mirror of
https://github.com/hanzoai/insights.git
synced 2026-11-03 00:52:32 +00:00
zeekay
1b367e2b62
The tree was a debranded snapshot of upstream taken around 2026-02-20, and upstream has moved about five months since. This replaces it with upstream master, mechanically re-derived: bin/debrand does the rename, and the pipeline and squashed migrations are restored from the shipped tree. Not a merge. The previous import replaced the tree wholesale in a single commit, so there is no shared content history to merge against. Reproducing the rename on the new tree is what actually converges. What survives: .hanzo/workflows, the Dockerfiles and bin/docker* entrypoints, compose files, LLM.md, and insights/migrations. The squashed initial migration is load-bearing — the live database records it by name, so upstream's ~918 individual migrations cannot come back. Nine products are dropped as unreferenced. Twenty-one more that we do not run are kept anyway because the core app still imports them; unpicking those is a refactor and does not belong in a rebase. This base does not build yet, and the reasons are not in the rename: upstream now wants Python 3.13.13 against our 3.12.12 and Django 5.2 against our 4.2, and the dependency manifests still name upstream packages that we republish under our own names. Those are decisions, not renames.
3.8 KiB
3.8 KiB
File Classification
Classify the PR file list before planning frontend QA. When in doubt, include a small browser smoke rather than pretending coverage is complete.
| File pattern | Frontend target | Notes |
|---|---|---|
frontend/src/**/*.tsx, frontend/src/**/*.ts |
Browser flow | Use route-finding.md for scene files. For kea logic, find the scene importing it. |
products/*/frontend/**/*.tsx, products/*/frontend/**/*.ts |
Browser flow | Read the product manifest.tsx and route entries. |
products/*/manifest.tsx |
Browser flow | Directly inspect routes, urls, and tree items. |
frontend/**/*.css, frontend/**/*.scss, Tailwind class-only changes |
Visual screenshot | Capture route screenshots. Do not call it pixel-perfect regression. |
| Backend, query, model, or service files | Comment-only unless a frontend route is obvious | This skill is for frontend QA. If a UI flow clearly exercises the change, smoke that route. |
insights/migrations/**, insights/datastore/migrations/** |
Comment-only by default | Warn that local stack may be stale. Do not autonomously edit migrations. |
pnpm-lock.yaml, package.json, requirements*.txt, pyproject.toml |
Comment-only unless user confirms stack updated | Dependency changes can invalidate the local stack. |
.github/**, Dockerfiles, k8s, infra |
Usually comment-only | Frontend QA may be meaningless for local app behavior. |
docs/**, *.md, pure copy outside app UI |
No frontend target | Post "nothing meaningful to frontend QA" if a PR comment is requested. |
Planning Rules
- Use
test-case-design.mdto turn file categories into behavior-focused cases. - Mixed frontend/backend PR: focus on routes where the frontend visibly uses the changed behavior.
- Product-scoped frontend PR: read
products/<product>/manifest.tsxfirst. - Shared component PR: find importing scenes with
rgand choose 1-3 high-signal routes. - If no URL maps, keep a coverage-gap item in the final report.
- If a file exposes user-facing strings, include a screenshot of the route that renders them.
Test Plan Shape
Each planned case should be concrete enough for a later agent to execute:
{
"kind": "browser",
"diff_behavior": "Dashboard save flow now writes filter state through the new field",
"risk": "User saves a dashboard and loses the selected filter on reload",
"expected_behavior": "Saved filter remains visible and no error toast appears",
"oracle_source": "existing dashboard save behavior and nearby filter persistence tests",
"oracle_confidence": "high",
"setup": "Use a dashboard with at least one insight",
"route": "/dashboard/:id",
"action": "Change filter, save, reload dashboard",
"evidence": "Screenshot after reload plus console/network check"
}
Use coverage_gap for changed files that could not be mapped. Coverage gaps are honest output, not failures.