| Dimension | Shared | Only in A | Only in B | Overlap |
|---|---|---|---|---|
| Sections | 0 | 2 | 4 | 0% |
| Commands | 0 | 0 | 7 | 0% |
| Section tags | 2 | 2 | 2 | 33% |
What each file covers
Sections
0 shared · 2 only in A · 4 only in B- − ML Workspace Instructions
- − Rules
- + GitHub Copilot Instructions
- + Always Follow
- + Stack And Commands
- + Workflow
Commands
0 shared · 0 only in A · 7 only in B- + npm run lint:frontend
- + npm run typecheck:frontend
- + npm test --workspace frontend
- + npm --workspace frontend run build
- + python -m ruff check .
- + python -m pytest
- + python -m mypy app
Section tags
2 shared · 2 only in A · 2 only in B- − monorepo
- − do-not
- + test
- + lint-format
- code-style
- agent-behaviour
Line diff
Aledon8/OpenLeukemia · ml/CLAUDE.md
@@ −1 @@
1# ML Workspace Instructions
2
3This directory is for training pipelines, inference experiments, evaluations, and notebooks.
4
5## Rules
6
7- Keep structured ML workflows separate from LLM explanation workflows.
8- Prefer reproducible scripts and documented evaluation outputs over one-off notebook state.
9- Do not treat model output as diagnosis, urgent clinical judgment, or treatment recommendation.
10- Track assumptions about input data, labels, validation, and limitations close to the code or report that uses them.
11- Avoid committing patient-identifying or sensitive medical data.
12- Prefer measurable success criteria: metric, split, threshold, calibration, and known failure modes.
13
Aledon8/OpenLeukemia · .github/copilot-instructions.md
@@ +1 @@
1# GitHub Copilot Instructions
2
3OpenLeukemia is a patient-centered leukemia data platform. It organizes medical data, tracks changes over time, extracts structured information from documents, and explains validated model outputs in plain language.
4
5## Always Follow
6
7- Preserve explicit consent, privacy-by-default behavior, and patient ownership of data.
8- Registration is not blanket consent.
9- AI output may summarize, extract, and explain; it must not diagnose, recommend treatment, or make urgent clinical judgments.
10- Treat LLM extraction as candidate data until validated or confirmed.
11- Prefer minimal, focused changes that match nearby code.
12- Do not introduce dependencies, API changes, schema changes, or consent-flow changes without clear need.
13- Add or update tests when behavior changes.
14- Run or recommend the relevant check after changes.
15
16## Stack And Commands
17
18- Frontend: React 19, TypeScript, Vite, Vitest, ESLint.
19- Platform: Supabase Auth, Postgres, Storage, RLS, Edge Functions.
20- AI service: Python 3.12, FastAPI, Pydantic, Uvicorn.
21- Frontend checks: `npm run lint:frontend`, `npm run typecheck:frontend`, `npm test --workspace frontend`, `npm --workspace frontend run build`.
22- AI-service checks from `ai-service/`: `python -m ruff check .`, `python -m pytest`, `python -m mypy app`.
23
24## Workflow
25
26- Read nearby files before editing and follow the existing pattern.
27- For multi-file or risky changes, identify touched files and verification steps before implementation.
28- Use artifacts as primary evidence: errors, logs, screenshots, issues, plans, and referenced files.
29- Verify work with tests, build, lint, typecheck, or visual comparison when practical.
30
@@ −1 +1 @@
1−# ML Workspace Instructions
1+# GitHub Copilot Instructions
22
3−This directory is for training pipelines, inference experiments, evaluations, and notebooks.
3+OpenLeukemia is a patient-centered leukemia data platform. It organizes medical data, tracks changes over time, extracts structured information from documents, and explains validated model outputs in plain language.
44
5−## Rules
5+## Always Follow
66
7−- Keep structured ML workflows separate from LLM explanation workflows.
8−- Prefer reproducible scripts and documented evaluation outputs over one-off notebook state.
9−- Do not treat model output as diagnosis, urgent clinical judgment, or treatment recommendation.
10−- Track assumptions about input data, labels, validation, and limitations close to the code or report that uses them.
11−- Avoid committing patient-identifying or sensitive medical data.
12−- Prefer measurable success criteria: metric, split, threshold, calibration, and known failure modes.
7+- Preserve explicit consent, privacy-by-default behavior, and patient ownership of data.
8+- Registration is not blanket consent.
9+- AI output may summarize, extract, and explain; it must not diagnose, recommend treatment, or make urgent clinical judgments.
10+- Treat LLM extraction as candidate data until validated or confirmed.
11+- Prefer minimal, focused changes that match nearby code.
12+- Do not introduce dependencies, API changes, schema changes, or consent-flow changes without clear need.
13+- Add or update tests when behavior changes.
14+- Run or recommend the relevant check after changes.
15+
16+## Stack And Commands
17+
18+- Frontend: React 19, TypeScript, Vite, Vitest, ESLint.
19+- Platform: Supabase Auth, Postgres, Storage, RLS, Edge Functions.
20+- AI service: Python 3.12, FastAPI, Pydantic, Uvicorn.
21+- Frontend checks: `npm run lint:frontend`, `npm run typecheck:frontend`, `npm test --workspace frontend`, `npm --workspace frontend run build`.
22+- AI-service checks from `ai-service/`: `python -m ruff check .`, `python -m pytest`, `python -m mypy app`.
23+
24+## Workflow
25+
26+- Read nearby files before editing and follow the existing pattern.
27+- For multi-file or risky changes, identify touched files and verification steps before implementation.
28+- Use artifacts as primary evidence: errors, logs, screenshots, issues, plans, and referenced files.
29+- Verify work with tests, build, lint, typecheck, or visual comparison when practical.
1330
