KOPR Live Theta-86 → Theta-110
Developer history of the APCHK redesign and the reference-star source/background correctness chain through the current Theta-110 / 0.1.0.dev625 contract.
Historical attribution caveat: Theta-97 and Theta-99→110 are directly backed by retained release/evidence artifacts. Theta-98 is identified directly as the R2E2B baseline for Theta-99. Theta-86→96 release-by-release attribution is reconstructed from retained current QA contracts and the monotonic implementation sequence; verify original standalone release reports before publishing a formal external per-release history.
Current Theta-110 contract
- Manual aperture adequacy is a non-blocking observer advisory, not a physical coma-edge detector.
- The contamination-resistant automatic path is authoritative for ACF measurement and uses the frozen observer ACF context plus per-radius canonical local background.
- Review follows filtered Curve-of-Growth evidence, stops sequentially at the first evidence boundary, scans contamination already inside the observer aperture and bootstraps tiny apertures.
- Qualified V1 calibration remains 0.10 mag / 20% projected horizon.
- Try suggested aperture is preview-only; remeasurement requires explicit Perform Comet Measure.
- Reference-star measurement overlays are source-local; frozen background provenance is bound to the true source image.
- Invalid reference background is a per-star quality rejection; the canonical ensemble solver owns whole-observation sufficiency. Comet background remains strict fail-closed.
Release sequence
| Release | Package | Main change |
|---|---|---|
| Theta-86 | 0.1.0.dev601 | APCHK-A2 diagnostic background/geometry preparation; historical foundation. |
| Theta-87 | 0.1.0.dev602 | APCHK-A3 shared typed adequacy evaluator. |
| Theta-88 | 0.1.0.dev603 | Public AUTOPHOT science reuse boundary. |
| Theta-89 | 0.1.0.dev604 | Strict same-Comet-image FWHM provenance; later removed as an APCHK dependency. |
| Theta-90 | 0.1.0.dev605 | Provisional manual-comet transaction/candidate model. |
| Theta-91 | 0.1.0.dev606 | WCCD shadow integration and Check manual comet aperture setting. |
| Theta-92 | 0.1.0.dev607 | Actionable recommendation transaction; later preview semantics supersede auto-remeasure. |
| Theta-93 | 0.1.0.dev608 | Transient review overlay and display-only Auto-stretch for aperture check. |
| Theta-94 | 0.1.0.dev609 | R2A measurement-equivalent sweep; removed WCS/FWHM/Star Measure as required review inputs. |
| Theta-95 | 0.1.0.dev610 | Cautious observer wording: recommendation rather than claimed correct/physical aperture. |
| Theta-96 | 0.1.0.dev611 | R2C active WCCD measurement-equivalent integration; later superseded by ACF/CoG path. |
| Theta-97 | 0.1.0.dev612 | Independent-background cross-check changed from hard gate to advisory. |
| Theta-98 | 0.1.0.dev613 | Same-measurement comparator with canonical local background sampled separately for every radius. |
| Theta-99 | 0.1.0.dev614 | Live WCCD cutover to frozen same-measurement context; old “unstable independent background” wording removed. |
| Theta-100 | 0.1.0.dev615 | ACF/CoG evidence model: raw versus used increments, clipping, trend/residual and Δmag. |
| Theta-101 | 0.1.0.dev616 | Sequential ACF shadow policy and V1 candidate calibration 0.10 mag / 20%. |
| Theta-102 | 0.1.0.dev617 | Live sequential ACF/CoG cutover; Ordinary comparator removed from authoritative adequacy path. |
| Theta-103 | 0.1.0.dev618 | Observer-prefix contamination preflight prevents a contaminated large aperture from normalizing the baseline. |
| Theta-104 | 0.1.0.dev619 | Observer-facing evidence-boundary interpretation and adversarial/real-data qualification. |
| Theta-105 | 0.1.0.dev620 | Tiny-aperture CoG bootstrap. |
| Theta-106 | 0.1.0.dev621 | Compact warning UX and Comet-measure viewport preservation. |
| Theta-107 | 0.1.0.dev622 | Try suggested aperture becomes preview-only with no automatic remeasure/commit. |
| Theta-108 | 0.1.0.dev623 | Reference source_id/source_role provenance and image-local measurement overlays. |
| Theta-109 | 0.1.0.dev624 | Frozen reference-background validity/geometry/sample provenance; source-specific legacy fallback. |
| Theta-110 | 0.1.0.dev625 | Per-star reference-background quality gate; whole-observation decision returned to canonical ensemble solver. |
Qualification/audit milestones without a separate release
- Post-Theta-97 R2E1 real-FITS audit demonstrated that the 10P independent background was stable and exposed the distant frozen-sky/raw-annulus failure mode.
- Post-Theta-104 four-comet RQ covered 10P, C/2023 R1, 169P and C/2025 M2; no new product defect and no calibration change.
- Post-Theta-110 combined RQ: 13/13 explicit combined workflow PASS, 89/89 broader directly affected WCCD PASS, architecture 322 files / 802 edges / 0 SCC / 0 violations, no code change and therefore no Theta-111.
- Native GUI smoke on the qualification host remains HOLD because PyQt5 was unavailable.
Superseded behavior
Do not describe the current product as a physical coma-radius detector, plateau finder, long full-field frozen-background sweep, raw-annulus authoritative classifier, WCS/FWHM/Star-Measure-dependent APCHK, automatic remeasure after Try suggested aperture, global cross-image star-overlay restorer, or whole-WCCD abort on the first invalid reference background.
Related documentation
Current WCCD architecture · User workflow · ACF science · Qualification