forrtproject / forrtproject/flora-extractor
Per-release recall monitor against flora.csv in the dashboard rules view (supersedes #151)
- Dominant language
- Python
- Stars
- 2
- Forks
- 1
- Avg merge
- 1d 6h
- Merged PRs (30d)
- 4
Description
`filter/spec/CONVENTIONS.md` now bans any LLM-labelled corpus as a recall denominator. That settles the core ask of #151 by policy: recall is not measured against labels the pipeline itself produced.
**The unbuilt remainder.** A per-release recall check of the routing piles against the human-curated `flora.csv` — which known FLoRA replications does each release discard, or leave in `pending`? — surfaced in the Stage 4 dashboard's rules view, so the number moves with each release rather than being computed by hand.
Supersedes #151.
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with filter/spec/CONVENTIONS.md to confirm the recall denominator policy, then trace the Stage 4 dashboard's rules view and how releases expose routing piles. Use the human-curated flora.csv as the reference and consider the work complete when each release shows which known FLoRA replications it discards or leaves pending.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- analytics, data
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 55/100