Repository navigation
Description
Task
T-079· Milestone M6 — Analysis + risk · Size M · Priority P1 · Epic #6Implements FR-EXP-003 · PR-10
Branch
feat/t-079-experiment-html-report-design-table— one task = one branch = one PR (doc 32 §6)
Summary
Implement render_experiment_report(...) per doc 19 §7: design table, per-metric box plots, discovery curves with 95 % bands, Kaplan–Meier, statistics table (median, IQR, p, p_Holm, Â12 + verbal effect), case gallery, threats to validity, environment manifest, raw-data Parquet link.
Context / motivation
The deliverable that answers the research question (doc 19 §1).
Scope
In scope
- Reports all strategies and metrics, including where S-PROBE is not better (no cherry-picking, doc 19 §5).
Implementation guidance
roboprobe/reporting/experiment.py:def render_experiment_report(experiment_id: UUID, uow, store, *, out=None) -> Path.- Data from the experiment comparison service (T-102b) — the same DTO the API returns for
/experiments/{id}/comparison(doc 12 §3.5), so the HTML and dashboard can never disagree. - Sections per doc 19 §7: design table; per-metric box plots (all strategies); discovery curves with 95 % bootstrap bands (
experiments.bootstrap_resamples); Kaplan–Meier curves for MET-03 (censored at budget); statistics table: median, IQR, MWU p, Holm-adjusted p, Â12 with verbal magnitude (negligible < 0.56 ≤ small < 0.64 ≤ medium < 0.71 ≤ large; thresholds inexperiments.a12_thresholds, new); case gallery (top modes per strategy); threats to validity (doc 19 §8 text + experiment notes); environment manifest; link/hash of raw Parquet. - Pre-registration banner: shows the frozen S-PROBE hyper-parameter hash (T-061) and flags any deviation.
- Standards: respect the layering contracts (doc 16 §2) and conventions in doc 34; no hard-coded tunables — thresholds/defaults come from typed settings (doc 27); no magic numbers (ruff
PLR2004); public APIs typed (mypy strict) and documented.
Configuration keys
From the configuration reference (doc 27). Keys marked (new) must be added to the typed settings models and appear in the generated reference (T-003c).
experiments.bootstrap_resamplesexperiments.alpha- experiments.a12_thresholds (new)
Acceptance criteria
- Report for a small analytic experiment renders offline.
- All strategies and metrics shown (no filtering) — test asserts every strategy appears in every table.
- Holm-adjusted p ≥ raw p for all rows (sanity test).
- Renders offline.
- Tests listed below added and green in CI; coverage of new code meets the gate in doc 18 §3.
- Docs, config reference and CHANGELOG
Unreleasedupdated where behaviour, keys or interfaces changed. - Definition of Done (doc 32 §3) satisfied.
Tests required
tests/integration/reporting/test_experiment_report.py.
Dependencies
Blocked by: #140 (T-078b), #107 (T-101b), #109 (T-102b)
Blocks: #157 (T-084a), #207 (T-105a), #208 (T-105b)
Parallelization notes
Files / paths owned by this task — avoid editing other paths; shared files (pyproject.toml entry points, cli/main.py, api/routers/__init__.py, router.tsx, docker/compose.yaml, ci.yml) are append-only:
backend/src/roboprobe/reporting/experiment.pybackend/src/roboprobe/reporting/templates/experiment/**
Independent.
Definition of Done
Per doc 32 §3 and AGENTS.md §5: lint, format, mypy strict and import-linter green; tests added and passing; docs updated; PR template completed with Closes #<this issue>; Conventional Commit messages referencing T-079 (doc 16 §6); commits authored only by the repository owner identity (AGENTS.md §4); CI green; reviewer checklist (doc 32 §5.1) satisfied.
References
Design docs: doc 19 §7 · doc 19 §5 · doc 19 §8 · doc 19 §3 · doc 12 §3.5
Requirements: FR-EXP-003 · PR-10 (traceability: doc 28)
Metadata
Metadata
Assignees
Labels
Projects
- StatusShow more project fieldsBacklog
Activity
roboprobe report <id> -o report.html#140experiment run|report(concurrency-limited) #109