chore: release v2.49.2

* chore: update MainNavigation component to use nested navigation literals

* simplify

---------

Co-authored-by: Marc Klingen <git@marcklingen.com>

.env.dev.example

@@ -47,5 +47,5 @@ REDIS_PORT=6379
 REDIS_AUTH="myredissecret"
 
 LANGFUSE_WORKER_PASSWORD=mybasicauthsecret
-# openssl rand -base64 32 used only here
+# openssl rand -hex 32 used only here
 ENCRYPTION_KEY=6c16874e5c0f0cc74ddec00425fa99fbe9ffbe412b7d5d906a4b00005df91403 

.env.prod.example

@@ -58,6 +58,11 @@ LANGFUSE_CSP_ENFORCE_HTTPS="true"
 # AUTH_AUTH0_CLIENT_SECRET=
 # AUTH_AUTH0_ISSUER=
 # AUTH_AUTH0_ALLOW_ACCOUNT_LINKING=false
+# AUTH_COGNITO_CLIENT_ID=
+# AUTH_COGNITO_CLIENT_SECRET=
+# AUTH_COGNITO_ISSUER=
+# AUTH_COGNITO_ALLOW_ACCOUNT_LINKING=false
+
 
 # Transactional email, optional
 # Defines the email address to use as the from address.
@@ -96,7 +101,7 @@ LANGFUSE_CSP_ENFORCE_HTTPS="true"
 
 ### START Langfuse Cloud Config
 # Used for Langfuse Cloud deployments
-# Not recommended for self-hosted deployments as these are NOT COVERED BY SEMVER
+# Not recommended for self-hosted deployments as these are NOT COVERED BY SEMANTIC VERSIONING
 
 # NEXT_PUBLIC_LANGFUSE_CLOUD_REGION="US"
 # NEXTAUTH_COOKIE_DOMAIN=".langfuse.com"

.github/workflows/pipeline.yml

@@ -40,19 +40,27 @@ jobs:
       - name: Checkout
         uses: actions/checkout@v3
 
-      - name: Build Web Docker image
-        uses: docker/build-push-action@v4
-        with:
-          context: .
-          file: ./web/Dockerfile
-          push: false
+      - name: Build and run both images from compose
+        run: |
+          docker compose -f docker-compose.build.yml up -d
+          sleep 5 # Wait for PostgreSQL to accept connections
 
-      - name: Build Worker Docker image
-        uses: docker/build-push-action@v4
-        with:
-          context: .
-          file: ./worker/Dockerfile
-          push: false
+      - name: Check server health
+        run: |
+          timeout 10 bash -c 'until curl -f http://localhost:3000/api/public/health; do sleep 2; done'
+
+      - name: Check worker health
+        run: |
+          timeout 10 bash -c 'until curl -f http://localhost:3030/api/health; do sleep 2; done'
+
+      - name: Ensure no unhealthy status
+        run: |
+          if docker-compose ps | grep "(unhealthy)"; then
+            echo "One or more services are unhealthy"
+            exit 1
+          else
+            echo "All services are healthy"
+          fi
 
   tests-web:
     runs-on: ubuntu-latest
@@ -109,9 +117,6 @@ jobs:
     strategy:
       matrix:
         node-version: [20]
-    env:
-      OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-      ENCRYPTION_KEY: ${{ secrets.ENCRYPTION_KEY }}
     steps:
       - name: Set Swap Space
         uses: pierotofy/set-swap-space@master

.github/workflows/porter_stack_worker.yml

@@ -0,0 +1,29 @@
+"on":
+  push:
+    branches:
+    - main
+name: Deploy to worker
+jobs:
+  porter-deploy:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout code
+      uses: actions/checkout@v3
+    - name: Set Github tag
+      id: vars
+      run: echo "sha_short=$(git rev-parse --short HEAD)" >> $GITHUB_OUTPUT
+    - name: Setup porter
+      uses: porter-dev/setup-porter@v0.1.0
+    - name: Deploy stack
+      timeout-minutes: 30
+      run: exec porter apply
+      env:
+        PORTER_CLUSTER: "3959"
+        PORTER_DEPLOYMENT_TARGET_ID: d2bb23e4-1d77-48f2-a113-383d076959b3
+        PORTER_HOST: https://dashboard.getporter.dev
+        PORTER_PR_NUMBER: ${{ github.event.number }}
+        PORTER_PROJECT: "12565"
+        PORTER_REPO_NAME: ${{ github.event.repository.name }}
+        PORTER_STACK_NAME: worker
+        PORTER_TAG: ${{ steps.vars.outputs.sha_short }}
+        PORTER_TOKEN: ${{ secrets.PORTER_STACK_12565_3959 }}

CONTRIBUTING.md

@@ -191,6 +191,16 @@ Requirements
     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
+
 ## Monorepo quickstart
 
 - Available packages and their dependencies
@@ -290,7 +300,77 @@ When a new release is tagged on the `main` branch (excluding prereleases), it tr
 1. The Docker image is published to GitHub Packages with the version number and `latest` tag.
 2. The deployment is carried out on Langfuse Cloud. This is done by force pushing the `main` branch to the `production` branch during every release, using the [`release.yml`](.github/workflows/release.yml) GitHub Action.
 
-## Using secrets stored in 1Password
+## Theming
+
+At Langfuse, we utilize CSS variables to manage our theme settings across the platform.
+
+Our approach leverages separate CSS variables for backgrounds (--background) and foregrounds (--foreground), fully adhering to the [shadcn/ui](https://ui.shadcn.com/docs/theming) color conventions. The background suffix can be omitted if the variable is used for the background color of the component. We recommend using HSL values for these colors to enhance consistency and customization. There is no need to manually handle dark mode styling with "dark:" prefixes, as next-themes automatically manages the theme switching.
+
+Given the following CSS variables:
+
+```
+--primary: 222.2 47.4% 11.2%; // e.g. background-color
+--primary-foreground: 210 40% 98%; // e.g. text-color
+```
+
+The background color of the following component will be `hsl(var(--primary))` and the foreground color will be `hsl(var(--primary-foreground))`.
+
+```
+<div class="bg-primary text-primary-foreground">Hello</div>
+```
+
+### Color Variables
+
+| Variable                 | Description                                                        | Examples                         |
+| ------------------------ | ------------------------------------------------------------------ | -------------------------------- |
+| --background             | Background color                                                   | Default background color of body |
+| --foreground             | Foreground color                                                   | Default text color of body       |
+| --muted                  | Muted background color                                             | TabsList, Skeleton and Switch    |
+| --muted-foreground       | Muted foreground color                                             |                                  |
+| --popover                | Popover background color                                           | DropdownMenu, HoverCard, Popover |
+| --popover-foreground     | Popover foreground color                                           |                                  |
+| --card                   | Card background color                                              | Card                             |
+| --card-foreground        | Card foreground color                                              |                                  |
+| --border                 | Border color                                                       | Default border color             |
+| --input                  | Input field border color                                           | Input, Select, Textarea          |
+| --primary                | Primary button background colors                                   | Button variant="primary"         |
+| --primary-foreground     | Primary button foreground color                                    |                                  |
+| --secondary              | Secondary button background color                                  | Button variant="secondary"       |
+| --secondary-foreground   | Secondary button foreground color                                  |                                  |
+| --accent                 | Used for accents such as hover effects                             | DropdownMenuItem, SelectItem     |
+| --accent-foreground      | Used for texts on hover effects                                    | DropdownMenuItem, SelectItem     |
+| --destructive            | Destructive action color for background                            | Button variant="destructive"     |
+| --destructive-foreground | Destructive action color for text                                  |                                  |
+| --ring                   | Focus ring color                                                   | MultiSelect                      |
+| --primary-accent         | Primary accent color used for branding                             | Layout                           |
+| --hover-primary-accent   | Primary accent color used for hover effects for links              | SignIn and AuthCloudRegionSwitch |
+| --muted-green            | Muted green for Event label                                        | ObservationTree                  |
+| --muted-orange           | Muted orange for Generation label                                  | ObservationTree                  |
+| --muted-blue             | Muted blue for Span label                                          | ObservationTree                  |
+| --muted-gray             | Muted gray for disabled status badges                              | StatusBadge                      |
+| --accent-light-green     | Light green accent for background of output and assistant messages | IOPreview, Generations, Traces   |
+| --accent-dark-green      | Dark green accent for border of output and assistant messages      | CodeJsonViewer and IOPReview     |
+| --light-red              | Light red for error background                                     | level-color and StatusBadge      |
+| --dark-red               | Dark red for error text and error badge dot color                  | level-color and ErrorPage        |
+| --light-yellow           | Light yellow for warning background                                | LevelColor                       |
+| --dark-yellow            | Dark yellow for warning text                                       | LevelColor                       |
+| --light-green            | Light green for success status badge background                    | StatusBadge                      |
+| --dark-green             | Dark green for success status badge text and dot                   | StatusBadge                      |
+| --light-blue             | Light blue for background of Staging label                         | LangfuseLogo                     |
+| --dark-blue              | Dark blue for text and border of Staging label                     | LangfuseLogo                     |
+| --accent-light-blue      | Light blue accent for table link hover effect                      | TableLink                        |
+| --accent-dark-blue       | Dark blue accent for table link text                               | TableLink                        |
+
+### Adding New Colors
+
+1. Global Definitions: Add new CSS variable definitions in the global.css file.
+2. Tailwind Configuration: Reflect these new colors in the tailwind.config.js to maintain alignment with Tailwind's utility classes.
+
+By following these guidelines, you can ensure that any contributions to our theme are consistent, maintainable, and aligned with our design system.
+
+## Maintainers
+
+### Using secrets stored in 1Password
 
 When applying changes to non-local environments, you may need to use secrets stored in 1Password. We use the 1Password CLI for this purpose.
 

README.md

@@ -106,22 +106,26 @@ See the [→ Quickstart](https://langfuse.com/docs/get-started) to integrate Lan
 
 ### Integrations
 
-| Integration                                                      | Supports      | Description                                                 |
-| ---------------------------------------------------------------- | ------------- | ----------------------------------------------------------- |
-| [**SDK** - _recommended_](https://langfuse.com/docs/sdk)         | Python, JS/TS | Manual instrumentation using the SDKs for full flexibility. |
-| [OpenAI SDK](https://langfuse.com/docs/integrations/openai)      | Python, JS/TS | Automated instrumentation of OpenAI SDK.                    |
-| [Langchain](https://langfuse.com/docs/integrations/langchain)    | Python, JS/TS | Instrumentation via Langchain callbacks.                    |
-| [LlamaIndex](https://langfuse.com/docs/integrations/llama-index) | Python        | Automated instrumentation via LlamaIndex callback system.   |
-| [API](https://langfuse.com/docs/api)                             |               | Directly call the public API. OpenAPI spec available.       |
+| Integration                                              | Supports                   | Description                                                                                                                                      |
+| -------------------------------------------------------- | -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ |
+| [SDK](/docs/sdk)                                         | Python, JS/TS              | Manual instrumentation using the SDKs for full flexibility.                                                                                      |
+| [OpenAI](/docs/integrations/openai)                      | Python, JS/TS              | Automated instrumentation using drop-in replacement of OpenAI SDK.                                                                               |
+| [Langchain](/docs/integrations/langchain)                | Python, JS/TS              | Automated instrumentation by passing callback handler to Langchain application.                                                                  |
+| [LlamaIndex](/docs/integrations/llama-index/get-started) | Python                     | Automated instrumentation via LlamaIndex callback system.                                                                                        |
+| [Haystack](/docs/integrations/haystack)                  | Python                     | Automated instrumentation via Haystack content tracing system.                                                                                   |
+| [LiteLLM](/docs/integrations/litellm)                    | Python, JS/TS (proxy only) | Use any LLM as a drop in replacement for GPT. Use Azure, OpenAI, Cohere, Anthropic, Ollama, VLLM, Sagemaker, HuggingFace, Replicate (100+ LLMs). |
+| [API](/docs/api)                                         |                            | Directly call the public API. OpenAPI spec available.                                                                                            |
 
-External projects/packages that integrate with Langfuse:
+Packages that integrate with Langfuse:
 
-| Name                                                            | Description                                                                                                                                      |
-| --------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ |
-| [LiteLLM](https://langfuse.com/docs/integrations/litellm)       | Use any LLM as a drop in replacement for GPT. Use Azure, OpenAI, Cohere, Anthropic, Ollama, VLLM, Sagemaker, HuggingFace, Replicate (100+ LLMs). |
-| [Flowise](https://langfuse.com/docs/integrations/flowise)       | JS/TS no-code builder for customized LLM flows.                                                                                                  |
-| [Langflow](https://langfuse.com/docs/integrations/langflow)     | Python-based UI for LangChain, designed with react-flow to provide an effortless way to experiment and prototype flows.                          |
-| [Superagent](https://langfuse.com/docs/integrations/superagent) | Open Source AI Assistant Framework & API for prototyping and deployment of agents.                                                               |
+| Name                                                       | Description                                                                                                             |
+| ---------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
+| [Instructor](/docs/integrations/instructor)                | Library to get structured LLM outputs (JSON, Pydantic)                                                                  |
+| [Mirascope](/docs/integrations/mirascope)                  | Python toolkit for building LLM applications.                                                                           |
+| [AI SDK by Vercel](/docs/sdk/typescript/example-vercel-ai) | Typescript SDK that makes streaming LLM outputs super easy.                                                             |
+| [Flowise](/docs/integrations/flowise)                      | JS/TS no-code builder for customized LLM flows.                                                                         |
+| [Langflow](/docs/integrations/langflow)                    | Python-based UI for LangChain, designed with react-flow to provide an effortless way to experiment and prototype flows. |
+| [Superagent](/docs/integrations/superagent)                | Open Source AI Assistant Framework & API for prototyping and deployment of agents.                                      |
 
 ## Questions and feedback
 

careers

@@ -0,0 +1,14 @@
+_We are Hiring_
+Join us in scaling Langfuse in Berlin, Germany. We are an open source company, we hire in person, we are only hiring technical talent.
+
+_Open Roles_
+
+- Backend Engineer, 70-130k EUR, 0.25-0.75% Equity, https://www.ycombinator.com/companies/langfuse/jobs/mnrdwla-backend-engineer
+- Product Engineer, 70-130k EUR, 0.25-0.75% Equity, https://www.ycombinator.com/companies/langfuse/jobs/aAvmoFB-product-engineer
+- Developer Advocate, 60-110k EUR, 0.25-0.5% Equity, https://www.ycombinator.com/companies/langfuse/jobs/uHysbKH-developer-advocate-devrel
+
+_More Info_
+
+- https://langfuse.com/careers
+- https://langfuse.com/docs
+- https://langfuse.com/changelog

docker-compose.build.yml

@@ -20,6 +20,11 @@ services:
       - LANGFUSE_WORKER_HOST=${LANGFUSE_WORKER_HOST:-worker}
       - LANGFUSE_WORKER_PASSWORD=${LANGFUSE_WORKER_PASSWORD:-mybasicauthsecret}
     restart: always
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:3000/api/public/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
 
   worker:
     build:
@@ -41,6 +46,11 @@ services:
       - REDIS_AUTH=${REDIS_AUTH:-myredissecret}
       - LANGFUSE_WORKER_PASSWORD=${LANGFUSE_WORKER_PASSWORD:-mybasicauthsecret}
     restart: always
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:3030/api/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
 
   redis:
     image: redis:7.2.4

docker-compose.yml

@@ -4,7 +4,8 @@ services:
   langfuse-server:
     image: ghcr.io/langfuse/langfuse:latest
     depends_on:
-      - db
+      db:
+        condition: service_healthy
     ports:
       - "3000:3000"
     environment:
@@ -18,6 +19,11 @@ services:
   db:
     image: postgres
     restart: always
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U postgres"]
+      interval: 3s
+      timeout: 3s
+      retries: 10
     environment:
       - POSTGRES_USER=postgres
       - POSTGRES_PASSWORD=postgres

ee/package.json

@@ -33,7 +33,7 @@
     "@repo/eslint-config": "*",
     "@repo/typescript-config": "*",
     "@types/node": "^20.11.29",
-    "@typescript-eslint/parser": "^7.7.0",
+    "@typescript-eslint/parser": "^7.12.0",
     "eslint": "^8.57.0",
     "eslint-config-prettier": "^9.1.0",
     "eslint-config-standard": "^17.1.0",

ee/src/sso/index.ts

@@ -2,10 +2,11 @@ import { type Provider } from "next-auth/providers/index";
 import GoogleProvider from "next-auth/providers/google";
 import GitHubProvider from "next-auth/providers/github";
 import OktaProvider from "next-auth/providers/okta";
+import CognitoProvider from "next-auth/providers/cognito";
 import Auth0Provider from "next-auth/providers/auth0";
 import AzureADProvider from "next-auth/providers/azure-ad";
 import { isEeAvailable } from "..";
-import { prisma } from "@langfuse/shared/src/db";
+import { type SsoConfig, prisma } from "@langfuse/shared/src/db";
 import { encrypt, decrypt } from "@langfuse/shared/encryption";
 import { SsoProviderSchema } from "./types";
 import { type NextApiRequest, type NextApiResponse } from "next";
@@ -25,14 +26,36 @@ let cachedSsoConfigs: {
 async function getSsoConfigs(): Promise<SsoProviderSchema[]> {
   if (!isEeAvailable) return [];
   const CACHE_TTL = 60 * 1000; // 1 minute
+  const DB_MAX_WAIT = 2000; // 2 seconds
+  const DB_TIMEOUT = 3000; // 3 seconds
 
   // Set/refresh the cache if it's empty or expired
   if (
     cachedSsoConfigs === null ||
     Date.now() - cachedSsoConfigs.timestamp > CACHE_TTL
   ) {
+    // findMany with custom timeout via $transaction
+    let dbConfigs: SsoConfig[] = [];
+    try {
+      dbConfigs = await prisma.$transaction(
+        async (prisma) => prisma.ssoConfig.findMany(),
+        {
+          maxWait: DB_MAX_WAIT,
+          timeout: DB_TIMEOUT,
+        }
+      );
+    } catch (e) {
+      // cache empty array to prevent repeated DB calls on error
+      cachedSsoConfigs = {
+        data: [],
+        timestamp: Date.now(),
+      };
+
+      // caught and logged in the caller
+      throw e;
+    }
+
     // transform into zod object
-    const dbConfigs = await prisma.ssoConfig.findMany();
     const parsedSsoConfigs = dbConfigs
       .map((v) => {
         try {
@@ -144,6 +167,12 @@ const dbToNextAuthProvider = (provider: SsoProviderSchema): Provider | null => {
       ...provider.authConfig,
       clientSecret: decrypt(provider.authConfig.clientSecret),
     });
+  else if (provider.authProvider === "cognito")
+    return CognitoProvider({
+      id: getAuthProviderIdForSsoConfig(provider), // use the domain as the provider id as we use domain-specific credentials
+      ...provider.authConfig,
+      clientSecret: decrypt(provider.authConfig.clientSecret),
+    });
   else {
     // Type check to ensure we handle all providers
     // eslint-disable-next-line no-unused-vars

ee/src/sso/types.ts

@@ -64,11 +64,24 @@ export const AzureAdProviderSchema = base.extend({
     .nullish(),
 });
 
+export const CognitoProviderSchema = base.extend({
+  authProvider: z.literal("cognito"),
+  authConfig: z
+    .object({
+      clientId: z.string(),
+      clientSecret: z.string(),
+      issuer: z.string(),
+      allowDangerousEmailAccountLinking: z.boolean().optional().default(false),
+    })
+    .nullish(),
+});
+
 export type GoogleProviderSchema = z.infer<typeof GoogleProviderSchema>;
 export type GithubProviderSchema = z.infer<typeof GithubProviderSchema>;
 export type Auth0ProviderSchema = z.infer<typeof Auth0ProviderSchema>;
 export type OktaProviderSchema = z.infer<typeof OktaProviderSchema>;
 export type AzureAdProviderSchema = z.infer<typeof AzureAdProviderSchema>;
+export type CognitoProviderSchema = z.infer<typeof CognitoProviderSchema>;
 
 export const SsoProviderSchema = z.discriminatedUnion("authProvider", [
   GoogleProviderSchema,
@@ -76,6 +89,7 @@ export const SsoProviderSchema = z.discriminatedUnion("authProvider", [
   Auth0ProviderSchema,
   OktaProviderSchema,
   AzureAdProviderSchema,
+  CognitoProviderSchema,
 ]);
 
 export type SsoProviderSchema = z.infer<typeof SsoProviderSchema>;

fern/apis/server/definition/commons.yml

@@ -209,8 +209,8 @@ types:
       - ARCHIVED
   ScoreSource:
     enum:
+      - ANNOTATION
       - API
-      - REVIEW
       - EVAL
 
 errors:

fern/apis/server/definition/metrics.yml

@@ -48,12 +48,25 @@ types:
       date: date
       countTraces: integer
       countObservations: integer
-      totalCost: double
+      totalCost:
+        type: double
+        docs: Total model cost in USD
       usage: list<UsageByModel>
   UsageByModel:
     docs: Daily usage of a given model. Usage corresponds to the unit set for the specific model (e.g. tokens).
     properties:
       model: optional<string>
-      inputUsage: integer
-      outputUsage: integer
-      totalUsage: integer
+      inputUsage:
+        type: integer
+        docs: Total number of generation input units (e.g. tokens)
+      outputUsage:
+        type: integer
+        docs: Total number of generation output units (e.g. tokens)
+      totalUsage:
+        type: integer
+        docs: Total number of generation total units (e.g. tokens)
+      countTraces: integer
+      countObservations: integer
+      totalCost:
+        type: double
+        docs: Total model cost in USD

fern/apis/server/definition/prompts.yml

@@ -70,14 +70,24 @@ types:
       name: string
       prompt: list<ChatMessage>
       config: optional<unknown>
-      labels: optional<list<string>>
+      labels:
+        type: optional<list<string>>
+        docs: List of deployment labels of this prompt version.
+      tags:
+        type: optional<list<string>>
+        docs: List of tags to apply to all versions of this prompt.
 
   CreateTextPromptRequest:
     properties:
       name: string
       prompt: string
       config: optional<unknown>
-      labels: optional<list<string>>
+      labels:
+        type: optional<list<string>>
+        docs: List of deployment labels of this prompt version.
+      tags:
+        type: optional<list<string>>
+        docs: List of tags to apply to all versions of this prompt.
 
   Prompt:
     union:
@@ -89,7 +99,12 @@ types:
       name: string
       version: integer
       config: unknown
-      labels: list<string>
+      labels:
+        type: list<string>
+        docs: List of deployment labels of this prompt version.
+      tags:
+        type: list<string>
+        docs: List of tags. Used to filter via UI and API. The same across versions of a prompt.
 
   ChatMessage:
     properties:

fern/apis/server/definition/sessions.yml

@@ -7,7 +7,7 @@ service:
   base-path: /api/public
   endpoints:
     get:
-      docs: Get a session
+      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>`
       method: GET
       path: /sessions/{sessionId}
       path-parameters:

fern/apis/server/definition/trace.yml

@@ -30,6 +30,7 @@ service:
             docs: Limit of items per page. If you encounter api issues due to too large page sizes, try to reduce the limit.
           userId: optional<string>
           name: optional<string>
+          sessionId: optional<string>
           fromTimestamp:
             type: optional<datetime>
             docs: Retrieve only traces newer than this datetime (ISO 8601).

package.json

@@ -1,6 +1,6 @@
 {
   "name": "langfuse",
-  "version": "2.41.1",
+  "version": "2.49.2",
   "author": "engineering@langfuse.com",
   "license": "MIT",
   "private": true,
@@ -28,7 +28,7 @@
   },
   "devDependencies": {
     "@release-it/bumper": "^6.0.1",
-    "dotenv-cli": "^7.4.1",
+    "dotenv-cli": "^7.4.2",
     "prettier": "^3.2.5",
     "release-it": "^17.2.1",
     "turbo": "^1.13.3"

packages/config-eslint/package.json

@@ -9,7 +9,7 @@
   ],
   "devDependencies": {
     "@typescript-eslint/eslint-plugin": "^7.1.0",
-    "@typescript-eslint/parser": "^7.7.0",
+    "@typescript-eslint/parser": "^7.12.0",
     "@vercel/style-guide": "^6.0.0",
     "eslint-config-next": "^14.2.3",
     "eslint-config-prettier": "^9.1.0",

packages/shared/package.json

@@ -44,14 +44,15 @@
     "seed": "ts-node -r tsconfig-paths/register -r dotenv/config --compiler-options {\"module\":\"CommonJS\"} prisma/seed.ts"
   },
   "dependencies": {
-    "@langchain/anthropic": "^0.1.16",
+    "@langchain/anthropic": "^0.1.21",
     "@langchain/core": "^0.1.61",
-    "@langchain/openai": "^0.0.28",
+    "@langchain/openai": "^0.0.33",
     "@prisma/client": "^5.13.0",
     "@types/bcryptjs": "^2.4.6",
     "bcryptjs": "^2.4.3",
     "kysely": "^0.27.3",
     "langchain": "^0.1.36",
+    "lodash": "^4.17.21",
     "prisma-extension-kysely": "^2.1.0",
     "zod": "^3.22.4",
     "zod-to-json-schema": "^3.22.5"
@@ -59,11 +60,11 @@
   "devDependencies": {
     "@repo/eslint-config": "*",
     "@repo/typescript-config": "*",
-    "@types/lodash": "^4.14.202",
+    "@types/lodash": "^4.17.4",
     "@types/node": "^20.11.29",
-    "@types/pg": "^8.11.5",
+    "@types/pg": "^8.11.6",
     "@types/uuid": "^9.0.8",
-    "@typescript-eslint/parser": "^7.7.0",
+    "@typescript-eslint/parser": "^7.12.0",
     "eslint": "^8.57.0",
     "eslint-config-prettier": "^9.1.0",
     "eslint-config-standard": "^17.1.0",

packages/shared/prisma/generated/types.ts

@@ -25,22 +25,17 @@ export const ObservationLevel = {
 } as const;
 export type ObservationLevel = (typeof ObservationLevel)[keyof typeof ObservationLevel];
 export const ScoreSource = {
+    ANNOTATION: "ANNOTATION",
     API: "API",
-    REVIEW: "REVIEW",
     EVAL: "EVAL"
 } as const;
 export type ScoreSource = (typeof ScoreSource)[keyof typeof ScoreSource];
-export const PricingUnit = {
-    PER_1000_TOKENS: "PER_1000_TOKENS",
-    PER_1000_CHARS: "PER_1000_CHARS"
+export const ScoreDataType = {
+    CATEGORICAL: "CATEGORICAL",
+    NUMERIC: "NUMERIC",
+    BOOLEAN: "BOOLEAN"
 } as const;
-export type PricingUnit = (typeof PricingUnit)[keyof typeof PricingUnit];
-export const TokenType = {
-    PROMPT: "PROMPT",
-    COMPLETION: "COMPLETION",
-    TOTAL: "TOTAL"
-} as const;
-export type TokenType = (typeof TokenType)[keyof typeof TokenType];
+export type ScoreDataType = (typeof ScoreDataType)[keyof typeof ScoreDataType];
 export const DatasetStatus = {
     ACTIVE: "ACTIVE",
     ARCHIVED: "ARCHIVED"
@@ -246,6 +241,7 @@ export type Observation = {
     status_message: string | null;
     version: string | null;
     created_at: Generated<Timestamp>;
+    updated_at: Generated<Timestamp>;
     model: string | null;
     internal_model: string | null;
     modelParameters: unknown | null;
@@ -293,6 +289,7 @@ export type ObservationView = {
     calculated_output_cost: string | null;
     calculated_total_cost: string | null;
     latency: number | null;
+    time_to_first_token: number | null;
 };
 export type PosthogIntegration = {
     project_id: string;
@@ -302,14 +299,6 @@ export type PosthogIntegration = {
     enabled: boolean;
     created_at: Generated<Timestamp>;
 };
-export type Pricing = {
-    id: string;
-    model_name: string;
-    pricing_unit: Generated<PricingUnit>;
-    price: string;
-    currency: Generated<string>;
-    token_type: TokenType;
-};
 export type Project = {
     id: string;
     created_at: Generated<Timestamp>;
@@ -342,13 +331,32 @@ export type Prompt = {
 export type Score = {
     id: string;
     timestamp: Generated<Timestamp>;
-    project_id: string | null;
+    project_id: string;
     name: string;
     value: number;
     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;
+    created_at: Generated<Timestamp>;
+    updated_at: Generated<Timestamp>;
+    data_type: Generated<ScoreDataType>;
+};
+export type ScoreConfig = {
+    id: string;
+    created_at: Generated<Timestamp>;
+    updated_at: Generated<Timestamp>;
+    project_id: string;
+    name: string;
+    data_type: ScoreDataType;
+    is_archived: Generated<boolean>;
+    min_value: number | null;
+    max_value: number | null;
+    categories: unknown | null;
+    description: string | null;
 };
 export type Session = {
     id: string;
@@ -444,10 +452,10 @@ export type DB = {
     observations: Observation;
     observations_view: ObservationView;
     posthog_integrations: PosthogIntegration;
-    pricings: Pricing;
     project_memberships: ProjectMembership;
     projects: Project;
     prompts: Prompt;
+    score_configs: ScoreConfig;
     scores: Score;
     Session: Session;
     sso_configs: SsoConfig;

packages/shared/prisma/migrations/20240512155021_scores_drop_fk_on_traces_and_observations/migration.sql

@@ -0,0 +1,5 @@
+-- DropForeignKey
+ALTER TABLE "scores" DROP CONSTRAINT "scores_observation_id_fkey";
+
+-- DropForeignKey
+ALTER TABLE "scores" DROP CONSTRAINT "scores_trace_id_fkey";

packages/shared/prisma/migrations/20240512155022_scores_non_null_and_add_fk_project_id/migration.sql

@@ -0,0 +1,11 @@
+/*
+  Warnings:
+
+  - Made the column `project_id` on table `scores` required. This step will fail if there are existing NULL values in that column.
+
+*/
+-- AlterTable
+ALTER TABLE "scores" ALTER COLUMN "project_id" SET NOT NULL;
+
+-- AddForeignKey
+ALTER TABLE "scores" ADD CONSTRAINT "scores_project_id_fkey" FOREIGN KEY ("project_id") REFERENCES "projects"("id") ON DELETE CASCADE ON UPDATE CASCADE;

packages/shared/prisma/migrations/20240513082203_scores_unique_id_and_projectid_instead_of_id_and_traceid/migration.sql

@@ -0,0 +1,8 @@
+/*
+  Warnings:
+
+  - A unique constraint covering the columns `[id,project_id]` on the table `scores` will be added. If there are existing duplicate values, this will fail.
+
+*/
+-- DropIndex
+DROP INDEX "scores_id_trace_id_key";

packages/shared/prisma/migrations/20240513082204_scores_unique_id_and_projectid_instead_of_id_and_traceid_index/migration.sql

@@ -0,0 +1,8 @@
+/*
+  Warnings:
+
+  - A unique constraint covering the columns `[id,project_id]` on the table `scores` will be added. If there are existing duplicate values, this will fail.
+
+*/
+-- CreateIndex
+CREATE UNIQUE INDEX CONCURRENTLY "scores_id_project_id_key" ON "scores"("id", "project_id");

packages/shared/prisma/migrations/20240513082205_observations_view_add_time_to_first_token/migration.sql

@@ -0,0 +1,61 @@
+CREATE OR REPLACE VIEW "observations_view" AS
+SELECT
+    o.*,
+    m.id AS "model_id",
+    m.start_date AS "model_start_date",
+    m.input_price,
+    m.output_price,
+    m.total_price,
+    m.tokenizer_config AS "tokenizer_config",
+    CASE 
+        WHEN o.input_cost IS NULL AND o.output_cost IS NULL AND o.total_cost IS NULL THEN
+            o.prompt_tokens::decimal * m.input_price
+        ELSE
+            o.input_cost
+    END AS "calculated_input_cost",
+    CASE 
+        WHEN o.input_cost IS NULL AND o.output_cost IS NULL AND o.total_cost IS NULL THEN
+            o.completion_tokens::decimal * m.output_price
+        ELSE
+            o.output_cost
+    END AS "calculated_output_cost",
+    CASE 
+        WHEN o.input_cost IS NULL AND o.output_cost IS NULL AND o.total_cost IS NULL THEN
+            CASE 
+                WHEN m.total_price IS NOT NULL AND o.total_tokens IS NOT NULL THEN
+                    m.total_price * o.total_tokens
+                ELSE
+                    o.prompt_tokens::decimal * m.input_price + 
+                    o.completion_tokens::decimal * m.output_price
+            END
+        ELSE
+            o.total_cost
+    END AS "calculated_total_cost",
+    CASE WHEN o.end_time IS NULL THEN NULL ELSE (EXTRACT(EPOCH FROM o."end_time") - EXTRACT(EPOCH FROM o."start_time"))::double precision END AS "latency",
+    CASE WHEN o.completion_start_time IS NOT NULL AND o.start_time IS NOT NULL THEN EXTRACT(EPOCH FROM (completion_start_time - start_time))::double precision ELSE NULL END as "time_to_first_token"
+    
+FROM
+    observations o
+LEFT JOIN LATERAL (
+    SELECT
+        models.*
+    FROM
+        models
+    WHERE (models.project_id = o.project_id OR models.project_id IS NULL)
+    AND models.model_name = o.internal_model
+    AND (models.start_date < o.start_time OR models.start_date IS NULL)
+    AND o.unit::TEXT = models.unit
+    ORDER BY
+        models.project_id ASC, -- in postgres, NULLs are sorted last when ordering ASC
+        models.start_date DESC NULLS LAST -- now, NULLs are sorted last when ordering DESC as well
+    LIMIT 1
+) m ON TRUE
+
+
+-- requirements:
+-- 1. The view should return all columns from the observations table
+-- 2. The view should match with only one model for each observation if:
+--     a. The model has the same project_id as the observation, otherwise the model without project_id. 
+--     b. The model has the same model_name as the observation
+--     c. The model has a start_date that is less than the observation start_time, otherwise the model without start_date
+--     d. The model has the same unit as the observation

packages/shared/prisma/migrations/20240522081254_scores_add_author_user_id/migration.sql

@@ -0,0 +1,2 @@
+-- AlterTable
+ALTER TABLE "scores" ADD COLUMN     "author_user_id" TEXT;

packages/shared/prisma/migrations/20240522095738_scores_add_author_user_id_index/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "scores_author_user_id_idx" ON "scores"("author_user_id");

packages/shared/prisma/migrations/20240523142425_score_config_add_table/migration.sql

@@ -0,0 +1,42 @@
+-- CreateEnum
+CREATE TYPE "ScoreDataType" AS ENUM ('CATEGORICAL', 'NUMERIC');
+
+-- AlterTable
+ALTER TABLE "scores" ADD COLUMN     "config_id" TEXT,
+ADD COLUMN     "data_type" "ScoreDataType" NOT NULL DEFAULT 'NUMERIC',
+ADD COLUMN     "string_value" TEXT;
+
+-- CreateTable
+CREATE TABLE "score_configs" (
+    "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,
+    "name" TEXT NOT NULL,
+    "data_type" "ScoreDataType" NOT NULL,
+    "is_archived" BOOLEAN NOT NULL DEFAULT false,
+    "min_value" DOUBLE PRECISION,
+    "max_value" DOUBLE PRECISION,
+    "categories" JSONB,
+    "description" TEXT,
+
+    CONSTRAINT "score_configs_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateIndex
+CREATE INDEX "score_configs_data_type_idx" ON "score_configs"("data_type");
+
+-- CreateIndex
+CREATE INDEX "score_configs_is_archived_idx" ON "score_configs"("is_archived");
+
+-- CreateIndex
+CREATE INDEX "score_configs_project_id_idx" ON "score_configs"("project_id");
+
+-- CreateIndex
+CREATE INDEX "score_configs_categories_idx" ON "score_configs"("categories");
+
+-- CreateIndex
+CREATE UNIQUE INDEX "score_configs_id_project_id_key" ON "score_configs"("id", "project_id");
+
+-- AddForeignKey
+ALTER TABLE "score_configs" ADD CONSTRAINT "score_configs_project_id_fkey" FOREIGN KEY ("project_id") REFERENCES "projects"("id") ON DELETE CASCADE ON UPDATE CASCADE;

packages/shared/prisma/migrations/20240523142524_scores_add_config_id_idx/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "scores_config_id_idx" ON "scores"("config_id");

packages/shared/prisma/migrations/20240523142610_scores_add_fk_scores_config_id/migration.sql

@@ -0,0 +1,2 @@
+-- AddForeignKey
+ALTER TABLE "scores" ADD CONSTRAINT "scores_config_id_fkey" FOREIGN KEY ("config_id") REFERENCES "score_configs"("id") ON DELETE SET NULL ON UPDATE CASCADE;

packages/shared/prisma/migrations/20240524154058_scores_source_enum_add_annotation/migration.sql

@@ -0,0 +1,2 @@
+-- AlterEnum
+ALTER TYPE "ScoreSource" ADD VALUE 'ANNOTATION';

packages/shared/prisma/migrations/20240524156058_scores_source_backfill_annotation_for_review/migration.sql

@@ -0,0 +1,4 @@
+-- Backfill the scores source for 'REVIEW' to be 'ANNOTATION'
+UPDATE "scores"
+SET "source" = 'ANNOTATION'::"ScoreSource"
+WHERE "source" = 'REVIEW'::"ScoreSource";

packages/shared/prisma/migrations/20240524165931_scores_source_enum_drop_review/migration.sql

@@ -0,0 +1,14 @@
+/*
+  Warnings:
+
+  - The values [REVIEW] on the enum `ScoreSource` will be removed. If these variants are still used in the database, this will fail.
+
+*/
+-- AlterEnum
+BEGIN;
+CREATE TYPE "ScoreSource_new" AS ENUM ('ANNOTATION', 'API', 'EVAL');
+ALTER TABLE "scores" ALTER COLUMN "source" TYPE "ScoreSource_new" USING ("source"::text::"ScoreSource_new");
+ALTER TYPE "ScoreSource" RENAME TO "ScoreSource_old";
+ALTER TYPE "ScoreSource_new" RENAME TO "ScoreSource";
+DROP TYPE "ScoreSource_old";
+COMMIT;

packages/shared/prisma/migrations/20240524190433_job_executions_add_fk_index_config_id/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "job_executions_job_configuration_id_idx" ON "job_executions"("job_configuration_id");

packages/shared/prisma/migrations/20240524190434_job_executions_add_fk_index_score_id/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "job_executions_job_output_score_id_idx" ON "job_executions"("job_output_score_id");

packages/shared/prisma/migrations/20240524190435_job_executions_add_fk_index_trace_id/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "job_executions_job_input_trace_id_idx" ON "job_executions"("job_input_trace_id");

packages/shared/prisma/migrations/20240524190436_job_executions_index_created_at/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "job_executions_created_at_idx" ON "job_executions"("created_at");

packages/shared/prisma/migrations/20240528214726_add_cursor_new_columns_observations/migration.sql

@@ -0,0 +1,2 @@
+-- AlterTable
+ALTER TABLE "observations" ADD COLUMN     "updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP;

packages/shared/prisma/migrations/20240528214727_add_cursor_new_columns_scores/migration.sql

@@ -0,0 +1,4 @@
+
+-- AlterTable
+ALTER TABLE "scores" ADD COLUMN     "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+ADD COLUMN     "updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP;

packages/shared/prisma/migrations/20240528214728_add_cursor_index_01/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "audit_logs_updated_at_idx" ON "audit_logs"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_02/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_items_created_at_idx" ON "dataset_items"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_03/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_items_updated_at_idx" ON "dataset_items"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_04/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_run_items_created_at_idx" ON "dataset_run_items"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_05/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_run_items_updated_at_idx" ON "dataset_run_items"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_06/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_runs_created_at_idx" ON "dataset_runs"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_07/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_runs_updated_at_idx" ON "dataset_runs"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_08/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "datasets_created_at_idx" ON "datasets"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_09/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "datasets_updated_at_idx" ON "datasets"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_10/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "job_executions_updated_at_idx" ON "job_executions"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_11/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "observations_updated_at_idx" ON "observations"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_12/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "prompts_created_at_idx" ON "prompts"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_13/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "prompts_updated_at_idx" ON "prompts"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_14/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "score_configs_created_at_idx" ON "score_configs"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_15/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "score_configs_updated_at_idx" ON "score_configs"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_16/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "scores_created_at_idx" ON "scores"("created_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_17/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "scores_updated_at_idx" ON "scores"("updated_at");

packages/shared/prisma/migrations/20240528214728_add_cursor_index_18/migration.sql

@@ -0,0 +1,3 @@
+
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "trace_sessions_updated_at_idx" ON "trace_sessions"("updated_at");

packages/shared/prisma/migrations/20240530212419_model_price_anthropic_via_google_vertex/migration.sql

@@ -0,0 +1,7 @@
+-- Google Vertex uses @ to separate model name and version
+
+UPDATE "public"."models" SET "match_pattern" = '(?i)^(claude-3-haiku(-|@)?20240307)$' WHERE "id" = 'cltr0w45b000008k1407o9qv1';
+
+UPDATE "public"."models" SET "match_pattern" = '(?i)^(claude-3-opus(-|@)?20240229)$' WHERE "id" = 'cltgy0iuw000008le3vod1hhy';
+
+UPDATE "public"."models" SET "match_pattern" = '(?i)^(claude-3-sonnet(-|@)?20240229)$' WHERE "id" = 'cltgy0pp6000108le56se7bl3';

packages/shared/prisma/migrations/20240603212024_dataset_items_add_index_source_trace_id/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX CONCURRENTLY "dataset_items_source_trace_id_idx" ON "dataset_items" USING HASH ("source_trace_id");

packages/shared/prisma/migrations/20240604133338_scores_add_index_name/migration.sql

@@ -0,0 +1,2 @@
+-- CreateIndex
+CREATE INDEX "scores_project_id_name_idx" ON "scores"("project_id", "name");

packages/shared/prisma/migrations/20240604133339_score_data_type_add_boolean/migration.sql

@@ -0,0 +1,2 @@
+-- AlterEnum
+ALTER TYPE "ScoreDataType" ADD VALUE 'BOOLEAN';

packages/shared/prisma/migrations/20240604133340_backfill_manual_scores/migration.sql

@@ -0,0 +1,50 @@
+-- Create manual-score config for any project with manual scores and link config_id to scores
+BEGIN;
+WITH project_configs AS (
+INSERT INTO score_configs (id,
+		project_id,
+		name,
+		data_type,
+		is_archived,
+		min_value,
+		max_value,
+		description)
+SELECT
+	gen_random_uuid () AS id,
+	s.project_id,
+	'manual-score',
+	'NUMERIC',
+	FALSE,
+	- 1,
+	1,
+	'Langfuse legacy annotation score.'
+FROM ( SELECT DISTINCT
+		project_id
+	FROM
+		scores
+	WHERE
+		name = 'manual-score'
+		AND source = 'ANNOTATION') s
+WHERE
+	NOT EXISTS (
+		SELECT
+			1
+		FROM
+			score_configs sc
+		WHERE
+			sc.name = 'manual-score'
+			AND sc.project_id = s.project_id)
+	RETURNING
+		id,
+		project_id
+)
+UPDATE
+	scores
+SET
+	config_id = pc.id
+FROM
+	project_configs pc
+WHERE
+	scores.project_id = pc.project_id
+	AND scores.name = 'manual-score';
+COMMIT;

packages/shared/prisma/migrations/20240606090858_pricings_add_latest_gemini_models/migration.sql

@@ -0,0 +1,5 @@
+-- does not include pricing yet, will be added as soon as it is calculated at ingestion time
+
+INSERT INTO "public"."models" ("id", "model_name", "match_pattern", "unit") VALUES ('clx30djsn0000w9mzebiv41we', 'gemini-1.5-flash', '(?i)^(gemini-1.5-flash)(@[a-zA-Z0-9]+)?$', 'CHARACTERS');
+
+INSERT INTO "public"."models" ("id", "model_name", "match_pattern", "unit") VALUES ('clx30hkrx0000w9mz7lqi0ial', 'gemini-1.5-pro', '(?i)^(gemini-1.5-pro)(@[a-zA-Z0-9]+)?$', 'CHARACTERS');

packages/shared/prisma/migrations/20240606093356_drop_unused_pricings_table/migration.sql

@@ -0,0 +1,14 @@
+/*
+  Warnings:
+
+  - You are about to drop the `pricings` table. If the table is not empty, all the data it contains will be lost.
+
+*/
+-- DropTable
+DROP TABLE "pricings";
+
+-- DropEnum
+DROP TYPE "PricingUnit";
+
+-- DropEnum
+DROP TYPE "TokenType";

packages/shared/prisma/migrations/20240606133011_remove_trace_fkey_datasetrunitems/migration.sql

@@ -0,0 +1,3 @@
+-- DropForeignKey
+ALTER TABLE "dataset_run_items" DROP CONSTRAINT "dataset_run_items_trace_id_fkey";
+ALTER TABLE "dataset_run_items" DROP CONSTRAINT "dataset_run_items_observation_id_fkey";

packages/shared/prisma/schema.prisma

@@ -113,6 +113,8 @@ model Project {
     JobExecution       JobExecution[]
     LlmApiKeys         LlmApiKeys[]
     PosthogIntegration PosthogIntegration[]
+    Score              Score[]
+    scoreConfig        ScoreConfig[]
 
     @@map("projects")
 }
@@ -204,6 +206,7 @@ model TraceSession {
     @@id([id, projectId])
     @@index([projectId])
     @@index([createdAt])
+    @@index([updatedAt])
     @@map("trace_sessions")
 }
 
@@ -230,10 +233,8 @@ model Trace {
     createdAt  DateTime      @default(now()) @map("created_at")
     updatedAt  DateTime      @default(now()) @updatedAt @map("updated_at")
 
-    scores          Score[]
-    DatasetRunItems DatasetRunItems[]
-    DatasetItem     DatasetItem[]
-    JobExecution    JobExecution[]
+    DatasetItem  DatasetItem[]
+    JobExecution JobExecution[]
 
     @@index([projectId])
     @@index([sessionId])
@@ -292,25 +293,24 @@ model Observation {
     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?
-    internalModel       String?           @map("internal_model")
+    internalModel       String?       @map("internal_model")
     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")
+    promptTokens        Int           @default(0) @map("prompt_tokens")
+    completionTokens    Int           @default(0) @map("completion_tokens")
+    totalTokens         Int           @default(0) @map("total_tokens")
     unit                String?
-    inputCost           Decimal?          @map("input_cost")
-    outputCost          Decimal?          @map("output_cost")
-    totalCost           Decimal?          @map("total_cost")
-    completionStartTime DateTime?         @map("completion_start_time")
-    scores              Score[]
-    project             Project           @relation(fields: [projectId], references: [id], onDelete: Cascade)
+    inputCost           Decimal?      @map("input_cost")
+    outputCost          Decimal?      @map("output_cost")
+    totalCost           Decimal?      @map("total_cost")
+    completionStartTime DateTime?     @map("completion_start_time")
+    project             Project       @relation(fields: [projectId], references: [id], onDelete: Cascade)
     derivedDatasetItems DatasetItem[]
-    datasetRunItems     DatasetRunItems[]
 
     promptId String? @map("prompt_id")
     prompt   Prompt? @relation(fields: [promptId], onDelete: SetNull, references: [id])
@@ -324,6 +324,7 @@ model Observation {
     @@index([type])
     @@index(startTime)
     @@index(createdAt)
+    @@index(updatedAt)
     @@index(projectId)
     @@index(parentObservationId)
     @@index(model)
@@ -374,6 +375,7 @@ view ObservationView {
     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")
 }
@@ -394,54 +396,73 @@ enum ObservationLevel {
 model Score {
     id            String         @id @default(cuid())
     timestamp     DateTime       @default(now())
-    projectId     String?        @map("project_id")
+    projectId     String         @map("project_id")
+    project       Project        @relation(fields: [projectId], references: [id], onDelete: Cascade)
     name          String
     value         Float
     source        ScoreSource
+    authorUserId  String?        @map("author_user_id")
     comment       String?
     traceId       String         @map("trace_id")
-    trace         Trace          @relation(fields: [traceId], references: [id], onDelete: Cascade)
     observationId String?        @map("observation_id")
-    observation   Observation?   @relation(fields: [observationId], references: [id], onDelete: SetNull)
+    configId      String?        @map("config_id")
+    stringValue   String?        @map("string_value")
+    createdAt     DateTime       @default(now()) @map("created_at")
+    updatedAt     DateTime       @default(now()) @updatedAt @map("updated_at")
+    dataType      ScoreDataType  @default(NUMERIC) @map("data_type")
     JobExecution  JobExecution[]
+    scoreConfig   ScoreConfig?   @relation(fields: [configId], references: [id], onDelete: SetNull)
 
-    @@unique([id, traceId]) // used for upsert
+    @@unique([id, projectId]) // used for upserts via prisma
     @@index(timestamp)
     @@index([value])
     @@index([projectId])
+    @@index([projectId, name])
+    @@index([authorUserId])
+    @@index([configId])
     @@index([traceId], type: Hash)
     @@index([observationId], type: Hash)
     @@index([source])
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("scores")
 }
 
 enum ScoreSource {
+    ANNOTATION
     API
-    REVIEW
     EVAL
 }
 
-enum PricingUnit {
-    PER_1000_TOKENS
-    PER_1000_CHARS
+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)
+    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")
+    description String?
+    score       Score[]
+
+    @@unique([id, projectId]) // used for upserts via prisma
+    @@index([dataType])
+    @@index([isArchived])
+    @@index([projectId])
+    @@index([categories])
+    @@index([createdAt])
+    @@index([updatedAt])
+    @@map("score_configs")
 }
 
-enum TokenType {
-    PROMPT
-    COMPLETION
-    TOTAL
-}
-
-model Pricing {
-    id          String      @id @default(cuid())
-    modelName   String      @map("model_name")
-    pricingUnit PricingUnit @default(PER_1000_TOKENS) @map("pricing_unit")
-    price       Decimal
-    currency    String      @default("USD")
-    tokenType   TokenType   @map("token_type")
-
-    @@index(modelName)
-    @@map("pricings")
+enum ScoreDataType {
+    CATEGORICAL
+    NUMERIC
+    BOOLEAN
 }
 
 model CronJobs {
@@ -467,6 +488,8 @@ model Dataset {
 
     @@unique([projectId, name])
     @@index([projectId], type: Hash)
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("datasets")
 }
 
@@ -486,8 +509,11 @@ model DatasetItem {
     updatedAt           DateTime          @default(now()) @updatedAt @map("updated_at")
     datasetRunItems     DatasetRunItems[]
 
+    @@index([sourceTraceId], type: Hash)
     @@index([sourceObservationId], type: Hash)
     @@index([datasetId], type: Hash)
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("dataset_items")
 }
 
@@ -509,26 +535,28 @@ model DatasetRuns {
 
     @@unique([datasetId, name])
     @@index([datasetId], type: Hash)
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("dataset_runs")
 }
 
 model DatasetRunItems {
-    id            String       @id @default(cuid())
-    datasetRunId  String       @map("dataset_run_id")
-    datasetRun    DatasetRuns  @relation(fields: [datasetRunId], references: [id], onDelete: Cascade)
-    datasetItemId String       @map("dataset_item_id")
-    datasetItem   DatasetItem  @relation(fields: [datasetItemId], references: [id], onDelete: Cascade)
-    traceId       String       @map("trace_id")
-    trace         Trace        @relation(fields: [traceId], references: [id], onDelete: Cascade)
-    observationId String?      @map("observation_id")
-    observation   Observation? @relation(fields: [observationId], references: [id], onDelete: Cascade)
-    createdAt     DateTime     @default(now()) @map("created_at")
-    updatedAt     DateTime     @default(now()) @updatedAt @map("updated_at")
+    id            String      @id @default(cuid())
+    datasetRunId  String      @map("dataset_run_id")
+    datasetRun    DatasetRuns @relation(fields: [datasetRunId], references: [id], onDelete: Cascade)
+    datasetItemId String      @map("dataset_item_id")
+    datasetItem   DatasetItem @relation(fields: [datasetItemId], references: [id], onDelete: Cascade)
+    traceId       String      @map("trace_id")
+    observationId String?     @map("observation_id")
+    createdAt     DateTime    @default(now()) @map("created_at")
+    updatedAt     DateTime    @default(now()) @updatedAt @map("updated_at")
 
     @@index([datasetRunId], type: Hash)
     @@index([datasetItemId], type: Hash)
     @@index([observationId], type: Hash)
     @@index([traceId])
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("dataset_run_items")
 }
 
@@ -571,6 +599,8 @@ model Prompt {
     @@index([projectId, name, version])
     @@index([projectId, id])
     @@index([projectId])
+    @@index([createdAt])
+    @@index([updatedAt])
     @@index([tags(ops: ArrayOps)], type: Gin)
     @@map("prompts")
 }
@@ -618,6 +648,7 @@ model AuditLog {
 
     @@index([projectId])
     @@index([createdAt])
+    @@index([updatedAt])
     @@map("audit_logs")
 }
 
@@ -711,6 +742,11 @@ model JobExecution {
     @@index([projectId, status])
     @@index([projectId, id])
     @@index([projectId])
+    @@index([jobConfigurationId])
+    @@index([jobOutputScoreId])
+    @@index([jobInputTraceId])
+    @@index([createdAt])
+    @@index([updatedAt])
     @@map("job_executions")
 }
 

packages/shared/prisma/seed.ts

@@ -4,6 +4,7 @@ import {
   type Prisma,
   ObservationType,
   ScoreSource,
+  ScoreDataType,
 } from "../src/index";
 import { hash } from "bcryptjs";
 import { parseArgs } from "node:util";
@@ -12,6 +13,7 @@ import { chunk } from "lodash";
 import { v4 } from "uuid";
 import { ModelUsageUnit } from "../src";
 import { getDisplaySecretKey, hashSecretKey } from "../src/server/auth";
+import { encrypt } from "../src/encryption";
 
 const LOAD_TRACE_VOLUME = 10_000;
 
@@ -38,6 +40,7 @@ async function main() {
       name: "Demo User",
       email: "demo@langfuse.com",
       password: await hash("password", 12),
+      image: "https://static.langfuse.com/langfuse-dev%2Fexample-avatar.png",
     },
   });
 
@@ -143,6 +146,11 @@ async function main() {
       });
     }
 
+    const configIdsAndNames = await generateConfigsForProject([
+      project1,
+      project2,
+    ]);
+
     const promptIds = await generatePromptsForProject([project1, project2]);
 
     const envTags = [null, "development", "staging", "production"];
@@ -156,7 +164,8 @@ async function main() {
       colorTags,
       project1,
       project2,
-      promptIds
+      promptIds,
+      configIdsAndNames
     );
 
     console.log(
@@ -165,6 +174,24 @@ async function main() {
 
     await uploadObjects(traces, observations, scores, sessions, events);
 
+    // If openai key is in environment, add it to the projects LLM API keys
+    const OPENAI_API_KEY = process.env.OPENAI_API_KEY;
+
+    if (OPENAI_API_KEY) {
+      await prisma.llmApiKeys.create({
+        data: {
+          projectId: project1.id,
+          secretKey: encrypt(OPENAI_API_KEY),
+          displaySecretKey: getDisplaySecretKey(OPENAI_API_KEY),
+          provider: "openai",
+        },
+      });
+    } else {
+      console.warn(
+        "No OPENAI_API_KEY found in environment. Skipping seeding LLM API key."
+      );
+    }
+
     // add eval objects
     const evalTemplate = await prisma.evalTemplate.upsert({
       where: {
@@ -434,13 +461,15 @@ function createObjects(
   colorTags: (string | null)[],
   project1: Project,
   project2: Project,
-  promptIds: Map<string, string[]>
+  promptIds: Map<string, string[]>,
+  configIdsAndNames: Map<string, { name: string; id: string }[]>
 ) {
   const traces: Prisma.TraceCreateManyInput[] = [];
   const observations: Prisma.ObservationCreateManyInput[] = [];
   const scores: Prisma.ScoreCreateManyInput[] = [];
   const sessions: Prisma.TraceSessionCreateManyInput[] = [];
   const events: Prisma.ObservationCreateManyInput[] = [];
+  const configs: Prisma.ScoreConfigCreateManyInput[] = [];
 
   for (let i = 0; i < traceVolume; i++) {
     // print progress to console with a progress bar that refreshes every 10 iterations
@@ -459,7 +488,7 @@ function createObjects(
     const session =
       Math.random() > 0.3
         ? {
-            id: `session-${i % 10}`,
+            id: `session-${i % 3}`,
             projectId: projectId,
           }
         : undefined;
@@ -492,16 +521,28 @@ function createObjects(
 
     traces.push(trace);
 
+    const configArray = configIdsAndNames.get(projectId) ?? [];
+    const randomIndex = Math.floor(Math.random() * 3);
+    const config =
+      configArray.length >= randomIndex - 1 && configArray[randomIndex];
+    const { name: annotationScoreName, id: configId } = config || {
+      name: "manual-score",
+      id: undefined,
+    };
+
     const traceScores = [
       ...(Math.random() > 0.5
         ? [
             {
               traceId: trace.id,
-              name: "manual-score",
+              name: annotationScoreName,
               value: Math.floor(Math.random() * 3) - 1,
               timestamp: traceTs,
-              source: ScoreSource.REVIEW,
+              source: ScoreSource.ANNOTATION,
               projectId,
+              authorUserId: `user-${i}`,
+              dataType: ScoreDataType.NUMERIC,
+              ...(configId ? { configId } : {}),
             },
           ]
         : []),
@@ -514,6 +555,7 @@ function createObjects(
               timestamp: traceTs,
               source: ScoreSource.API,
               projectId,
+              dataType: ScoreDataType.NUMERIC,
             },
           ]
         : []),
@@ -528,9 +570,9 @@ function createObjects(
       const spanTsStart = new Date(
         traceTs.getTime() + Math.floor(Math.random() * 30)
       );
-      // random duration of upto 30ms
+      // random duration of upto 5000ms
       const spanTsEnd = new Date(
-        spanTsStart.getTime() + Math.floor(Math.random() * 30)
+        spanTsStart.getTime() + Math.floor(Math.random() * 5000)
       );
 
       const span = {
@@ -574,6 +616,13 @@ function createObjects(
                 (spanTsEnd.getTime() - generationTsStart.getTime())
             )
         );
+        // somewhere in the middle
+        const generationTsCompletionStart = new Date(
+          generationTsStart.getTime() +
+            Math.floor(
+              (generationTsEnd.getTime() - generationTsStart.getTime()) / 3
+            )
+        );
 
         const promptTokens = Math.floor(Math.random() * 1000) + 300;
         const completionTokens = Math.floor(Math.random() * 500) + 100;
@@ -602,6 +651,8 @@ function createObjects(
           id: `generation-${v4()}`,
           startTime: generationTsStart,
           endTime: generationTsEnd,
+          completionStartTime:
+            Math.random() > 0.5 ? generationTsCompletionStart : undefined,
           name: `generation-${i}-${j}-${k}`,
           projectId: trace.projectId,
           promptId: promptId,
@@ -743,6 +794,7 @@ function createObjects(
     traces,
     observations,
     scores,
+    configs,
     sessions: uniqueSessions,
     events,
   };
@@ -925,3 +977,80 @@ async function generatePrompts(project: Project) {
   }
   return promptIds;
 }
+
+async function generateConfigsForProject(projects: Project[]) {
+  const projectIdsToConfigs: Map<string, { name: string; id: string }[]> =
+    new Map();
+
+  await Promise.all(
+    projects.map(async (project) => {
+      const configNameAndId = await generateConfigs(project);
+      projectIdsToConfigs.set(project.id, configNameAndId);
+    })
+  );
+  return projectIdsToConfigs;
+}
+
+async function generateConfigs(project: Project) {
+  const configNameAndId: { name: string; id: string }[] = [];
+
+  const configs = [
+    {
+      id: `config-${v4()}`,
+      name: "manual-score",
+      dataType: ScoreDataType.NUMERIC,
+      projectId: project.id,
+      isArchived: false,
+    },
+    {
+      id: `config-${v4()}`,
+      projectId: project.id,
+      name: "Accuracy",
+      dataType: ScoreDataType.CATEGORICAL,
+      categories: [
+        { label: "Incorrect", value: 0 },
+        { label: "Partially Correct", value: 1 },
+        { label: "Correct", value: 2 },
+      ],
+      isArchived: false,
+    },
+    {
+      id: `config-${v4()}`,
+      projectId: project.id,
+      name: "Toxicity",
+      dataType: ScoreDataType.BOOLEAN,
+      categories: [
+        { label: "True", value: 1 },
+        { label: "False", value: 0 },
+      ],
+      description:
+        "Used to indicate if text was harmful or offensive in nature.",
+      isArchived: false,
+    },
+  ];
+
+  for (const config of configs) {
+    await prisma.scoreConfig.upsert({
+      where: {
+        id_projectId: {
+          projectId: config.projectId,
+          id: config.id,
+        },
+      },
+      create: {
+        id: config.id,
+        projectId: config.projectId,
+        name: config.name,
+        dataType: config.dataType,
+        categories: config.categories,
+        isArchived: config.isArchived,
+      },
+      update: {
+        id: config.id,
+      },
+    });
+    configNameAndId.push({ name: config.name, id: config.id });
+  }
+
+  return configNameAndId;
+}

packages/shared/src/features/evals/types.ts

@@ -1,5 +1,5 @@
 import z from "zod";
-import { ModelProvider } from "../..";
+import { ModelProvider, UIModelParams } from "../..";
 
 export const langfuseObjects = [
   "trace",
@@ -77,30 +77,30 @@ export const availableEvalVariables = [
   },
 ];
 
-export const evalLLMModels = [
+export const evalLLMModels: UIModelParams[] = [
   {
-    provider: ModelProvider.OpenAI,
-    model: "gpt-3.5-turbo",
-    temperature: 1,
-    maxTemperature: 2,
-    max_tokens: 256,
-    top_p: 1,
+    provider: { value: ModelProvider.OpenAI, enabled: true },
+    model: { value: "gpt-3.5-turbo", enabled: true },
+    temperature: { value: 1, enabled: true },
+    maxTemperature: { value: 2, enabled: true },
+    max_tokens: { value: 256, enabled: true },
+    top_p: { value: 1, enabled: true },
   },
   {
-    provider: ModelProvider.OpenAI,
-    model: "gpt-4-turbo-preview",
-    temperature: 1,
-    maxTemperature: 2,
-    max_tokens: 256,
-    top_p: 1,
-  } as const,
+    provider: { value: ModelProvider.OpenAI, enabled: true },
+    model: { value: "gpt-4-turbo-preview", enabled: true },
+    temperature: { value: 1, enabled: true },
+    maxTemperature: { value: 2, enabled: true },
+    max_tokens: { value: 256, enabled: true },
+    top_p: { value: 1, enabled: true },
+  },
   {
-    provider: ModelProvider.OpenAI,
-    model: "gpt-4o",
-    temperature: 1,
-    maxTemperature: 2,
-    max_tokens: 256,
-    top_p: 1,
+    provider: { value: ModelProvider.OpenAI, enabled: true },
+    model: { value: "gpt-4o", enabled: true },
+    temperature: { value: 1, enabled: true },
+    maxTemperature: { value: 2, enabled: true },
+    max_tokens: { value: 256, enabled: true },
+    top_p: { value: 1, enabled: true },
   },
 ] as const;
 

packages/shared/src/features/ingestion/types.ts

@@ -1,9 +1,9 @@
 import lodash from "lodash";
 import { z } from "zod";
 
-import { ModelUsageUnit } from "@langfuse/shared";
-import { NonEmptyString, jsonSchema } from "@/src/utils/zod";
-import { ObservationLevel } from "@langfuse/shared";
+import { NonEmptyString, jsonSchema } from "../../utils/zod";
+import { ModelUsageUnit } from "../../constants";
+import { ObservationLevel } from "@prisma/client";
 
 export const Usage = z.object({
   input: z.number().int().nullish(),
@@ -109,7 +109,7 @@ export const CreateGenerationBody = CreateSpanBody.extend({
       z.string(),
       z
         .union([z.string(), z.number(), z.boolean(), z.array(z.string())])
-        .nullish(),
+        .nullish()
     )
     .nullish(),
   usage: usage,
@@ -131,7 +131,7 @@ export const UpdateGenerationBody = UpdateSpanBody.extend({
       z.string(),
       z
         .union([z.string(), z.number(), z.boolean(), z.array(z.string())])
-        .nullish(),
+        .nullish()
     )
     .nullish(),
   usage: usage,
@@ -195,7 +195,7 @@ export const LegacyGenerationsCreateSchema = z.object({
   modelParameters: z
     .record(
       z.string(),
-      z.union([z.string(), z.number(), z.boolean()]).nullish(),
+      z.union([z.string(), z.number(), z.boolean()]).nullish()
     )
     .nullish(),
   prompt: jsonSchema.nullish(),
@@ -219,7 +219,7 @@ export const LegacyGenerationPatchSchema = z.object({
   modelParameters: z
     .record(
       z.string(),
-      z.union([z.string(), z.number(), z.boolean()]).nullish(),
+      z.union([z.string(), z.number(), z.boolean()]).nullish()
     )
     .nullish(),
   prompt: jsonSchema.nullish(),
@@ -243,7 +243,7 @@ export const LegacyObservationBody = z.object({
   modelParameters: z
     .record(
       z.string(),
-      z.union([z.string(), z.number(), z.boolean()]).nullish(),
+      z.union([z.string(), z.number(), z.boolean()]).nullish()
     )
     .nullish(),
   input: jsonSchema.nullish(),

packages/shared/src/filterToPrisma.ts

@@ -1,7 +1,7 @@
 import { Prisma } from "@prisma/client";
 import {
   ColumnDefinition,
-  type TableNames as TableName,
+  type TableNames,
 } from "./interfaces/tableDefinition";
 import { FilterState } from "./types";
 import { filterOperators } from "./interfaces/filters";
@@ -27,7 +27,7 @@ const arrayOperatorReplacements = {
 export function tableColumnsToSqlFilterAndPrefix(
   filters: FilterState,
   tableColumns: ColumnDefinition[],
-  table: TableName
+  table: TableNames
 ): Prisma.Sql {
   const sql = tableColumnsToSqlFilter(filters, tableColumns, table);
   if (sql === Prisma.empty) {
@@ -43,7 +43,7 @@ export function tableColumnsToSqlFilterAndPrefix(
 export function tableColumnsToSqlFilter(
   filters: FilterState,
   tableColumns: ColumnDefinition[],
-  table: TableName
+  table: TableNames
 ): Prisma.Sql {
   const internalFilters = filters.map((filter) => {
     // Get column definition to map column to internal name, e.g. "t.id"
@@ -153,7 +153,7 @@ export function tableColumnsToSqlFilter(
 
 const castValueToPostgresTypes = (
   column: ColumnDefinition,
-  table: TableName
+  table: TableNames
 ) => {
   return column.name === "type" &&
     (table === "observations" ||

packages/shared/src/index.ts

@@ -9,6 +9,9 @@ export * from "./filterToPrisma";
 export * from "./tracesTable";
 export * from "./server/auth";
 export * from "./observationsTable";
+export * from "./features/ingestion/types";
+export * from "./utils/zod";
+export * from "./utils/json";
 export { env } from "./env";
 
 // llm api

packages/shared/src/interfaces/tableDefinition.ts

@@ -42,6 +42,7 @@ export const tableNames = [
   "traces_parent_observation_scores",
   "sessions",
   "prompts",
+  "users",
 ] as const;
 
 export type TableNames = (typeof tableNames)[number];

packages/shared/src/server/llm/types.ts

@@ -20,19 +20,23 @@ export enum ChatMessageRole {
   Assistant = "assistant",
 }
 export type ModelParams = AnthropicModelParams | OpenAIModelParams;
-export type UIModelParams = Required<
-  AnthropicModelParams | OpenAIModelParams
-> & {
-  maxTemperature: number;
+
+type RecordWithEnabledFlag<T> = {
+  [K in keyof T]: { value: T[K]; enabled: boolean };
 };
+export type UIModelParams = RecordWithEnabledFlag<
+  Required<AnthropicModelParams | OpenAIModelParams> & {
+    maxTemperature: number;
+  }
+>;
 
 // Generic config
 export type ModelConfig = z.infer<typeof ZodModelConfig>;
 
 export const ZodModelConfig = z.object({
-  max_tokens: z.number().optional(),
-  temperature: z.number().optional(),
-  top_p: z.number().optional(),
+  max_tokens: z.coerce.number().optional(),
+  temperature: z.coerce.number().optional(),
+  top_p: z.coerce.number().optional(),
 });
 
 // OpenAI

packages/shared/src/types.ts

@@ -30,4 +30,5 @@ export type TableName =
   | "sessions"
   | "scores"
   | "prompts"
-  | "dashboard";
+  | "dashboard"
+  | "users";

packages/shared/src/utils/json.ts

@@ -1,6 +1,6 @@
-import { type jsonSchema } from "@/src/utils/zod";
 import { type z } from "zod";
 import lodash from "lodash";
+import { jsonSchema } from "./zod";
 
 export const parseJson = (input: string) => {
   try {
@@ -37,7 +37,7 @@ export function deepParseJson(json: unknown): unknown {
         // Ensure we only iterate over the object's own properties
         if (Object.prototype.hasOwnProperty.call(json, key)) {
           (json as Record<string, unknown>)[key] = deepParseJson(
-            (json as Record<string, unknown>)[key],
+            (json as Record<string, unknown>)[key]
           );
         }
       }
@@ -50,7 +50,7 @@ export function deepParseJson(json: unknown): unknown {
 
 export const mergeJson = (
   json1?: z.infer<typeof jsonSchema>,
-  json2?: z.infer<typeof jsonSchema>,
+  json2?: z.infer<typeof jsonSchema>
 ) => {
   if (json1 === undefined) {
     return json2;

packages/shared/src/utils/zod.ts

@@ -26,7 +26,7 @@ const jsonSchemaNullable: z.ZodType<JsonNested> = z.lazy(() =>
     nestedLiteralSchema,
     z.array(jsonSchemaNullable),
     z.record(jsonSchemaNullable),
-  ]),
+  ])
 );
 
 // Root schema that does not allow nulls at the root level
@@ -35,17 +35,17 @@ export const jsonSchema: z.ZodType<Json> = z.lazy(() =>
     rootLiteralSchema,
     z.array(jsonSchemaNullable),
     z.record(jsonSchemaNullable),
-  ]),
+  ])
 );
 
 export const paginationZod = {
   page: z.preprocess(
     (x) => (x === "" ? undefined : x),
-    z.coerce.number().default(1),
+    z.coerce.number().default(1)
   ),
   limit: z.preprocess(
     (x) => (x === "" ? undefined : x),
-    z.coerce.number().lte(100).default(50),
+    z.coerce.number().lte(100).default(50)
   ),
 };
 

web/Dockerfile

@@ -5,7 +5,7 @@ FROM node:20-alpine AS alpine
 RUN apk update && apk upgrade --no-cache libcrypto3 libssl3 libc6-compat
 
 FROM alpine AS base
-RUN npm install turbo --global
+RUN npm install turbo@^1.13.3 --global
 ENV PNPM_HOME="/pnpm"
 ENV PATH="$PNPM_HOME:$PATH"
 RUN corepack enable

web/package.json

@@ -1,6 +1,6 @@
 {
   "name": "web",
-  "version": "2.41.1",
+  "version": "2.49.2",
   "private": true,
   "license": "MIT",
   "engines": {
@@ -25,12 +25,14 @@
     "@aws-sdk/lib-storage": "^3.568.0",
     "@aws-sdk/s3-request-presigner": "^3.554.0",
     "@codemirror/lang-json": "^6.0.1",
-    "@headlessui/react": "^1.7.19",
+    "@codemirror/lint": "^6.8.0",
+    "@headlessui/react": "1.7.18",
+    "@headlessui/tailwindcss": "0.2.0",
     "@heroicons/react": "^2.1.3",
     "@hookform/resolvers": "^3.3.4",
-    "@langchain/anthropic": "^0.1.16",
+    "@langchain/anthropic": "^0.1.21",
     "@langchain/core": "^0.1.61",
-    "@langchain/openai": "^0.0.28",
+    "@langchain/openai": "^0.0.33",
     "@langfuse/ee": "workspace:*",
     "@langfuse/shared": "workspace:*",
     "@marsidev/react-turnstile": "^0.5.4",
@@ -54,26 +56,28 @@
     "@radix-ui/react-switch": "^1.0.3",
     "@radix-ui/react-tabs": "^1.0.4",
     "@radix-ui/react-toggle": "^1.0.3",
+    "@radix-ui/react-toggle-group": "^1.0.4",
     "@radix-ui/react-tooltip": "^1.0.7",
-    "@react-email/components": "^0.0.17",
-    "@react-email/render": "^0.0.13",
+    "@react-email/components": "^0.0.18",
+    "@react-email/render": "^0.0.14",
+    "@remixicon/react": "^4.2.0",
     "@repo/eslint-config": "workspace:*",
     "@repo/typescript-config": "workspace:*",
     "@sentry/nextjs": "^7.113.0",
     "@sentry/node": "^7.113.0",
     "@sentry/profiling-node": "^7.113.0",
     "@sentry/types": "^7.113.0",
-    "@t3-oss/env-nextjs": "^0.8.0",
+    "@t3-oss/env-nextjs": "^0.10.1",
     "@tailwindcss/container-queries": "^0.1.1",
-    "@tailwindcss/forms": "^0.5.7",
     "@tanstack/react-query": "^4.36.1",
     "@tanstack/react-table": "^8.11.8",
-    "@tremor/react": "3.11.1",
+    "@tremor/react": "3.16.2",
     "@trpc/client": "^10.45.0",
     "@trpc/next": "^10.45.0",
     "@trpc/react-query": "^10.45.0",
     "@trpc/server": "^10.45.0",
     "@uiw/codemirror-theme-github": "^4.21.25",
+    "@uiw/codemirror-theme-tokyo-night": "^4.22.1",
     "@uiw/react-codemirror": "^4.21.25",
     "ai": "^3.0.23",
     "bcryptjs": "^2.4.3",
@@ -84,7 +88,7 @@
     "cors": "^2.8.5",
     "date-fns": "^3.3.1",
     "decimal.js": "^10.4.3",
-    "dompurify": "^3.1.2",
+    "dompurify": "^3.1.4",
     "exponential-backoff": "^3.1.1",
     "js-tiktoken": "^1.0.12",
     "kysely": "^0.27.3",
@@ -94,6 +98,7 @@
     "next": "^14.2.3",
     "next-auth": "^4.24.7",
     "next-query-params": "^5.0.0",
+    "next-themes": "^0.3.0",
     "nodemailer": "^6.9.13",
     "posthog-js": "^1.122.0",
     "posthog-node": "^3.6.3",
@@ -101,8 +106,8 @@
     "react": "18.2.0",
     "react-day-picker": "^8.10.1",
     "react-dom": "18.2.0",
-    "react-hook-form": "^7.51.3",
-    "react-icons": "^5.0.1",
+    "react-hook-form": "^7.51.5",
+    "react-icons": "^5.2.1",
     "react-responsive": "^10.0.0",
     "react18-json-view": "^0.2.8-canary.6",
     "sonner": "^1.4.41",
@@ -111,29 +116,31 @@
     "tailwindcss-animate": "^1.0.7",
     "use-query-params": "^2.2.1",
     "uuid": "^9.0.1",
+    "vaul": "^0.9.1",
     "zod": "^3.22.4"
   },
   "devDependencies": {
     "@jedmao/location": "^3.0.0",
     "@mermaid-js/mermaid-cli": "^10.7.0",
     "@playwright/test": "^1.43.1",
-    "@testing-library/jest-dom": "^6.4.2",
-    "@testing-library/react": "^15.0.6",
+    "@tailwindcss/forms": "^0.5.7",
+    "@testing-library/jest-dom": "^6.4.5",
+    "@testing-library/react": "^15.0.7",
     "@types/bcryptjs": "^2.4.6",
     "@types/cors": "^2.8.17",
     "@types/dompurify": "^3.0.5",
     "@types/eslint": "^8.56.7",
     "@types/jest": "^29.5.12",
-    "@types/lodash": "^4.14.202",
+    "@types/lodash": "^4.17.4",
     "@types/node": "20.10.5",
-    "@types/nodemailer": "^6.4.14",
+    "@types/nodemailer": "^6.4.15",
     "@types/react": "^18.2.79",
     "@types/react-dom": "^18.2.25",
     "@types/uuid": "^9.0.8",
     "@typescript-eslint/eslint-plugin": "^6.21.0",
-    "@typescript-eslint/parser": "^7.7.0",
+    "@typescript-eslint/parser": "^7.12.0",
     "autoprefixer": "^10.4.19",
-    "dotenv-cli": "^7.3.0",
+    "dotenv-cli": "^7.4.2",
     "eslint": "^8.56.0",
     "eslint-config-next": "^14.2.3",
     "jest": "^29.7.0",

web/public/generated/api/openapi.yml

@@ -1137,7 +1137,10 @@ paths:
       security: *ref_0
   /api/public/sessions/{sessionId}:
     get:
-      description: Get a session
+      description: >-
+        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>`
       operationId: sessions_get
       tags:
         - Sessions
@@ -1262,6 +1265,12 @@ paths:
           schema:
             type: string
             nullable: true
+        - name: sessionId
+          in: query
+          required: false
+          schema:
+            type: string
+            nullable: true
         - name: fromTimestamp
           in: query
           description: Retrieve only traces newer than this datetime (ISO 8601).
@@ -1825,8 +1834,8 @@ components:
       title: ScoreSource
       type: string
       enum:
+        - ANNOTATION
         - API
-        - REVIEW
         - EVAL
     CreateDatasetItemRequest:
       title: CreateDatasetItemRequest
@@ -2499,6 +2508,7 @@ components:
         totalCost:
           type: number
           format: double
+          description: Total model cost in USD
         usage:
           type: array
           items:
@@ -2521,14 +2531,28 @@ components:
           nullable: true
         inputUsage:
           type: integer
+          description: Total number of generation input units (e.g. tokens)
         outputUsage:
           type: integer
+          description: Total number of generation output units (e.g. tokens)
         totalUsage:
           type: integer
+          description: Total number of generation total units (e.g. tokens)
+        countTraces:
+          type: integer
+        countObservations:
+          type: integer
+        totalCost:
+          type: number
+          format: double
+          description: Total model cost in USD
       required:
         - inputUsage
         - outputUsage
         - totalUsage
+        - countTraces
+        - countObservations
+        - totalCost
     Observations:
       title: Observations
       type: object
@@ -2654,6 +2678,13 @@ components:
           items:
             type: string
           nullable: true
+          description: List of deployment labels of this prompt version.
+        tags:
+          type: array
+          items:
+            type: string
+          nullable: true
+          description: List of tags to apply to all versions of this prompt.
       required:
         - name
         - prompt
@@ -2672,6 +2703,13 @@ components:
           items:
             type: string
           nullable: true
+          description: List of deployment labels of this prompt version.
+        tags:
+          type: array
+          items:
+            type: string
+          nullable: true
+          description: List of tags to apply to all versions of this prompt.
       required:
         - name
         - prompt
@@ -2713,11 +2751,20 @@ components:
           type: array
           items:
             type: string
+          description: List of deployment labels of this prompt version.
+        tags:
+          type: array
+          items:
+            type: string
+          description: >-
+            List of tags. Used to filter via UI and API. The same across
+            versions of a prompt.
       required:
         - name
         - version
         - config
         - labels
+        - tags
     ChatMessage:
       title: ChatMessage
       type: object

web/public/generated/postman/collection.json

@@ -699,7 +699,7 @@
             "auth": null,
             "body": {
               "mode": "raw",
-              "raw": "{\n    \"type\": \"chat\",\n    \"name\": \"example\",\n    \"prompt\": [\n        {\n            \"role\": \"example\",\n            \"content\": \"example\"\n        }\n    ],\n    \"config\": \"UNKNOWN\",\n    \"labels\": [\n        \"example\"\n    ]\n}",
+              "raw": "{\n    \"type\": \"chat\",\n    \"name\": \"example\",\n    \"prompt\": [\n        {\n            \"role\": \"example\",\n            \"content\": \"example\"\n        }\n    ],\n    \"config\": \"UNKNOWN\",\n    \"labels\": [\n        \"example\"\n    ],\n    \"tags\": [\n        \"example\"\n    ]\n}",
               "options": {
                 "raw": {
                   "language": "json"
@@ -890,7 +890,7 @@
           "_type": "endpoint",
           "name": "Get",
           "request": {
-            "description": "Get a session",
+            "description": "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>`",
             "url": {
               "raw": "{{baseUrl}}/api/public/sessions/:sessionId",
               "host": [
@@ -963,7 +963,7 @@
           "request": {
             "description": "Get list of traces.",
             "url": {
-              "raw": "{{baseUrl}}/api/public/traces?page=&limit=&userId=&name=&fromTimestamp=&orderBy=&tags=",
+              "raw": "{{baseUrl}}/api/public/traces?page=&limit=&userId=&name=&sessionId=&fromTimestamp=&orderBy=&tags=",
               "host": [
                 "{{baseUrl}}"
               ],
@@ -993,6 +993,11 @@
                   "value": "",
                   "description": null
                 },
+                {
+                  "key": "sessionId",
+                  "value": "",
+                  "description": null
+                },
                 {
                   "key": "fromTimestamp",
                   "value": "",

web/sentry.server.config.ts

@@ -2,7 +2,7 @@ import { prisma } from "@langfuse/shared/src/db";
 import * as Sentry from "@sentry/nextjs";
 import { ProfilingIntegration } from "@sentry/profiling-node";
 import type { SamplingContext, TransactionEvent } from "@sentry/types";
-import { jsonSchema } from "@/src/utils/zod";
+import { jsonSchema } from "@langfuse/shared";
 
 if (process.env.NEXT_PUBLIC_SENTRY_DSN)
   Sentry.init({
@@ -18,12 +18,21 @@ if (process.env.NEXT_PUBLIC_SENTRY_DSN)
         samplingContext.request.url &&
         samplingContext.request.url.includes("api/trpc")
       ) {
-        return 0.3;
+        return 0.1;
       }
-      return 0.1;
+      if (
+        samplingContext.request &&
+        samplingContext.request.url &&
+        samplingContext.request.url.includes("api/auth") &&
+        samplingContext.transactionContext.status !== "ok" &&
+        samplingContext.transactionContext.status !== "unauthenticated"
+      ) {
+        return 0.1;
+      }
+      return 0.01;
     },
 
-    profilesSampleRate: 0.2, // Profiling sample rate is relative to tracesSampleRate
+    profilesSampleRate: 0.1,
     integrations: [
       // Add profiling integration to list of integrations
       new ProfilingIntegration(),

web/src/__tests__/ingestion.servertest.ts

@@ -281,6 +281,7 @@ describe("/api/public/ingestion API Endpoint", () => {
       expect(dbScore?.name).toBe("score-name");
       expect(dbScore?.value).toBe(100.5);
       expect(dbScore?.observationId).toBeNull();
+      expect(dbScore?.projectId).toBe("7a88fb47-b4e2-43b8-a06c-a5ce950dc53a");
     });
   });
 
@@ -726,6 +727,7 @@ describe("/api/public/ingestion API Endpoint", () => {
             metadata: { key: "value" },
             release: "1.0.0",
             version: "2.0.0",
+            tags: ["tag-1", "tag-2", "tag-2"],
           },
         },
       ],
@@ -742,6 +744,7 @@ describe("/api/public/ingestion API Endpoint", () => {
             id: traceId,
             name: "trace-name",
             userId: "user-2",
+            tags: ["tag-1", "tag-4", "tag-3"],
           },
         },
       ],
@@ -762,6 +765,8 @@ describe("/api/public/ingestion API Endpoint", () => {
     expect(dbTrace[0]?.externalId).toBeNull();
     expect(dbTrace[0]?.version).toBe("2.0.0");
     expect(dbTrace[0]?.projectId).toBe("7a88fb47-b4e2-43b8-a06c-a5ce950dc53a");
+    expect(dbTrace[0]?.tags).toEqual(["tag-1", "tag-2", "tag-3", "tag-4"]);
+    expect(dbTrace[0]?.tags.length).toBe(4);
   });
 
   it("should fail for wrong event formats", async () => {
@@ -816,57 +821,6 @@ describe("/api/public/ingestion API Endpoint", () => {
     expect(dbTrace.length).toBe(1);
   });
 
-  it("should fail for auth errors", async () => {
-    const traceId = v4();
-    const scoreId = v4();
-
-    const responseOne = await makeAPICall("POST", "/api/public/ingestion", {
-      batch: [
-        {
-          id: v4(),
-          type: "trace-create",
-          timestamp: new Date().toISOString(),
-          body: {
-            id: traceId,
-            name: "trace-name",
-            userId: "user-1",
-            metadata: { key: "value" },
-            release: "1.0.0",
-            version: "2.0.0",
-          },
-        },
-        {
-          id: v4(),
-          type: "score-create",
-          timestamp: new Date().toISOString(),
-          body: {
-            id: scoreId,
-            name: "score-name",
-            value: 100.5,
-            traceId: "some-random-id",
-          },
-        },
-      ],
-    });
-
-    expect(responseOne.status).toBe(207);
-
-    expect("errors" in responseOne.body).toBe(true);
-    // eslint-disable-next-line @typescript-eslint/no-unsafe-member-access
-    expect(responseOne.body.errors.length).toBe(1);
-    expect("successes" in responseOne.body).toBe(true);
-    // eslint-disable-next-line @typescript-eslint/no-unsafe-member-access
-    expect(responseOne.body.successes.length).toBe(1);
-
-    const dbTrace = await prisma.trace.findMany({
-      where: {
-        name: "trace-name",
-      },
-    });
-
-    expect(dbTrace.length).toBe(1);
-  });
-
   it("should fail for resource not found", async () => {
     const traceId = v4();
 
@@ -1653,121 +1607,6 @@ IB Home   /   . . .   /   News   /   News about the IB   /   Why ChatGPT is an o
     expect(dbScore?.value).toBe(scoreValue);
   });
 
-  it("should error on Bearer Auth for a trace from different project", async () => {
-    const otherProjectId = "other_project_id";
-    const traceId = "trace_id";
-    const bearerAuth = "Bearer pk-lf-1234567890";
-
-    const scoreId = "score_id";
-    const scoreEventId = "score_event_id";
-    const scoreName = "score-name";
-    const scoreValue = 100.5;
-
-    // Seed db with a trace to be scored
-    try {
-      await prisma.project.create({
-        data: {
-          id: otherProjectId,
-          name: "another-project",
-        },
-      });
-
-      await prisma.trace.create({
-        data: {
-          id: traceId,
-          name: "trace-name",
-          project: { connect: { id: otherProjectId } },
-        },
-      });
-
-      const response = await makeAPICall(
-        "POST",
-        "/api/public/ingestion",
-        {
-          batch: [
-            {
-              id: scoreEventId,
-              type: "score-create",
-              timestamp: new Date().toISOString(),
-              body: {
-                id: scoreId,
-                name: scoreName,
-                value: scoreValue,
-                traceId: traceId,
-              },
-            },
-          ],
-        },
-        bearerAuth,
-      );
-
-      expect(response.status).toBe(207);
-      expect(response.body.successes.length).toBe(0);
-      expect(response.body).toHaveProperty("errors");
-      expect(response.body.errors.length).toEqual(1);
-      expect(response.body.errors[0]?.id).toEqual(scoreEventId);
-
-      expect(await prisma.trace.count()).toBe(1);
-      expect(await prisma.trace.count({ where: { id: traceId } })).toBe(1);
-
-      const dbScore = await prisma.score.findUnique({
-        where: {
-          id: scoreId,
-        },
-      });
-
-      expect(dbScore).toBeNull();
-    } finally {
-      await prisma.project.delete({ where: { id: otherProjectId } });
-    }
-  });
-
-  it("should error on Bearer Auth with a trace that does not exist", async () => {
-    const traceId = "trace_id";
-    const bearerAuth = "Bearer pk-lf-1234567890";
-
-    const scoreId = "score_id";
-    const scoreEventId = "score_event_id";
-    const scoreName = "score-name";
-    const scoreValue = 100.5;
-
-    const response = await makeAPICall(
-      "POST",
-      "/api/public/ingestion",
-      {
-        batch: [
-          {
-            id: scoreEventId,
-            type: "score-create",
-            timestamp: new Date().toISOString(),
-            body: {
-              id: scoreId,
-              name: scoreName,
-              value: scoreValue,
-              traceId: traceId,
-            },
-          },
-        ],
-      },
-      bearerAuth,
-    );
-
-    expect(response.status).toBe(207);
-    expect(response.body.successes.length).toBe(0);
-    expect(response.body.errors.length).toBe(1);
-    expect(response.body.errors[0]?.id).toBe(scoreEventId);
-
-    expect(await prisma.trace.count()).toBe(0);
-
-    const dbScore = await prisma.score.findUnique({
-      where: {
-        id: scoreId,
-      },
-    });
-
-    expect(dbScore).toBeNull();
-  });
-
   it("should error on wrong input", async () => {
     const traceId = "trace_id";
     const bearerAuth = "Bearer pk-lf-1234567890";

web/src/__tests__/metrics-api.servertest.ts

@@ -68,12 +68,18 @@ describe("/api/public/metrics/daily API Endpoint", () => {
         inputUsage: 333,
         outputUsage: 0,
         totalUsage: 333,
+        countObservations: 1,
+        countTraces: 1,
+        totalCost: 0,
       },
       {
         model: "modelC",
         inputUsage: 666,
         outputUsage: 777,
         totalUsage: 1443,
+        countObservations: 1,
+        countTraces: 1,
+        totalCost: 1024.22,
       },
     ]);
 
@@ -88,6 +94,9 @@ describe("/api/public/metrics/daily API Endpoint", () => {
         inputUsage: 100,
         outputUsage: 200,
         totalUsage: 300,
+        countObservations: 1,
+        countTraces: 1,
+        totalCost: 0,
       },
     ]);
   });

web/src/__tests__/prompts.v2.servertest.ts

@@ -680,6 +680,71 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(fetchedPrompt.body.createdBy).toBe("API");
       expect(fetchedPrompt.body.config).toEqual({});
     });
+
+    it("should update tags across versions", async () => {
+      const promptName = "prompt-name" + nanoid();
+
+      const createPromptVersion = async (tags?: string[]) => {
+        await makeAPICall("POST", baseURI, {
+          name: promptName,
+          prompt: "This is a test prompt",
+          type: PromptType.Text,
+          ...(tags !== undefined && { tags: tags }),
+        });
+      };
+
+      const fetchPromptVersion = async (version: number) => {
+        const fetchedPrompt = await makeAPICall(
+          "GET",
+          `${baseURI}/${promptName}?version=${version}`,
+          undefined,
+        );
+        expect(fetchedPrompt.status).toBe(200);
+        if (!isPrompt(fetchedPrompt.body)) {
+          throw new Error("Expected body to be a prompt");
+        }
+        return fetchedPrompt.body;
+      };
+
+      // Create version 1 with ["tag"]
+      await createPromptVersion(["tag"]);
+      let fetchedPrompt1 = await fetchPromptVersion(1);
+      expect(fetchedPrompt1.tags).toEqual(["tag"]);
+      expect(fetchedPrompt1.version).toBe(1);
+
+      // Create version 2 with no tags provided (should use tags from version 1)
+      await createPromptVersion();
+      let fetchedPrompt2 = await fetchPromptVersion(2);
+      expect(fetchedPrompt2.tags).toEqual(["tag"]);
+      expect(fetchedPrompt2.version).toBe(2);
+
+      // Create version 3 with ["tag1", "tag2", "tag3"] (should update tags across versions)
+      await createPromptVersion(["tag1", "tag2", "tag3"]);
+      fetchedPrompt1 = await fetchPromptVersion(1);
+      fetchedPrompt2 = await fetchPromptVersion(2);
+      let fetchedPrompt3 = await fetchPromptVersion(3);
+      expect(fetchedPrompt1.tags).toEqual(["tag1", "tag2", "tag3"]);
+      expect(fetchedPrompt1.version).toBe(1);
+      expect(fetchedPrompt2.tags).toEqual(["tag1", "tag2", "tag3"]);
+      expect(fetchedPrompt2.version).toBe(2);
+      expect(fetchedPrompt3.tags).toEqual(["tag1", "tag2", "tag3"]);
+      expect(fetchedPrompt3.version).toBe(3);
+
+      // remove tags
+      await createPromptVersion([]);
+      fetchedPrompt1 = await fetchPromptVersion(1);
+      fetchedPrompt2 = await fetchPromptVersion(2);
+      fetchedPrompt3 = await fetchPromptVersion(3);
+      let fetchedPrompt4 = await fetchPromptVersion(4);
+      expect(fetchedPrompt1.tags).toEqual([]);
+      expect(fetchedPrompt1.version).toBe(1);
+      expect(fetchedPrompt2.tags).toEqual([]);
+      expect(fetchedPrompt2.version).toBe(2);
+      expect(fetchedPrompt3.tags).toEqual([]);
+      expect(fetchedPrompt3.version).toBe(3);
+      expect(fetchedPrompt4.tags).toEqual([]);
+      expect(fetchedPrompt4.version).toBe(4);
+    });
   });
 
   describe("when fetching a prompt list", () => {
@@ -784,10 +849,17 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(promptMeta3.tags).toEqual(["tag-1"]);
 
       // Validate pagination
-      expect(body.pagination.page).toBe(1);
-      expect(body.pagination.limit).toBe(10);
-      expect(body.pagination.totalPages).toBe(1);
-      expect(body.pagination.totalItems).toBe(3);
+      expect(body.meta.page).toBe(1);
+      expect(body.meta.limit).toBe(10);
+      expect(body.meta.totalPages).toBe(1);
+      expect(body.meta.totalItems).toBe(3);
+
+      // Validate pagination backwards compatibility
+      // https://github.com/langfuse/langfuse/issues/2068
+      expect(body.pagination?.page).toBe(1);
+      expect(body.pagination?.limit).toBe(10);
+      expect(body.pagination?.totalPages).toBe(1);
+      expect(body.pagination?.totalItems).toBe(3);
     });
 
     it("should fetch a prompt meta list with name filter", async () => {
@@ -802,10 +874,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(body.data[0].tags).toEqual([]);
 
       // Validate pagination
-      expect(body.pagination.page).toBe(1);
-      expect(body.pagination.limit).toBe(10);
-      expect(body.pagination.totalPages).toBe(1);
-      expect(body.pagination.totalItems).toBe(1);
+      expect(body.meta.page).toBe(1);
+      expect(body.meta.limit).toBe(10);
+      expect(body.meta.totalPages).toBe(1);
+      expect(body.meta.totalItems).toBe(1);
 
       // Test with a different name
       const response2 = await makeAPICall("GET", `${baseURI}?name=prompt-2`);
@@ -819,10 +891,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(body2.data[0].tags).toEqual([]);
 
       // Validate pagination
-      expect(body2.pagination.page).toBe(1);
-      expect(body2.pagination.limit).toBe(10);
-      expect(body2.pagination.totalPages).toBe(1);
-      expect(body2.pagination.totalItems).toBe(1);
+      expect(body2.meta.page).toBe(1);
+      expect(body2.meta.limit).toBe(10);
+      expect(body2.meta.totalPages).toBe(1);
+      expect(body2.meta.totalItems).toBe(1);
 
       // Return 200 with empty list if name does not exist
       const response3 = await makeAPICall(
@@ -846,10 +918,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(body.data[0].tags).toEqual(["tag-1"]);
 
       // Validate pagination
-      expect(body.pagination.page).toBe(1);
-      expect(body.pagination.limit).toBe(10);
-      expect(body.pagination.totalPages).toBe(1);
-      expect(body.pagination.totalItems).toBe(1);
+      expect(body.meta.page).toBe(1);
+      expect(body.meta.limit).toBe(10);
+      expect(body.meta.totalPages).toBe(1);
+      expect(body.meta.totalItems).toBe(1);
 
       // Return 200 with empty list if tag does not exist
       const response3 = await makeAPICall("GET", `${baseURI}?tag=non-existent`);
@@ -875,10 +947,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       ).toBe(true);
 
       // Validate pagination
-      expect(body.pagination.page).toBe(1);
-      expect(body.pagination.limit).toBe(10);
-      expect(body.pagination.totalPages).toBe(1);
-      expect(body.pagination.totalItems).toBe(3);
+      expect(body.meta.page).toBe(1);
+      expect(body.meta.limit).toBe(10);
+      expect(body.meta.totalPages).toBe(1);
+      expect(body.meta.totalItems).toBe(3);
 
       // Test with a different label
       const response2 = await makeAPICall("GET", `${baseURI}?label=dev`);
@@ -892,10 +964,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
       expect(body2.data[0].tags).toEqual([]);
 
       // Validate pagination
-      expect(body2.pagination.page).toBe(1);
-      expect(body2.pagination.limit).toBe(10);
-      expect(body2.pagination.totalPages).toBe(1);
-      expect(body2.pagination.totalItems).toBe(1);
+      expect(body2.meta.page).toBe(1);
+      expect(body2.meta.limit).toBe(10);
+      expect(body2.meta.totalPages).toBe(1);
+      expect(body2.meta.totalItems).toBe(1);
 
       // Return 200 with empty list if label does not exist
       const response3 = await makeAPICall(
@@ -914,10 +986,10 @@ describe("/api/public/v2/prompts API Endpoint", () => {
     const body = response.body as unknown as PromptsMetaResponse;
 
     expect(body.data).toHaveLength(1);
-    expect(body.pagination.page).toBe(1);
-    expect(body.pagination.limit).toBe(1);
-    expect(body.pagination.totalPages).toBe(3);
-    expect(body.pagination.totalItems).toBe(3);
+    expect(body.meta.page).toBe(1);
+    expect(body.meta.limit).toBe(1);
+    expect(body.meta.totalPages).toBe(3);
+    expect(body.meta.totalItems).toBe(3);
   });
 });
 

web/src/__tests__/query-builder.servertest.ts

@@ -50,7 +50,7 @@ describe("Build valid SQL queries", () => {
         table: "traces_scores",
         values: ["project-id"],
         strings: [
-          ' FROM  traces t JOIN scores s ON t.id = s.trace_id  WHERE  t."project_id" = ',
+          ` FROM  traces t JOIN scores s ON t.id = s.trace_id AND s.data_type != 'CATEGORICAL' AND t.project_id = s.project_id  WHERE  t."project_id" = `,
           ";",
         ],
       } as const,

web/src/__tests__/scores.servertest.ts

@@ -60,6 +60,7 @@ describe("/api/public/scores API Endpoint", () => {
     expect(dbScore?.observationId).toBeNull();
     expect(dbScore?.comment).toBe("comment");
     expect(dbScore?.source).toBe("API");
+    expect(dbScore?.projectId).toBe("7a88fb47-b4e2-43b8-a06c-a5ce950dc53a");
   });
 
   it("should create score for a trace with int", async () => {
@@ -488,16 +489,16 @@ describe("/api/public/scores API Endpoint", () => {
         totalPages: 1,
       });
       expect(getScore.body.data).toMatchObject([
-        {
-          id: scoreId_1,
-          name: scoreName,
-          value: 10.5,
-        },
         {
           id: scoreId_2,
           name: scoreName,
           value: 50.5,
         },
+        {
+          id: scoreId_1,
+          name: scoreName,
+          value: 10.5,
+        },
       ]);
     });
     it("test operator >=", async () => {
@@ -513,16 +514,16 @@ describe("/api/public/scores API Endpoint", () => {
         totalPages: 1,
       });
       expect(getScore.body.data).toMatchObject([
-        {
-          id: scoreId_2,
-          name: scoreName,
-          value: 50.5,
-        },
         {
           id: scoreId_3,
           name: scoreName,
           value: 100.8,
         },
+        {
+          id: scoreId_2,
+          name: scoreName,
+          value: 50.5,
+        },
       ]);
     });
     it("test operator !=", async () => {
@@ -538,16 +539,16 @@ describe("/api/public/scores API Endpoint", () => {
         totalPages: 1,
       });
       expect(getScore.body.data).toMatchObject([
-        {
-          id: scoreId_1,
-          name: scoreName,
-          value: 10.5,
-        },
         {
           id: scoreId_3,
           name: scoreName,
           value: 100.8,
         },
+        {
+          id: scoreId_1,
+          name: scoreName,
+          value: 10.5,
+        },
       ]);
     });
     it("test operator =", async () => {

web/src/__tests__/traces.servertest.ts

@@ -182,6 +182,32 @@ describe("/api/public/traces API Endpoint", () => {
       endTime: "2021-01-01T00:20:00.000Z",
     });
 
+    // Simulate scores on the trace
+    const scoreId1 = uuidv4();
+    await makeAPICall("POST", "/api/public/scores", {
+      id: scoreId1,
+      name: "score-1",
+      value: 75.0,
+      traceId: traceId,
+      comment: "First score",
+    });
+    const scoreId2 = uuidv4();
+    await makeAPICall("POST", "/api/public/scores", {
+      id: scoreId2,
+      name: "score-2",
+      value: 85.5,
+      traceId: traceId,
+      comment: "Second score",
+    });
+    const scoreId3 = uuidv4();
+    await makeAPICall("POST", "/api/public/scores", {
+      id: scoreId3,
+      name: "score-3",
+      value: 95.0,
+      traceId: traceId,
+      comment: "Third score",
+    });
+
     // GET traces
     // Retrieve the trace with totalCost and latency
     const traces = await makeAPICall<GetTracesAPIResponse>(
@@ -212,4 +238,62 @@ describe("/api/public/traces API Endpoint", () => {
     expect(trace.body.htmlPath).toContain(`/traces/${traceId}`);
     expect(trace.body.htmlPath).toContain(`/project/`); // do not know the projectId
   });
+
+  it("should filter traces by session ID", async () => {
+    const sessionId = "test-session-id";
+    const anotherSessionId = "another-session-id";
+
+    // Create traces with different session IDs
+    await makeAPICall("POST", "/api/public/traces", {
+      id: "trace-1",
+      name: "test-trace-1",
+      sessionId,
+      userId: "user-1",
+      projectId: "project-1",
+      metadata: { key: "value" },
+      release: "1.0.0",
+      version: "1.0.0",
+    });
+
+    await makeAPICall("POST", "/api/public/traces", {
+      id: "trace-2",
+      name: "test-trace-2",
+      sessionId: anotherSessionId,
+      userId: "user-2",
+      projectId: "project-1",
+      metadata: { key: "value" },
+      release: "1.0.0",
+      version: "1.0.0",
+    });
+
+    // Filter by session ID
+    const tracesBySessionId = await makeAPICall<GetTracesAPIResponse>(
+      "GET",
+      `/api/public/traces?sessionId=${sessionId}`,
+    );
+
+    expect(tracesBySessionId.status).toBe(200);
+    expect(tracesBySessionId.body.data).toHaveLength(1);
+    expect(tracesBySessionId.body.data[0].id).toBe("trace-1");
+
+    // Filter by another session ID
+    const tracesByAnotherSessionId = await makeAPICall<GetTracesAPIResponse>(
+      "GET",
+      `/api/public/traces?sessionId=${anotherSessionId}`,
+    );
+
+    expect(tracesByAnotherSessionId.status).toBe(200);
+    expect(tracesByAnotherSessionId.body.data).toHaveLength(1);
+    expect(tracesByAnotherSessionId.body.data[0].id).toBe("trace-2");
+
+    // Filter by non-existent session ID
+    const tracesByNonExistentSessionId =
+      await makeAPICall<GetTracesAPIResponse>(
+        "GET",
+        `/api/public/traces?sessionId=non-existent-session-id`,
+      );
+
+    expect(tracesByNonExistentSessionId.status).toBe(200);
+    expect(tracesByNonExistentSessionId.body.data).toHaveLength(0);
+  });
 });

web/src/__tests__/zod.servertest.ts

@@ -1,4 +1,4 @@
-import { paginationZod } from "@/src/utils/zod";
+import { paginationZod } from "@langfuse/shared";
 import { ZodError } from "zod";
 
 // Create test cases

web/src/app/api/chatCompletion/route.ts

@@ -0,0 +1,6 @@
+import chatCompletionHandler from "@/src/ee/features/playground/server/chatCompletionHandler";
+
+export const dynamic = "force-dynamic";
+export const maxDuration = 60;
+
+export const POST = chatCompletionHandler;

web/src/components/ChatMessages/ChatMessageComponent.tsx

@@ -63,6 +63,7 @@ export const ChatMessageComponent: React.FC<ChatMessageProps> = ({
 
         <Textarea
           ref={textAreaRef}
+          id={message.id}
           className="height-[auto] min-h-8 w-full pt-3  font-mono text-xs focus:outline-none"
           placeholder={placeholder}
           value={message.content}

web/src/components/LangfuseLogo.tsx

@@ -1,8 +1,105 @@
-import { AlertTriangle, Check } from "lucide-react";
+import {
+  AlertTriangle,
+  Check,
+  Github,
+  HardDriveDownload,
+  Newspaper,
+} from "lucide-react";
 
 import { VERSION } from "@/src/constants";
 import { env } from "@/src/env.mjs";
 import { cn } from "@/src/utils/tailwind";
+import {
+  DropdownMenu,
+  DropdownMenuTrigger,
+  DropdownMenuContent,
+  DropdownMenuItem,
+  DropdownMenuLabel,
+  DropdownMenuSeparator,
+} from "@/src/components/ui/dropdown-menu";
+import { ArrowUp } from "lucide-react";
+import { api } from "@/src/utils/api";
+import { Button } from "@/src/components/ui/button";
+import Link from "next/link";
+
+const VersionLabel = ({ className }: { className?: string }) => {
+  const checkUpdate = api.public.checkUpdate.useQuery(undefined, {
+    refetchOnMount: false,
+    refetchOnWindowFocus: false,
+    refetchOnReconnect: false,
+    enabled: !env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION, // do not check for updates on Langfuse Cloud
+    onError: (error) => console.error("checkUpdate error", error), // do not render default error message
+  });
+
+  const hasUpdate =
+    !env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION &&
+    checkUpdate.data &&
+    checkUpdate.data.updateType;
+
+  const color =
+    checkUpdate.data?.updateType === "major"
+      ? "text-dark-red"
+      : checkUpdate.data?.updateType === "minor"
+        ? "text-dark-yellow"
+        : undefined;
+
+  return (
+    <DropdownMenu>
+      <DropdownMenuTrigger asChild>
+        <Button variant="ghost" size="xs" className={className}>
+          {VERSION}
+          {hasUpdate && <ArrowUp className={`ml-1 h-3 w-3 ${color}`} />}
+        </Button>
+      </DropdownMenuTrigger>
+      <DropdownMenuContent>
+        {hasUpdate ? (
+          <>
+            <DropdownMenuLabel>
+              New {checkUpdate.data?.updateType} version:{" "}
+              {checkUpdate.data?.latestRelease}
+            </DropdownMenuLabel>
+            <DropdownMenuSeparator />
+          </>
+        ) : !env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION ? (
+          <>
+            <DropdownMenuLabel>This is the latest release</DropdownMenuLabel>
+            <DropdownMenuSeparator />
+          </>
+        ) : null}
+
+        <DropdownMenuItem asChild>
+          <Link
+            href="https://github.com/langfuse/langfuse/releases"
+            target="_blank"
+          >
+            <Github size={16} className="mr-2" />
+            Releases
+          </Link>
+        </DropdownMenuItem>
+        <DropdownMenuItem asChild>
+          <Link href="https://langfuse.com/changelog" target="_blank">
+            <Newspaper size={16} className="mr-2" />
+            Changelog
+          </Link>
+        </DropdownMenuItem>
+        {hasUpdate && (
+          <>
+            <DropdownMenuSeparator />
+            <DropdownMenuItem asChild>
+              <Link
+                href="https://langfuse.com/docs/deployment/self-host#update"
+                target="_blank"
+              >
+                <HardDriveDownload size={16} className="mr-2" />
+                Update
+              </Link>
+            </DropdownMenuItem>
+          </>
+        )}
+      </DropdownMenuContent>
+    </DropdownMenu>
+  );
+};
 
 export const LangfuseIcon = ({
   size = 32,
@@ -39,18 +136,18 @@ export const LangfuseLogo = ({
     {showEnvLabel && env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION && (
       <div
         className={cn(
-          "flex items-center gap-2 self-stretch rounded-md px-3 py-2 text-xs ring-1 lg:-mx-2",
+          "flex items-center gap-2 self-stretch rounded-md px-1 py-1 text-xs ring-1 sm:px-3 sm:py-2 lg:-mx-2",
           env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION === "STAGING"
-            ? "bg-blue-100 text-blue-500 ring-blue-500"
+            ? "bg-light-blue text-dark-blue ring-dark-blue"
             : env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION === "DEV"
-              ? "bg-green-100 text-green-500 ring-green-500"
-              : "bg-red-100 text-red-500 ring-red-500",
+              ? "bg-light-green text-dark-green ring-dark-green"
+              : "bg-light-red text-dark-red ring-dark-red",
         )}
       >
         {env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION === "DEV" ? (
-          <Check size={16} />
+          <Check size={16} className="hidden sm:block" />
         ) : (
-          <AlertTriangle size={16} />
+          <AlertTriangle size={16} className="hidden sm:block" />
         )}
         <span className="whitespace-nowrap">
           {["EU", "US"].includes(env.NEXT_PUBLIC_LANGFUSE_CLOUD_REGION)
@@ -65,22 +162,12 @@ export const LangfuseLogo = ({
       <span
         className={cn(
           "ml-2 font-mono font-semibold",
-          size === "sm" ? "text-sm" : "text-xl",
+          size === "sm" ? "text-sm" : "text-lg",
         )}
       >
         Langfuse
       </span>
-      {version && (
-        <a
-          href="https://github.com/langfuse/langfuse/releases"
-          target="_blank"
-          rel="noopener"
-          title="View releases on GitHub"
-          className="ml-2 text-xs text-gray-400"
-        >
-          {VERSION}
-        </a>
-      )}
+      {version && <VersionLabel className="ml-2" />}
     </div>
   </div>
 );

web/src/components/ModelParameters/LLMApiKeyComponent.tsx

@@ -0,0 +1,77 @@
+import Link from "next/link";
+
+import { Label } from "@/src/components/ui/label";
+import { useHasAccess } from "@/src/features/rbac/utils/checkAccess";
+import { api } from "@/src/utils/api";
+import { supportedModels, type UIModelParams } from "@langfuse/shared";
+import { ArrowTopRightIcon } from "@radix-ui/react-icons";
+
+export const LLMApiKeyComponent = (p: {
+  projectId: string;
+  modelParams: UIModelParams;
+}) => {
+  const hasAccess = useHasAccess({
+    projectId: p.projectId,
+    scope: "llmApiKeys:read",
+  });
+
+  if (!hasAccess) {
+    return (
+      <div>
+        <Label className="text-xs font-semibold">API key</Label>
+        <p className="text-sm text-muted-foreground">
+          LLM API Key only visible to Owner and Admin roles.
+        </p>
+      </div>
+    );
+  }
+
+  const apiKeys = api.llmApiKey.all.useQuery({
+    projectId: p.projectId,
+  });
+
+  if (apiKeys.isLoading) {
+    return (
+      <div>
+        <Label className="text-xs font-semibold">API key</Label>
+        <p className="text-sm text-muted-foreground">Loading...</p>
+      </div>
+    );
+  }
+
+  const model = p.modelParams.model.value;
+  const modelProvider = Object.entries(supportedModels).find((providerData) =>
+    (providerData[1] as any as string[]).includes(model),
+  )?.[0];
+
+  const apiKey = apiKeys.data?.data.find((k) => k.provider === modelProvider);
+
+  return (
+    <div className="space-y-2 text-xs">
+      <Label className="text-xs font-semibold">API key</Label>
+      <div>
+        {apiKey ? (
+          <span className="mr-2 rounded-sm bg-input p-1 text-xs">
+            {apiKey.displaySecretKey}
+          </span>
+        ) : undefined}
+      </div>
+      {/* Custom form message to include a link to the already existing prompt */}
+      {!apiKey ? (
+        <div className="flex flex-col font-medium text-destructive">
+          {`No LLM API key found for provider ${modelProvider}.`}
+
+          <Link
+            href={`/project/${p.projectId}/settings`}
+            className="flex flex-row"
+          >
+            Create a new LLM API key here. <ArrowTopRightIcon />
+          </Link>
+        </div>
+      ) : undefined}
+      <p className="text-muted-foreground">
+        The LLM API key is used for each execution and will incur costs.
+      </p>
+    </div>
+  );
+};

web/src/components/ModelParameters/index.tsx

@@ -1,8 +1,3 @@
-import {
-  supportedModels,
-  ModelProvider,
-  type UIModelParams,
-} from "@langfuse/shared";
 import { Input } from "@/src/components/ui/input";
 import {
   Select,
@@ -12,86 +7,111 @@ import {
   SelectValue,
 } from "@/src/components/ui/select";
 import { Slider } from "@/src/components/ui/slider";
+import { Switch } from "@/src/components/ui/switch";
+import useProjectIdFromURL from "@/src/hooks/useProjectIdFromURL";
+import { cn } from "@/src/utils/tailwind";
+import {
+  ModelProvider,
+  supportedModels,
+  type UIModelParams,
+} from "@langfuse/shared";
+
+import { LLMApiKeyComponent } from "./LLMApiKeyComponent";
 
 export type ModelParamsContext = {
   modelParams: UIModelParams;
   availableModels?: UIModelParams[];
-  updateModelParam: <Key extends keyof UIModelParams>(
+  updateModelParamValue: <Key extends keyof UIModelParams>(
     key: Key,
-    value: UIModelParams[Key],
+    value: UIModelParams[Key]["value"],
   ) => void;
-  disabled?: boolean;
+  setModelParamEnabled?: (key: keyof UIModelParams, enabled: boolean) => void;
+  formDisabled?: boolean;
 };
 
 export const ModelParameters: React.FC<ModelParamsContext> = ({
   modelParams,
   availableModels,
-  updateModelParam,
-  disabled,
+  updateModelParamValue,
+  setModelParamEnabled,
+  formDisabled = false,
 }) => {
+  const projectId = useProjectIdFromURL();
+
+  if (!projectId) return null;
+
   return (
     <div className="flex flex-col space-y-4">
       <p className="font-semibold">Model</p>
-      <div className="space-y-6">
+      <div className="space-y-4">
         <ModelParamsSelect
           title="Provider"
           modelParamsKey="provider"
-          disabled={disabled}
-          value={modelParams.provider}
+          disabled={formDisabled}
+          value={modelParams.provider.value}
           options={
             availableModels
-              ? [...new Set(availableModels.map((m) => m.provider))]
+              ? [...new Set(availableModels.map((m) => m.provider.value))]
               : Object.values(ModelProvider)
           }
-          updateModelParam={updateModelParam}
+          updateModelParam={updateModelParamValue}
         />
         <ModelParamsSelect
           title="Model name"
           modelParamsKey="model"
-          disabled={disabled}
-          value={modelParams.model}
+          disabled={formDisabled}
+          value={modelParams.model.value}
           options={Object.values(
             availableModels
               ? availableModels
-                  .filter((m) => m.provider === modelParams.provider)
-                  .map((m) => m.model)
-              : supportedModels[modelParams.provider],
+                  .filter(
+                    (m) => m.provider.value === modelParams.provider.value,
+                  )
+                  .map((m) => m.model.value)
+              : supportedModels[modelParams.provider.value],
           )}
-          updateModelParam={updateModelParam}
+          updateModelParam={updateModelParamValue}
         />
         <ModelParamsSlider
           title="Temperature"
           modelParamsKey="temperature"
-          disabled={disabled}
-          value={modelParams.temperature}
+          formDisabled={formDisabled}
+          enabled={modelParams.temperature.enabled}
+          setModelParamEnabled={setModelParamEnabled}
+          value={modelParams.temperature.value}
           min={0}
-          max={modelParams.maxTemperature}
+          max={modelParams.maxTemperature.value}
           step={0.01}
-          tooltip="The sampling temperature. Higher values will make the output more random, while lower values like will make it more focused and deterministic."
-          updateModelParam={updateModelParam}
+          tooltip="The sampling temperature. Higher values will make the output more random, while lower values will make it more focused and deterministic."
+          updateModelParam={updateModelParamValue}
         />
         <ModelParamsSlider
           title="Output token limit"
           modelParamsKey="max_tokens"
-          disabled={disabled}
-          value={modelParams.max_tokens}
+          formDisabled={formDisabled}
+          enabled={modelParams.max_tokens.enabled}
+          setModelParamEnabled={setModelParamEnabled}
+          value={modelParams.max_tokens.value}
           min={1}
           max={4096}
           step={1}
           tooltip="The maximum number of tokens that can be generated in the chat completion."
-          updateModelParam={updateModelParam}
+          updateModelParam={updateModelParamValue}
         />
         <ModelParamsSlider
           title="Top P"
           modelParamsKey="top_p"
-          disabled={disabled}
-          value={modelParams.top_p}
+          formDisabled={formDisabled}
+          enabled={modelParams.top_p.enabled}
+          setModelParamEnabled={setModelParamEnabled}
+          value={modelParams.top_p.value}
           min={0}
           max={1}
           step={0.01}
           tooltip="An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or temperature but not both."
-          updateModelParam={updateModelParam}
+          updateModelParam={updateModelParamValue}
         />
+        <LLMApiKeyComponent {...{ projectId, modelParams }} />
       </div>
     </div>
   );
@@ -102,7 +122,7 @@ type ModelParamsSelectProps = {
   modelParamsKey: keyof UIModelParams;
   value: string;
   options: string[];
-  updateModelParam: ModelParamsContext["updateModelParam"];
+  updateModelParam: ModelParamsContext["updateModelParamValue"];
   disabled?: boolean;
 };
 const ModelParamsSelect = ({
@@ -115,7 +135,14 @@ const ModelParamsSelect = ({
 }: ModelParamsSelectProps) => {
   return (
     <div className="space-y-2">
-      <p className="text-xs font-semibold">{title}</p>
+      <p
+        className={cn(
+          "text-xs font-semibold",
+          disabled && "text-muted-foreground",
+        )}
+      >
+        {title}
+      </p>
       <Select
         disabled={disabled}
         onValueChange={(value) =>
@@ -149,8 +176,10 @@ type ModelParamsSliderProps = {
   min: number;
   max: number;
   step: number;
-  updateModelParam: ModelParamsContext["updateModelParam"];
-  disabled?: boolean;
+  updateModelParam: ModelParamsContext["updateModelParamValue"];
+  enabled?: boolean;
+  formDisabled?: boolean;
+  setModelParamEnabled?: ModelParamsContext["setModelParamEnabled"];
 };
 const ModelParamsSlider = ({
   title,
@@ -161,30 +190,51 @@ const ModelParamsSlider = ({
   max,
   step,
   updateModelParam,
-  disabled,
+  setModelParamEnabled,
+  enabled,
+  formDisabled,
 }: ModelParamsSliderProps) => {
   return (
     <div className="space-y-3" title={tooltip}>
       <div className="flex flex-row">
-        <p className="flex-1 text-xs font-semibold">{title}</p>
-        <Input
-          className="h-6 w-14 appearance-none px-2 text-right"
-          type="number"
-          disabled={disabled}
-          min={min}
-          max={max}
-          step={step}
-          value={value}
-          onChange={(event) => {
-            updateModelParam(
-              modelParamsKey,
-              Math.max(Math.min(parseFloat(event.target.value), max), min),
-            );
-          }}
-        />
+        <p
+          className={cn(
+            "flex-1 text-xs font-semibold",
+            (!enabled || formDisabled) && "text-muted-foreground",
+          )}
+        >
+          {title}
+        </p>
+        <div className="flex flex-row space-x-3">
+          <Input
+            className="h-6 w-14 appearance-none px-2 text-right"
+            type="number"
+            disabled={!enabled || formDisabled}
+            min={min}
+            max={max}
+            step={step}
+            value={value}
+            onChange={(event) => {
+              updateModelParam(
+                modelParamsKey,
+                Math.max(Math.min(parseFloat(event.target.value), max), min),
+              );
+            }}
+          />
+          {setModelParamEnabled ? (
+            <Switch
+              title={`Control sending the ${title} parameter`}
+              disabled={formDisabled}
+              checked={enabled}
+              onCheckedChange={(checked) => {
+                setModelParamEnabled(modelParamsKey, checked);
+              }}
+            />
+          ) : null}
+        </div>
       </div>
       <Slider
-        disabled={disabled}
+        disabled={!enabled || formDisabled}
         min={min}
         max={max}
         step={step}

web/src/components/Slider.tsx

@@ -7,27 +7,23 @@ export const Slider = (props: {
   loading?: boolean;
   onChecked?: (checked: boolean) => void;
   isChecked?: boolean; // whether the slider is active
-}) => {
-  console.log("props", props);
-
-  return (
-    <Switch
-      checked={props.isChecked}
-      disabled={props.loading || props.disabled}
-      onChange={props.onChecked}
+}) => (
+  <Switch
+    checked={props.isChecked}
+    disabled={props.loading || props.disabled}
+    onChange={props.onChecked}
+    className={cn(
+      props.isChecked ? "bg-background" : "bg-input",
+      "relative inline-flex h-6 w-11 flex-shrink-0 cursor-pointer rounded-full border-2 border-transparent transition-colors duration-200 ease-in-out focus:outline-none focus:ring-2 focus:ring-foreground focus:ring-offset-2",
+    )}
+  >
+    <span className="sr-only">Use setting</span>
+    <span
+      aria-hidden="true"
       className={cn(
-        props.isChecked ? "bg-black" : "bg-gray-200",
-        "relative inline-flex h-6 w-11 flex-shrink-0 cursor-pointer rounded-full border-2 border-transparent transition-colors duration-200 ease-in-out focus:outline-none focus:ring-2 focus:ring-black focus:ring-offset-2",
+        props.isChecked ? "translate-x-5" : "translate-x-0",
+        "pointer-events-none inline-block h-5 w-5 transform rounded-full bg-background shadow ring-0 transition duration-200 ease-in-out",
       )}
-    >
-      <span className="sr-only">Use setting</span>
-      <span
-        aria-hidden="true"
-        className={cn(
-          props.isChecked ? "translate-x-5" : "translate-x-0",
-          "pointer-events-none inline-block h-5 w-5 transform rounded-full bg-white shadow ring-0 transition duration-200 ease-in-out",
-        )}
-      />
-    </Switch>
-  );
-};
+    />
+  </Switch>
+);

web/src/components/Support.tsx

@@ -8,14 +8,7 @@ import {
   CardTitle,
 } from "@/src/components/ui/card";
 import { chatAvailable, openChat } from "@/src/features/support-chat/chat";
-import {
-  Book,
-  Calendar,
-  Github,
-  Mail,
-  MessageSquare,
-  Slack,
-} from "lucide-react";
+import { Book, Github, Mail, MessageSquare, Slack } from "lucide-react";
 import Link from "next/link";
 import { SiDiscord } from "react-icons/si";
 
@@ -30,7 +23,7 @@ const supportChannels = [
   },
   {
     icon: Github,
-    title: "GitHub",
+    title: "GitHub Issues",
     description:
       "Create an issue on Github to report bugs or request new features.",
     href: "https://github.com/langfuse/langfuse/issues/new/choose",
@@ -38,12 +31,11 @@ const supportChannels = [
     primary: true,
   },
   {
-    icon: SiDiscord,
-    title: "Discord",
-    description:
-      "Get support from community and maintainers. Follow announcements to stay up to date with new features.",
-    href: "https://langfuse.com/discord",
-    buttonText: "Join Discord",
+    icon: Github,
+    title: "GitHub Support",
+    description: "Create a support ticket via GitHub discussions.",
+    href: "https://github.com/orgs/langfuse/discussions/categories/support",
+    buttonText: "Submit question",
     primary: true,
   },
   {
@@ -57,24 +49,25 @@ const supportChannels = [
   {
     icon: MessageSquare,
     title: "Chat",
-    description: "Get quick support directly from the team.",
+    description: "Get support directly from the team.",
     onClick: () => openChat(),
     available: chatAvailable,
     buttonText: "Launch Chat",
   },
   {
-    icon: Calendar,
-    title: "Schedule call",
-    description: "Schedule a call with one of the founders.",
-    href: "https://cal.com/marc-kl/office-hours",
-    buttonText: "Schedule Call",
+    icon: SiDiscord,
+    title: "Discord",
+    description:
+      "Get support from community. Follow announcements to stay up to date with new features.",
+    href: "https://langfuse.com/discord",
+    buttonText: "Join Discord",
   },
   {
     icon: Slack,
     title: "Slack Connect",
     description: "Get a dedicated support channel for you and your team.",
     href: "mailto:help@langfuse.com?subject=Slack%20Connect%20Request&body=I'd%20like%20to%20request%20a%20dedicated%20Slack%20Connect%20channel%20for%20me%20and%20my%20team.%0D%0A%0D%0AUsers%20(emails)%20to%20include%20besides%20mine%3A%0D%0A%0D%0A",
-    buttonText: "Request Slack Connect (via Email)",
+    buttonText: "Request via Email",
   },
 ];
 

web/src/components/deleteButton.tsx

@@ -53,16 +53,13 @@ export function DeleteButton({
     },
   });
 
-  if (!hasAccess) {
-    return null;
-  }
-
   return (
     <Popover key={itemId}>
       <PopoverTrigger asChild>
         <Button
           variant={isTableAction ? "ghost" : "outline"}
           size={isTableAction ? "xs" : "icon"}
+          disabled={!hasAccess}
           onClick={() =>
             type === "trace"
               ? capture("trace:delete_form_open", {

web/src/components/error-page.tsx

@@ -18,7 +18,7 @@ export const ErrorPage = ({
 
   return (
     <div className="flex h-screen flex-col items-center justify-center">
-      <AlertCircle className="mb-4 h-12 w-12 text-red-500" />
+      <AlertCircle className="text-dark-red mb-4 h-12 w-12" />
       <h1 className="mb-4 text-xl font-bold">{title}</h1>
       <p className="mb-8 text-center">{message}</p>
       {session.status === "unauthenticated" ? (

web/src/components/grouped-score-badge.tsx

@@ -4,9 +4,13 @@ import {
   HoverCardContent,
   HoverCardTrigger,
 } from "@/src/components/ui/hover-card";
+import {
+  isBooleanDataType,
+  isCategoricalDataType,
+} from "@/src/features/manual-scoring/lib/helpers";
 import { type ScoreSimplified } from "@/src/server/api/routers/generations/getAllQuery";
 import { cn } from "@/src/utils/tailwind";
-import { MessageCircle } from "lucide-react";
+import { MessageCircleMore } from "lucide-react";
 
 export const GroupedScoreBadges = ({
   scores,
@@ -34,11 +38,13 @@ export const GroupedScoreBadges = ({
     <div className={cn("text-xs", props.className)}>
       {props.scores.map((s, i) => (
         <span key={i} className="group/score ml-1 first:ml-0">
-          {s.value.toFixed(2)}
+          {isCategoricalDataType(s.dataType) || isBooleanDataType(s.dataType)
+            ? s.stringValue
+            : s.value.toFixed(2)}
           {s.comment && (
             <HoverCard>
               <HoverCardTrigger className="ml-1 inline-block cursor-pointer">
-                <MessageCircle size={12} />
+                <MessageCircleMore size={12} />
               </HoverCardTrigger>
               <HoverCardContent className="overflow-hidden whitespace-normal break-normal">
                 <p>{s.comment}</p>
@@ -58,7 +64,7 @@ export const GroupedScoreBadges = ({
           .sort(([a], [b]) => (a < b ? -1 : 1))
           .map(([name, scores]) => (
             <div key={name}>
-              <div className="text-xs text-gray-500">{name}</div>
+              <div className="text-xs text-muted-foreground">{name}</div>
               <ScoresOfGroup scores={scores} />
             </div>
           ))}
@@ -73,9 +79,12 @@ export const GroupedScoreBadges = ({
             <Badge
               variant="outline"
               key={name}
-              className="break-all font-normal"
+              className="grid grid-cols-[1fr,auto] gap-1 font-normal"
             >
-              {name}: <ScoresOfGroup scores={scores} className="ml-2" />
+              <p className="truncate" title={name}>
+                {name}:
+              </p>
+              <ScoresOfGroup scores={scores} />
             </Badge>
           ))}
       </>