Add inspection-only Golden Evidence Cohort - #120
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 292719b42b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| as_of_date = _text(row.get("as_of_date")) | ||
| provenance = tuple( | ||
| field | ||
| for field, value in (("source", source_id), ("source_reference", source_ref), ("as_of_date", as_of_date)) | ||
| if not value |
There was a problem hiding this comment.
Reject invalid fundamental as-of dates
When a fundamentals row contains a nonempty but malformed as_of_date (for example, from a bad manual import), this value is treated as complete provenance, so approved numeric fields can enter usable_evidence_lanes and may make saved DCF evidence usable. Validate the date before clearing the provenance blocker so temporally invalid evidence remains withheld.
Useful? React with 👍 / 👎.
| ticker | ||
| for ticker in candidate_groups | ||
| if ticker not in selected | ||
| and _asset_type(ticker, saved_by_ticker, proposed_by_ticker, universe_by_ticker) in OPERATING_ASSET_TYPES | ||
| ), |
There was a problem hiding this comment.
Restrict the control role to candidates with evidence gaps
When the first distinct operating-company promotion has complete provenance, rights, field scope, and price evidence, it still satisfies this selection and is emitted as evidence_gap_control despite having no independent blockers. In that case the advertised negative control is absent—and _member can even report state="insufficient_evidence" alongside an empty blocker list—so require the selected candidate group to contain an actual evidence blocker.
Useful? React with 👍 / 👎.
Summary
Cohort contract
Verification
Current data boundary
Saved readiness remains stale but tracked: prices 265/3538, fundamentals 23/3538, DCF 23/3538, peers 9/3538. The packet is inspection-only and does not activate any data or research conclusion.