LaserLab v0.3.1 fixture provenance and golden regressions

[ProhibitedTV]

Summary

• makes fixture catalog metadata enforceable, including phenomena, licensing, attribution, expected detector behavior, limitations, and redistribution status
• carries fixture provenance through ingestion, blinded samples, detector results, JSON reports, and visible HTML report tables
• expands the dashboard fixture catalog and adds a researcher-facing fixture catalog document
• adds compact golden scientific thresholds and a matched-null regression that cannot promote above the control evidence level
• retains HTML, JSON, processed images, and candidate crops when CI analysis fails
• bumps the package to v0.3.1

Why

Community review depends on knowing what each fixture is expected to demonstrate and on preventing detector changes from silently becoming more permissive. This patch makes fixture provenance auditable and puts conservative null behavior under CI protection.

Validation

• python -m compileall laserlab gui tests scripts
• python -m unittest discover -s tests (11 tests)
• one-frame bundled fixture demo with source provenance verified in report.json
• fixture catalog validates all required metadata and all bundled files
• git diff --check

Closes #16
Closes #17

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ce4a7a0451

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".

[chatgpt-codex-connector]

Preserve fixture metadata for generated controls

With the default control_generation setting, laser captures produce matched control samples via generate_matched_controls, but those records do not carry capture_metadata from the parent capture. As a result this newly added result field becomes {} for every generated control, so if a derived control is part of results.json or ranks as a top candidate, the fixture provenance is lost even though the parent laser capture had it.

Useful? React with 👍 / 👎.