 +835,7 @@ export const getTracesLatencies = async (
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...(timestampFilter
? { dateTimeFilterObservations: timestampFilter.value }
: {}),
@@ -584,8 +861,13 @@ export const getModelLatenciesOverTime = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const appliedFilter = chFilter.apply();
@@ -606,6 +888,7 @@ export const getModelLatenciesOverTime = async (
FROM observations o
${traceFilter ? "JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
AND ${appliedFilter.query}
GROUP BY provided_model_name, start_time_bucket
${orderByQuery};
@@ -617,7 +900,12 @@ export const getModelLatenciesOverTime = async (
quantiles: string[];
}>({
query,
params: { projectId, ...appliedFilter.params, ...orderByParams },
params: {
projectId,
...appliedFilter.params,
...environmentFilter.params,
...orderByParams,
},
tags: {
feature: "dashboard",
type: "modelLatenciesOverTime",
@@ -641,8 +929,13 @@ export const getNumericScoreTimeSeries = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const chFilterRes = chFilter.apply();
@@ -661,6 +954,7 @@ export const getNumericScoreTimeSeries = async (
FROM scores s final
${traceFilter ? "JOIN traces t ON s.trace_id = t.id AND s.project_id = t.project_id" : ""}
WHERE s.project_id = {projectId: String}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${chFilterRes?.query ? `AND ${chFilterRes.query}` : ""}
GROUP BY score_name, score_timestamp
${orderByQuery}
@@ -675,6 +969,7 @@ export const getNumericScoreTimeSeries = async (
params: {
projectId,
...(chFilterRes ? chFilterRes.params : {}),
...environmentFilter.params,
...orderByParams,
},
tags: {
@@ -696,8 +991,13 @@ export const getCategoricalScoreTimeSeries = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const chFilterRes = chFilter.apply();
@@ -719,6 +1019,7 @@ export const getCategoricalScoreTimeSeries = async (
FROM scores s final
${traceFilter ? "JOIN traces t ON s.trace_id = t.id AND s.project_id = t.project_id" : ""}
WHERE s.project_id = {projectId: String}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
${chFilterRes?.query ? `AND ${chFilterRes.query}` : ""}
GROUP BY score_name, score_data_type, score_source, score_value ${bucketSizeInSeconds ? ", score_timestamp" : ""}
${orderByQuery}
@@ -736,6 +1037,7 @@ export const getCategoricalScoreTimeSeries = async (
params: {
projectId,
...(chFilterRes ? chFilterRes.params : {}),
...environmentFilter.params,
...orderByParams,
},
tags: {
@@ -762,8 +1064,13 @@ export const getObservationsStatusTimeSeries = async (
projectId: string,
filter: FilterState,
) => {
const { envFilter, remainingFilters } =
extractEnvironmentFilterFromFilters(filter);
const environmentFilter = new FilterList(
convertEnvFilterToClickhouseFilter(envFilter),
).apply();
const chFilter = new FilterList(
createFilterFromFilterState(filter, dashboardColumnDefinitions),
createFilterFromFilterState(remainingFilters, dashboardColumnDefinitions),
);
const chFilterRes = chFilter.apply();
@@ -782,6 +1089,7 @@ export const getObservationsStatusTimeSeries = async (
FROM observations o
${traceFilter ? "JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
WHERE project_id = {projectId: String}
${environmentFilter.query ? `AND ${environmentFilter.query}` : ""}
AND o.level IS NOT NULL
AND ${chFilterRes?.query}
GROUP BY level ${bucketSizeInSeconds ? ", start_time_bucket" : ""}
@@ -797,6 +1105,7 @@ export const getObservationsStatusTimeSeries = async (
params: {
projectId,
...(chFilterRes ? chFilterRes.params : {}),
...environmentFilter.params,
...orderByParams,
},
tags: {
@@ -36,6 +36,7 @@ export const observationRecordBaseSchema = z.object({
project_id: z.string(),
type: z.string(),
parent_observation_id: z.string().nullish(),
environment: z.string().default("default"),
name: z.string().nullish(),
metadata: z.record(z.string()),
level: z.string().nullish(),
@@ -52,9 +53,6 @@ export const observationRecordBaseSchema = z.object({
prompt_version: z.number().nullish(),
is_deleted: z.number(),
});
export type ObservationRecordBaseType = z.infer<
typeof observationRecordBaseSchema
>;
export const observationRecordReadSchema = observationRecordBaseSchema.extend({
created_at: clickhouseStringDateSchema,
@@ -98,6 +96,7 @@ export const traceRecordBaseSchema = z.object({
release: z.string().nullish(),
version: z.string().nullish(),
project_id: z.string(),
environment: z.string().default("default"),
public: z.boolean(),
bookmarked: z.boolean(),
tags: z.array(z.string()),
@@ -106,7 +105,6 @@ export const traceRecordBaseSchema = z.object({
session_id: z.string().nullish(),
is_deleted: z.number(),
});
export type TraceRecordBaseType = z.infer<typeof traceRecordBaseSchema>;
export const traceRecordReadSchema = traceRecordBaseSchema.extend({
timestamp: clickhouseStringDateSchema,
@@ -129,6 +127,7 @@ export const scoreRecordBaseSchema = z.object({
project_id: z.string(),
trace_id: z.string(),
observation_id: z.string().nullish(),
environment: z.string().default("default"),
name: z.string(),
value: z.number().nullish(),
source: z.string(),
@@ -140,7 +139,6 @@ export const scoreRecordBaseSchema = z.object({
queue_id: z.string().nullish(),
is_deleted: z.number(),
});
export type ScoreRecordBaseType = z.infer<typeof scoreRecordBaseSchema>;
export const scoreRecordReadSchema = scoreRecordBaseSchema.extend({
created_at: clickhouseStringDateSchema,
@@ -246,6 +244,7 @@ export const convertPostgresTraceToInsert = (
: Array.isArray(trace.metadata)
? { metadata: trace.metadata }
: trace.metadata,
environment: trace.environment,
release: trace.release,
version: trace.version,
project_id: trace.project_id,
@@ -282,6 +281,7 @@ export const convertPostgresObservationToInsert = (
project_id: observation.project_id,
type: observation.type,
parent_observation_id: observation.parent_observation_id,
environment: observation.environment,
start_time: observation.start_time?.getTime(),
end_time: observation.end_time?.getTime(),
name: observation.name,
@@ -345,6 +345,7 @@ export const convertPostgresScoreToInsert = (
project_id: score.project_id,
trace_id: score.trace_id,
observation_id: score.observation_id,
environment: score.environment,
name: score.name,
value: score.value,
source: score.source,
@@ -0,0 +1,36 @@
import { queryClickhouse } from "./clickhouse";
export type EnvironmentFilterProps = {
projectId: string;
};
export const getEnvironmentsForProject = async (
props: EnvironmentFilterProps,
): Promise<{ environment: string }[]> => {
const { projectId } = props;
const query = `
SELECT environments
FROM project_environments FINAL
WHERE project_id = {projectId: String}
`;
const results = await queryClickhouse<{
environments: string[];
}>({
query,
params: { projectId },
tags: {
feature: "tracing",
type: "environment",
kind: "byId",
projectId,
},
});
const environments = results.length > 0 ? results[0].environments : [];
environments.push("default");
return Array.from(new Set(environments)).map((environment) => ({
environment,
}));
};
@@ -81,6 +81,135 @@ export const getEventLogByProjectIdBeforeDate = (
});
};
export const getEventLogByProjectIdAndEntityIds = (
projectId: string,
entityType: "observation" | "trace" | "score",
entityIds: string[],
): AsyncGenerator<EventLogRecordReadType> => {
const query = `
select *
from event_log
where project_id = {projectId: String}
and entity_type = {entityType: String}
and entity_id in ({entityIds: Array(String)})
`;
return queryClickhouseStream<EventLogRecordReadType>({
query,
params: {
projectId,
entityType,
entityIds,
},
clickhouseConfigs: {
request_timeout: 120_000, // 2 minutes
},
tags: {
feature: "eventLog",
kind: "list",
projectId,
},
});
};
export const getEventLogByProjectIdAndTraceIds = (
projectId: string,
traceIds: string[],
): AsyncGenerator<EventLogRecordReadType> => {
const query = `
with filtered_traces as (
select distinct
id as entity_id,
project_id as project_id,
'trace' as entity_type
from traces
where project_id = {projectId: String}
and id in ({traceIds: Array(String)})
), filtered_observations as (
select distinct
id as entity_id,
project_id as project_id,
'observation' as entity_type
from observations
where project_id = {projectId: String}
and trace_id in ({traceIds: Array(String)})
), filtered_scores as (
select distinct
id as entity_id,
project_id as project_id,
'score' as entity_type
from scores
where project_id = {projectId: String}
and trace_id in ({traceIds: Array(String)})
), filtered_events as (
select *
from filtered_traces
union all
select *
from filtered_observations
union all
select *
from filtered_scores
)
-- We use a semi join because we only use the 'filtered_events' as a filter.
-- There is no need to build the cartesian product (i.e. the combination) between the event log and the events.
select el.*
from event_log el
left semi join filtered_events fe
on el.project_id = fe.project_id and el.entity_id = fe.entity_id and el.entity_type = fe.entity_type
where el.project_id = {projectId: String}
`;
return queryClickhouseStream<EventLogRecordReadType>({
query,
params: {
projectId,
traceIds,
},
clickhouseConfigs: {
request_timeout: 120_000, // 2 minutes
},
tags: {
feature: "eventLog",
kind: "list",
projectId,
},
});
};
/**
* Deletes event log records by projectId and the _eventLog_.id
* @param projectId - Project ID
* @param ids - ID record of the event log table to be deleted
*/
export const deleteEventLogByProjectIdAndIds = async (
projectId: string,
ids: string[],
): Promise<void> => {
const query = `
delete from event_log
where project_id = {projectId: String}
and id in ({ids: Array(String)});
`;
await commandClickhouse({
query,
params: {
projectId,
ids,
},
clickhouseConfigs: {
request_timeout: 300_000, // 5 minutes
},
tags: {
feature: "eventLog",
kind: "delete",
projectId,
},
});
};
export const deleteEventLogByProjectId = async (
projectId: string,
): Promise<void> => {
@@ -10,3 +10,4 @@ export * from "./clickhouse";
export * from "./constants";
export * from "./trace-sessions";
export * from "./eventLog";
export * from "./environments";
@@ -5,7 +5,6 @@ import {
queryClickhouseStream,
upsertClickhouse,
} from "./clickhouse";
import { ObservationLevel } from "@prisma/client";
import { logger } from "../logger";
import { InternalServerError, LangfuseNotFoundError } from "../../errors";
import { prisma } from "../../db";
@@ -115,6 +114,7 @@ export const getObservationsViewForTrace = async (
project_id,
type,
parent_observation_id,
environment,
start_time,
end_time,
name,
@@ -178,6 +178,7 @@ export const getObservationForTraceIdByName = async (
project_id,
type,
parent_observation_id,
environment,
start_time,
end_time,
name,
@@ -343,6 +344,7 @@ const getObservationByIdInternal = async (
id,
trace_id,
project_id,
environment,
type,
parent_observation_id,
start_time,
@@ -508,6 +510,7 @@ const getObservationsTableInternal = async <T>(
o.provided_cost_details as "provided_cost_details",
o.cost_details as "cost_details",
o.level as level,
o.environment as "environment",
o.status_message as "status_message",
o.version as version,
o.parent_observation_id as "parent_observation_id",
@@ -1137,6 +1140,7 @@ export const getObservationMetricsForPrompts = async (
export const getLatencyAndTotalCostForObservations = async (
projectId: string,
observationIds: string[],
timestamp?: Date,
) => {
const query = `
SELECT
@@ -1145,7 +1149,8 @@ export const getLatencyAndTotalCostForObservations = async (
dateDiff('millisecond', start_time, end_time) AS latency_ms
FROM observations FINAL
WHERE project_id = {projectId: String}
AND id IN ({observationIds: Array(String)})
AND id IN ({observationIds: Array(String)})
${timestamp ? `AND start_time >= {timestamp: DateTime64(3)}` : ""}
`;
const rows = await queryClickhouse<{
id: string;
@@ -1156,6 +1161,9 @@ export const getLatencyAndTotalCostForObservations = async (
params: {
projectId,
observationIds,
...(timestamp
? { timestamp: convertDateToClickhouseDateTime(timestamp) }
: {}),
},
tags: {
feature: "tracing",
@@ -1175,6 +1183,7 @@ export const getLatencyAndTotalCostForObservations = async (
export const getLatencyAndTotalCostForObservationsByTraces = async (
projectId: string,
traceIds: string[],
timestamp?: Date,
) => {
const query = `
SELECT
@@ -1184,6 +1193,7 @@ export const getLatencyAndTotalCostForObservationsByTraces = async (
FROM observations FINAL
WHERE project_id = {projectId: String}
AND trace_id IN ({traceIds: Array(String)})
${timestamp ? `AND start_time >= {timestamp: DateTime64(3)}` : ""}
GROUP BY trace_id
`;
const rows = await queryClickhouse<{
@@ -1195,6 +1205,9 @@ export const getLatencyAndTotalCostForObservationsByTraces = async (
params: {
projectId,
traceIds,
...(timestamp
? { timestamp: convertDateToClickhouseDateTime(timestamp) }
: {}),
},
tags: {
feature: "tracing",
@@ -1311,6 +1324,59 @@ export const getTraceIdsForObservations = async (
}));
};
export const getObservationsForBlobStorageExport = function (
projectId: string,
minTimestamp: Date,
maxTimestamp: Date,
) {
const query = `
SELECT
id,
trace_id,
project_id,
environment,
type,
parent_observation_id,
start_time,
end_time,
name,
metadata,
level,
status_message,
version,
input,
output,
provided_model_name,
model_parameters,
usage_details,
cost_details,
completion_start_time,
prompt_name,
prompt_version
FROM observations FINAL
WHERE project_id = {projectId: String}
AND start_time >= {minTimestamp: DateTime64(3)}
AND start_time <= {maxTimestamp: DateTime64(3)}
`;
const records = queryClickhouseStream<Record<string, unknown>>({
query,
params: {
projectId,
minTimestamp: convertDateToClickhouseDateTime(minTimestamp),
maxTimestamp: convertDateToClickhouseDateTime(maxTimestamp),
},
tags: {
feature: "blobstorage",
type: "observation",
kind: "analytic",
projectId,
},
});
return records;
};
export const getGenerationsForPostHog = async function* (
projectId: string,
minTimestamp: Date,
@@ -1,14 +1,14 @@
import {
Observation,
ObservationView,
ObservationType,
ObservationLevel,
Prisma,
} from "@prisma/client";
import { Prisma } from "@prisma/client";
import Decimal from "decimal.js";
import { parseClickhouseUTCDateTimeFormat } from "./clickhouse";
import { ObservationRecordReadType } from "./definitions";
import { parseJsonPrioritised } from "../../utils/json";
import {
Observation,
ObservationView,
ObservationType,
ObservationLevelType,
} from "./types";
export const convertObservationToView = (
record: ObservationRecordReadType,
@@ -55,6 +55,7 @@ export const convertObservation = (
traceId: record.trace_id ?? null,
projectId: record.project_id,
type: record.type as ObservationType,
environment: record.environment,
parentObservationId: record.parent_observation_id ?? null,
startTime: parseClickhouseUTCDateTimeFormat(record.start_time),
endTime: record.end_time
@@ -69,7 +70,7 @@ export const convertObservation = (
val && parseJsonPrioritised(val),
]),
),
level: record.level as ObservationLevel,
level: record.level as ObservationLevelType,
statusMessage: record.status_message ?? null,
version: record.version ?? null,
input: (record.input
+164 -59
View File
@@ -1,4 +1,5 @@
import { Score, ScoreDataType, ScoreSource } from "@prisma/client";
import { ScoreDataType } from "@prisma/client";
import { Score, ScoreSourceType } from "./types";
import {
commandClickhouse,
parseClickhouseUTCDateTimeFormat,
@@ -6,7 +7,7 @@ import {
queryClickhouseStream,
upsertClickhouse,
} from "./clickhouse";
import { FilterList } from "../queries/clickhouse-sql/clickhouse-filter";
import { FilterList, orderByToClickhouseSql } from "../queries";
import { FilterCondition, FilterState, TimeFilter } from "../../types";
import {
createFilterFromFilterState,
@@ -17,7 +18,6 @@ import {
dashboardColumnDefinitions,
scoresTableUiColumnDefinitions,
} from "../../tableDefinitions";
import { orderByToClickhouseSql } from "../queries/clickhouse-sql/orderby-factory";
import {
convertScoreAggregation,
convertToScore,
@@ -77,7 +77,7 @@ export const searchExistingAnnotationScore = async (
export const getScoreById = async (
projectId: string,
scoreId: string,
source?: ScoreSource,
source?: ScoreSourceType,
) => {
const query = `
SELECT *
@@ -110,7 +110,7 @@ export const getScoreById = async (
export const getScoresByIds = async (
projectId: string,
scoreId: string[],
source?: ScoreSource,
source?: ScoreSourceType,
) => {
const query = `
SELECT *
@@ -281,7 +281,7 @@ export const getScoresGroupedByNameSourceType = async (
return rows.map((row) => ({
name: row.name,
source: row.source as ScoreSource,
source: row.source as ScoreSourceType,
dataType: row.data_type as ScoreDataType,
}));
};
@@ -346,9 +346,7 @@ export const getScoresUiCount = async (props: {
offset?: number;
}) => {
const rows = await getScoresUiGeneric<{ count: string }>({
select: `
count(*) as count
`,
select: "count",
tags: { kind: "count" },
...props,
});
@@ -372,6 +370,7 @@ export const getScoresUiTable = async (props: {
const rows = await getScoresUiGeneric<{
id: string;
project_id: string;
environment: string;
name: string;
value: number;
string_value: string | null;
@@ -393,9 +392,53 @@ export const getScoresUiTable = async (props: {
created_at: string;
updated_at: string;
}>({
select: `
select: "rows",
tags: { kind: "analytic" },
...props,
});
return rows.map((row) => ({
projectId: row.project_id,
environment: row.environment,
authorUserId: row.author_user_id,
traceId: row.trace_id,
observationId: row.observation_id,
traceUserId: row.user_id,
traceName: row.trace_name,
traceTags: row.trace_tags,
configId: row.config_id,
queueId: row.queue_id,
createdAt: parseClickhouseUTCDateTimeFormat(row.created_at),
updatedAt: parseClickhouseUTCDateTimeFormat(row.updated_at),
stringValue: row.string_value,
comment: row.comment,
dataType: row.data_type as ScoreDataType,
source: row.source as ScoreSourceType,
name: row.name,
value: row.value,
timestamp: parseClickhouseUTCDateTimeFormat(row.timestamp),
id: row.id,
}));
};
export const getScoresUiGeneric = async <T>(props: {
select: "count" | "rows";
projectId: string;
filter: FilterState;
orderBy: OrderByState;
limit?: number;
offset?: number;
tags?: Record<string, string>;
}): Promise<T[]> => {
const { projectId, filter, orderBy, limit, offset } = props;
const select =
props.select === "count"
? "count(*) as count"
: `
s.id,
s.project_id,
s.environment,
s.name,
s.value,
s.string_value,
@@ -417,62 +460,26 @@ export const getScoresUiTable = async (props: {
t.user_id,
t.name as trace_name,
t.tags as trace_tags
`,
tags: { kind: "analytic" },
...props,
`;
const { scoresFilter } = getProjectIdDefaultFilter(projectId, {
tracesPrefix: "t",
});
return rows.map((row) => ({
projectId: row.project_id,
authorUserId: row.author_user_id,
traceId: row.trace_id,
observationId: row.observation_id,
traceUserId: row.user_id,
traceName: row.trace_name,
traceTags: row.trace_tags,
configId: row.config_id,
queueId: row.queue_id,
createdAt: parseClickhouseUTCDateTimeFormat(row.created_at),
updatedAt: parseClickhouseUTCDateTimeFormat(row.updated_at),
stringValue: row.string_value,
comment: row.comment,
dataType: row.data_type as ScoreDataType,
source: row.source as ScoreSource,
name: row.name,
value: row.value,
timestamp: parseClickhouseUTCDateTimeFormat(row.timestamp),
id: row.id,
}));
};
export const getScoresUiGeneric = async <T>(props: {
select: string;
projectId: string;
filter: FilterState;
orderBy: OrderByState;
limit?: number;
offset?: number;
tags?: Record<string, string>;
}): Promise<T[]> => {
const { select, projectId, filter, orderBy, limit, offset } = props;
const { tracesFilter, scoresFilter, observationsFilter } =
getProjectIdDefaultFilter(projectId, { tracesPrefix: "t" });
scoresFilter.push(
...createFilterFromFilterState(filter, scoresTableUiColumnDefinitions),
);
const scoresFilterRes = scoresFilter.apply();
// TODO: Can we realistically apply a traces time filter here? Is an order by event_ts a risk?
// Only join traces for rows or if there is a trace filter on counts
const performTracesJoin =
props.select === "rows" ||
scoresFilter.some((f) => f.clickhouseTable === "traces");
const query = `
SELECT
${select}
FROM scores s final
LEFT JOIN traces t
ON s.trace_id = t.id
AND t.project_id = s.project_id
${performTracesJoin ? "LEFT JOIN traces t ON s.trace_id = t.id AND t.project_id = s.project_id" : ""}
WHERE s.project_id = {projectId: String}
${scoresFilterRes?.query ? `AND ${scoresFilterRes.query}` : ""}
${orderByToClickhouseSql(orderBy ?? null, scoresTableUiColumnDefinitions)}
@@ -547,17 +554,17 @@ export const getScoreNames = async (
}));
};
export const deleteScore = async (projectId: string, scoreId: string) => {
export const deleteScores = async (projectId: string, scoreIds: string[]) => {
const query = `
DELETE FROM scores
WHERE project_id = {projectId: String}
AND id = {scoreId: String};
AND id in ({scoreIds: Array(String)});
`;
await commandClickhouse({
query: query,
params: {
projectId,
scoreId,
scoreIds,
},
tags: {
feature: "tracing",
@@ -767,6 +774,37 @@ export const getScoreCountsByProjectInCreationInterval = async ({
}));
};
export const getScoreCountOfProjectsSinceCreationDate = async ({
projectIds,
start,
}: {
projectIds: string[];
start: Date;
}) => {
const query = `
SELECT
count(*) as count
FROM scores
WHERE project_id IN ({projectIds: Array(String)})
AND created_at >= {start: DateTime64(3)}
`;
const rows = await queryClickhouse<{ count: string }>({
query,
params: {
projectIds,
start: convertDateToClickhouseDateTime(start),
},
tags: {
feature: "tracing",
type: "score",
kind: "analytic",
},
});
return Number(rows[0]?.count ?? 0);
};
export const getDistinctScoreNames = async (
projectId: string,
cutoffCreatedAt: Date,
@@ -808,6 +846,49 @@ export const getDistinctScoreNames = async (
return rows.map((row) => row.name);
};
export const getScoresForBlobStorageExport = function (
projectId: string,
minTimestamp: Date,
maxTimestamp: Date,
) {
const query = `
SELECT
id,
timestamp,
project_id,
environment,
trace_id,
observation_id,
name,
value,
source,
comment,
data_type,
string_value
FROM scores FINAL
WHERE project_id = {projectId: String}
AND timestamp >= {minTimestamp: DateTime64(3)}
AND timestamp <= {maxTimestamp: DateTime64(3)}
`;
const records = queryClickhouseStream<Record<string, unknown>>({
query,
params: {
projectId,
minTimestamp: convertDateToClickhouseDateTime(minTimestamp),
maxTimestamp: convertDateToClickhouseDateTime(maxTimestamp),
},
tags: {
feature: "blobstorage",
type: "score",
kind: "analytic",
projectId,
},
});
return records;
};
export const getScoresForPostHog = async function* (
projectId: string,
minTimestamp: Date,
@@ -875,3 +956,27 @@ export const getScoresForPostHog = async function* (
};
}
};
export const hasAnyScore = async (projectId: string) => {
const query = `
SELECT 1
FROM scores
WHERE project_id = {projectId: String}
LIMIT 1
`;
const rows = await queryClickhouse<{ 1: number }>({
query,
params: {
projectId,
},
tags: {
feature: "tracing",
type: "score",
kind: "hasAny",
projectId,
},
});
return rows.length > 0;
};

Some files were not shown because too many files have changed in this diff Show More