Release 1.1.1: corroborate cloudy frames with local stellar attenuation
All checks were successful
Release AstroFrame / packages (push) Successful in 1m8s

This commit is contained in:
Elijah 2026-10-04 10:45:10 -07:00
parent a8e7a87b94
commit 65ded0b606
14 changed files with 731 additions and 9 deletions

114
AGENTS.md Normal file
View file

@ -0,0 +1,114 @@
# AstroFrame agent instructions
Always ask clarifying questions before proceeding when you need more information.
## Scope and installed-app protection
- Work on the source code and workspace artifacts. The user explicitly requires that development and release verification leave their locally installed AstroFrame untouched.
- Do not run installers or uninstallers, launch the installed application, apply an update to it, or change its installed files, preferences, registry entries, Start Menu shortcuts, or taskbar pins. Earlier installer testing removed the user's installation and shortcuts.
- Do not run packaged `--verify-update ... --apply` probes against the user's installation. Use fake update locators and workspace fixtures for update tests. Any installed-app interaction needs explicit new user authorization.
- Keep test output and disposable tools under `artifacts/` and `.tools/`. For preview QA, copy saved-project caches into a workspace test folder before rendering; read source exposures without modifying them.
- Publish only when a release is requested or included in the authorized task. Documentation-only changes do not need a version bump or release tag.
## Repository and version source
- Public Forgejo repository: <https://git.elijahkuntz.com/Elijah/AstroFrame>.
- Main development branch: `main`. Use the configured `origin`; inspect it rather than replacing its URL.
- `Directory.Build.props` contains the application version. `packaging/Set-Version.ps1` updates that value; it does not commit, tag, or push.
- At the time these instructions were added, `1.0.2` was published successfully. Read the current props, tags, and published releases before choosing the next version; do not assume this baseline stays current.
- Use a new stable `major.minor.patch` version without leading zeros and an exactly matching `v<version>` tag. Never overwrite a published version, its assets, or its tag. The initial `v1.0.0` tag is unpublished from a failed run; `1.0.1` was the first published release.
- Put release notes in `docs/release-notes/<version>.md`. The publisher uses this file as the release description, with a generic fallback if absent.
### Version numbering
- Use `major.minor.patch` according to the scope of the release, rather than counting commits.
- Increment the patch component (`1.1.x`) for singular or minor changes, fixes, and focused improvements.
- Increment the minor component (`1.x.1`) for major fixes, major features, or extensive changes within a bounded part of the program.
- Increment the major component (`x.1.1`) for entire-program overhauls, extensive changes impacting much of the program, or codebase rewrites.
- The examples identify the component to increment. Normally reset lower components to zero when incrementing a higher component; an explicitly requested version takes precedence. This focused algorithm correction is requested as `1.1.1`.
## Build and validation
AstroFrame is an Avalonia/.NET 8 application, published self-contained for `win-x64` and `linux-x64`. Windows uses a per-user Velopack installer; Linux uses an AppImage and optional per-user menu installer, targeting Debian and Arch. ARM is not currently required. Packages are currently unsigned.
Use PowerShell 7 for packaging and publication. Velopack CLI and application integration are pinned to `1.2.161`; its CLI requires a .NET 10 SDK/runtime. Tests require the .NET 8 runtime. Keep those versions compatible when changing dependencies.
From the repository root, keep temporary files, NuGet HTTP cache, and telemetry handling within the development scope:
```powershell
New-Item -ItemType Directory -Force artifacts/test-temp, .tools/http-cache | Out-Null
$env:TEMP = Join-Path $PWD 'artifacts/test-temp'
$env:TMP = $env:TEMP
$env:NUGET_HTTP_CACHE_PATH = Join-Path $PWD '.tools/http-cache'
$env:AVALONIA_TELEMETRY_OPTOUT = '1'
./Build.ps1 -Test
./tools/Verify-ReleaseAutomation.ps1
```
Set the temporary environment before starting test processes. `Build.ps1` also sets workspace .NET home and NuGet package directories. It prefers `.tools/dotnet/dotnet.exe` on Windows when available, otherwise `dotnet` on PATH.
- `Build.ps1 -Test` builds and runs engine and updater validation. `Verify-ReleaseAutomation.ps1` tests the publisher with a mocked API and does not publish a release.
- Render affected controls for visual changes and inspect the result; build/test success alone does not validate an overlay or layout. Use isolated source rendering that does not initialize the installed app's settings or updater.
- `Build.ps1 -Publish` means local `dotnet publish` into `artifacts/AstroFrame`; it does **not** publish a Forgejo release. `Build.ps1 -Package` invokes package creation.
- For optional local packages, use `./packaging/Release.ps1 -Platform Both`. Outputs default to `artifacts/releases/win-x64` and `artifacts/releases/linux-x64`. Existing versioned publish folders are rejected to prevent stale files. Use fresh `-PublishRoot` and `-Output` workspace paths for a rebuild.
- Linux packaging needs `mksquashfs` from `squashfs-tools`. `packaging/release.sh <version>` is the Linux-only alternative. Normal release packaging happens on the server, so a local installer build is not necessary.
## Preferred release workflow: server build and publish
Finish the change, verify it, and prepare release notes before pushing a release tag. Include all intended changes in the release commit, preserving unrelated user work. The publisher requires a clean checkout whose HEAD exists on the remote.
Example for the next release after the recorded baseline; substitute the actual new version:
```powershell
./packaging/Set-Version.ps1 -Version 1.0.3
# Create docs/release-notes/1.0.3.md and stage all intended source/test changes too.
git add Directory.Build.props docs/release-notes/1.0.3.md
git commit -m "Release 1.0.3"
git push origin main
git tag v1.0.3
git push origin v1.0.3
```
Run validation on the final versioned source before committing/tagging. Stage intended source/test files explicitly in addition to the two example files. Push the source commit before its tag. The tag must point at the committed source whose version matches it.
`.forgejo/workflows/release.yml` handles the rest:
1. A push to `main` builds and tests both platforms without publishing.
2. A push of `v<version>` builds, tests, and publishes that version. A manual workflow run on `main` is build-only; a manual run on a version tag publishes.
3. `packaging/Build-ServerRelease.ps1` checks the ref/version match, runs engine/updater/publisher validation, and cross-packages both platforms into `artifacts/server-release/packages`.
4. The job renders the packaged Linux settings page under Xvfb, using AppImage extraction mode.
5. `packaging/Publish.ps1 -Directory artifacts/server-release/packages -Publish -RequireTag` validates both channels, creates a draft, uploads all assets, and makes the release public only after every upload succeeds.
Known runner requirements that previously caused failures:
- Runner name: `unraid-runner`; advertised label: `ubuntu-latest`. `runs-on` selects the **label**, not the runner name. The `docker` label did not match this runner.
- Job container: `node:22-bookworm`. This Linux runner can create both the Windows Setup executable and Linux AppImage using Velopack cross-packaging.
- Checkout is pinned to `https://data.forgejo.org/actions/checkout@v4.4.0`. Runner v6.4.0 supports Node 20 actions; checkout v6 requires Node 24 and failed. Check runner compatibility before upgrading actions.
- The workflow installs the .NET 10 SDK, .NET 8 runtime, PowerShell 7.4.20, `squashfs-tools`, and Linux desktop/Xvfb dependencies. Preserve these prerequisites.
- The publisher uses Forgejo's repository-scoped `${{ github.token }}` as `ASTROFRAME_RELEASE_TOKEN`; no personal token is needed for the server workflow.
- Server Linux smoke tests do not establish Windows UI acceptance or separate Arch/Debian desktop acceptance. Report validation accurately and respect the installed-app protection rule.
## Assets required for automatic updates
Installed applications use Velopack's public Gitea release source, with separate `win-x64` and `linux-x64` channels. They check published stable releases, not Git commits, and never pull source code or need a Git client. No repository write token belongs in the application.
Each public release must include **both** platforms. Uploading only an installer breaks the automatic-update workflow. Normal builds produce 13 assets:
- For each channel: `releases.<channel>.json`, `RELEASES-<channel>`, `assets.<channel>.json`, and `AstroFrame-<version>-<channel>-full.nupkg`.
- Windows: `AstroFrame-win-x64-Setup.exe`.
- Linux: `AstroFrame-linux-x64.AppImage`, `install-linux.sh`, `astroframe.png`, and `AstroFrame-<version>-linux-x64.zip`.
- Include any generated delta packages referenced by the feed. Full packages support updates from older versions without deltas. Keep older published releases available.
After pushing the tag, check the **tagged** Actions run through successful publication, then verify the public release is stable, not a draft, contains both channels and expected assets, and both feeds reference the new full package with matching size/checksum. A green `main` build is not proof of publication. Inspect the public API/assets or use an isolated fake locator; do not update the user's installed copy as a test.
The agreed user experience is: check on launch and every six hours, notify when a newer stable release exists, then download/install/restart only when the user clicks **Install update and restart**. Settings shows version/status, automatic-check preference, and manual check. Installation is blocked during processing or saving and saves the project before restart. Source-tree builds are not self-updating. Preserve this behavior when modifying updates.
## Publisher recovery and local fallback
- `packaging/Publish.ps1` verifies version, clean source, remote commit, matching tag when required, package sizes, and SHA256 before publication. The correct commit API route is `/api/v1/repos/Elijah/AstroFrame/git/commits/<sha>`; `/commits/<sha>` returned 404.
- A failed upload leaves a draft, invisible to installed apps. Resuming is safe only with identical package bytes; existing uploads are downloaded and compared by SHA256. A fresh rebuild may differ and cause a conflict.
- Do not delete published releases or force-move tags to fix publication. Ask before deleting conflicting draft assets/releases or moving an unpublished tag; prefer a new version when requested by the user.
- Local fallback: supply a repository-write token through the process environment as `ASTROFRAME_RELEASE_TOKEN`, then run `./packaging/Publish.ps1 -Directory <package-root> -RequireTag` to create/resume a draft. Add `-Publish` only when immediate publication is authorized. Use the matching tagged checkout and both channel folders.
- Never commit, print, embed, or persist a release token in scripts or application files.
See [docs/releases.md](docs/releases.md) and the scripts/workflow above for implementation details. Treat those files as the current authority if these instructions become stale, and update this runbook when changing the workflow.

View file

@ -1,6 +1,6 @@
<Project> <Project>
<PropertyGroup> <PropertyGroup>
<Version>1.1.0</Version> <Version>1.1.1</Version>
<Authors>Elijah Kuntz</Authors> <Authors>Elijah Kuntz</Authors>
<Product>AstroFrame</Product> <Product>AstroFrame</Product>
</PropertyGroup> </PropertyGroup>

View file

@ -0,0 +1,65 @@
# Algorithm comparison for 1.1.1
Validation performed October 4, 2026. The baseline is the unchanged source at `a8e7a87` (1.1.0), with its validation executable preserved before changing the engine. Both algorithms processed original Sample Data FITS files into separate workspace artifact directories. Source files and the installed application were untouched.
## Results
| Session | Frames | 1.1.0 Keep / Review / Reject | 1.1.1 Keep / Review / Reject |
|---|---:|---|---|
| NGC 1432, September 20 | 490 | 358 / 55 / 77 | **358 / 32 / 100** |
| NGC 1432, October 2 | 563 | 547 / 16 / 0 | **547 / 16 / 0** |
| NGC 7000, August 17 | 147 | 141 / 6 / 0 | **141 / 6 / 0** |
| IC 1805, August 26 | 309 | 301 / 3 / 5 | **301 / 2 / 6** |
| Total | 1,509 | 1,347 / 80 / 82 | **1,347 / 56 / 106** |
All 24 differences are Review → Reject: 23 in the requested NGC 1432 session and one in IC 1805. No Keep or existing Reject changed verdict; no new Keep was introduced. The requested session's review workload falls by 23 frames (41.8%). All four supplied cloud examples now Reject. The two supplied clear-looking morphology examples remain Review because original-source stellar crops show shape differences and the matched-star evidence remains near the practical tolerance.
## What changed
- Missing reference stars are distinguished from failed geometry. Tight-match count, residuals and spatial distribution can validate the surviving star field. The nearly opaque example has 164 tight correspondences and a median residual of 0.423 original pixels, despite a detected-star match fraction of only 0.297.
- Broad coherent sky anomalies can be corroborated by attenuation measured specifically at reference apertures under positive excess-sky cells. Nearby comparisons use the same apertures, with reference scatter and local sky/background photometry uncertainty. This avoids dilution by unaffected stars or coarse whole-image regions.
- Matched bright stars cross-check elongation, with enough nearby reference frames and a conservative effect/uncertainty margin. An initial candidate Keep was not retained after original-source crops showed measurable elongation. The final comparison preserves every prior Keep/Reject membership.
- The analysis-version identifier changed, invalidating old automatic measurements while retaining human decisions. Display and binary-cache formats are unchanged; reanalysis generates the new diagnostics.
## Inspection of the differences
Every final changed exposure was inspected in source-derived previews and against its measured local attenuation. Each NGC 1432 difference shows a cloud band or patch, including the four supplied examples. A positive cloud structure and a consistent loss in the affected stellar apertures support these rejections. The changed IC 1805 frame (`20260826-012414`) shows cloud entering the lower part of the field; subsequent frames show that cloud moving across the right side. Its affected-aperture flux is approximately 73% of nearby reference flux. Adjacent preceding clear frames preserve the persistent nebula and remain Keep. This is transient cloud evidence rather than the nebula itself being rejected.
Additional visual controls include nine fixed-time samples in each nebula session, the nine weakest pre-existing NGC 1432 rejects, and the strongest remaining cloud Review frames. NGC 7000 has **zero decision changes across all 147 exposures**. All 301 IC 1805 Keeps and five existing Rejects are unchanged. The second NGC 1432 night also has zero changes across 563 exposures.
These inspections are an assistant assessment supported by measurements, not independent owner labels for the entire dataset. They support this focused improvement and found no incorrect new rejection in the audited differences. They do not establish a zero-error guarantee on future targets or substitute for controlled stacks. We retain Review where geometry, attenuation, or morphology is uncertain.
## Regression and release checks
- The focused regression loop initially failed three checks on the old policy and passes on the new policy: accurate geometry despite star disappearance; broad sky structure with local attenuation; and detection-cohort morphology corroboration.
- The full engine suite passes **77 checks**, including persistent nebulosity and gradients, prolonged moving cloud, no stable clear reference, low geometry confidence, insufficient/noisy apertures, measurable elongation, source-safe export, preview independence, and upgrade preservation of human decisions.
- A checked-in 46-frame measured fixture preserves four owner cloud examples and two owner-confirmed clear controls, without requiring the original FITS files in CI.
- All **9/9 owner-labeled real examples** pass. The **40/40 previously screened cloud-only samples available in these tested sessions** also pass. Those samples were previously used in tuning, so they are not an independent accuracy estimate.
- The updater and mocked publisher suites run on the final versioned source before publication. The publisher tests do not publish or interact with the installed app.
## Additional spatial and trail experiments
The three suggested topics were evaluated before publication. They are promising directions, but the tested prototypes did not establish a further safe verdict improvement, so they are not additional production rejection rules in 1.1.1.
**Cloud-scale smoothing and multiple scales.** All 1,509 cached aligned sky maps were evaluated at the existing cell scale and with 3×3 and 7×7 box filters. Each scale used scatter measured from equally filtered clear-reference residuals, retaining the original noise floor and overlap mask; uncertainty was not reduced by assuming neighboring cells were independent. These are bounded box-filter experiments, not an exhaustive evaluation of Gaussian, wavelet, or adaptive filters. The existing significance, coverage and connected-area gates were applied for comparison.
- No previously nonsuspicious Keep acquired a cloud candidate at either added scale.
- A 3×3 filter reduced suspicious sky Review candidates from 15 to 14 in NGC 1432, without introducing a new candidate; the removed candidate is from the otherwise unchanged October night.
- A 7×7 filter reduced NGC 1432 candidates from 15 to 11, and IC 1805 Review candidates from two to one. It also removed the sky-candidate gate from two existing IC 1805 Rejects (`20260826-012650` and `20260826-012721`), whose source-derived previews show cloud entering at the edge. Their affected-aperture flux ratios are approximately 60% and 59%.
- Requiring agreement at all scales could suppress real partial clouds. Taking their union retained the original candidate membership, giving no additional verified accuracy gain. Connected regions and affected-star corroboration were therefore retained without replacing the fine-scale evidence or promoting smoothed-away warnings to Keep.
Filtering a background mesh is an established measurement technique, but does not itself establish a safe frame rejection policy; [Photutils Background2D](https://photutils.readthedocs.io/en/latest/api/photutils.background.Background2D.html) documents mesh filtering independently of this classifier.
**Directional trails.** Original FITS stellar cores were measured for 104 selected exposures: morphology Review frames, nearby representative Keeps, and nine fixed-time Keep controls in each target. The exploratory measurement reused positive core moments, excluded detected close neighbors within ten superpixels, and computed the ellipticity-weighted double-angle resultant for isolated stars with ellipticity at least 0.15. This measures directional consistency from zero to one; it is not a calibrated PSF fit or an independent tracking-error label.
The median resultant was 0.987 for 33 NGC 1432 Reviews versus 0.874 for 41 Keeps, and 0.963 for six NGC 7000 Reviews versus 0.653 for 15 Keeps. However, nine IC 1805 Keep controls already had a median resultant of 0.930 and median isolated-star ellipticity of 0.280. Thus a high directional consistency alone does not distinguish tracking failure from the session's normal optics/sampling. The Reviews are already captured by elongation; there are no independently labeled trail acceptance boundaries in these samples to justify automatically discarding them. A future rule should compare matched-star shape tensors after accounting for frame rotation, spatial optical patterns, sampling and fit uncertainty, then validate against explicit acceptable/unacceptable tracking labels. This release retains conservative matched-star elongation corroboration and Review for unresolved shape changes.
[Siril's Dynamic PSF documentation](https://siril.readthedocs.io/en/latest/Dynamic-PSF.html) describes rotation-angle fitting and notes that elliptical stars can arise from optical aberrations, tracking, or blended stars. Its fitted Gaussian/Moffat profiles are more sophisticated than AstroFrame's core moments; we do not claim identical measurement accuracy or that the suggested cloud classifier is documented as a Siril feature.
Workspace experiment sources/results are `SpatialTopics.cs`, `spatial-topics.json`, `DirectionTopics.cs`, and `direction-topics.json` under `artifacts/algorithm-1.1.1/`; `smoothed-loss-1.png` contains the seven sky-candidate losses at the largest scale. These research experiments do not modify production measurements, the source FITS files, or the comparison verdicts above.
## Reproduction and artifacts
Baseline/new analysis uses the validation executable's `analyze <source> <output> --fresh` mode, with temporary/cache environment variables confined to the workspace as prescribed in AGENTS.md. The baseline executable is preserved at `artifacts/algorithm-1.1.1/baseline-cli/`; baseline and final projects, catalog caches, previews, manifests and reports are under `baseline-*` and `new-*` sibling folders. Final policy refinements reused identical stellar/sky/aperture measurements and rebuilt their reference diagnostics. All 1,509 source size/modification-time fingerprints remain unchanged.
`artifacts/algorithm-1.1.1/comparison.json` records the full per-exposure differences. `changes-1.png` through `changes-3.png` show every final changed frame; `nebula-controls-*.png` and `nebula-cloud-transition-1.png` show the nebula controls and the IC 1805 transition. `keep-stars.png` documents why the exploratory new Keep was retained as Review in the final policy. Compact session counts and per-change measurements are checked in as [quality-1.1.1-validation.json](diagnostics/quality-1.1.1-validation.json).

View file

@ -6,6 +6,8 @@ The supplied 1,920 originals cover four targets and five sessions, with 2160×38
The first algorithm failed real owner-reviewed examples: cloud structure inflated detected-star counts; flux only sampled surviving detections; rolling baselines absorbed prolonged cloud; a small ellipticity deviation reviewed a clear image. The v2 design replaces the cloud measurements, reference construction and classification policy. Rejection counts are not a target or a validation criterion. The first algorithm failed real owner-reviewed examples: cloud structure inflated detected-star counts; flux only sampled surviving detections; rolling baselines absorbed prolonged cloud; a small ellipticity deviation reviewed a clear image. The v2 design replaces the cloud measurements, reference construction and classification policy. Rejection counts are not a target or a validation criterion.
Version 1.1.1 refines this policy with geometric fit diagnostics, local cloud-aperture corroboration, and matched-star morphology checks. The original v2 results remain a baseline; see [the 1.1.1 comparison](algorithm-1.1.1-validation.md) for the measurements and limits of the new validation.
## Components ## Components
- **Engine:** FITS reader, independent `IFrameDetector` measurements, background reference construction, aperture photometry, named decision rules, persistence and export. No GUI, Python runtime, AI or GPU dependency. - **Engine:** FITS reader, independent `IFrameDetector` measurements, background reference construction, aperture photometry, named decision rules, persistence and export. No GUI, Python runtime, AI or GPU dependency.
@ -33,10 +35,13 @@ REJECT evidence routes are explicit:
- Strong spatial structure: at least 25× reference scatter at the 95th percentile, at least 8% affected overlap and a connected area of at least 3.5%. A lone sky signal must be overwhelming before discarding integration; weaker coherent structure needs stellar corroboration or routes to REVIEW. - Strong spatial structure: at least 25× reference scatter at the 95th percentile, at least 8% affected overlap and a connected area of at least 3.5%. A lone sky signal must be overwhelming before discarding integration; weaker coherent structure needs stellar corroboration or routes to REVIEW.
- Moderate coherent spatial structure plus substantial global flux loss, regional forced-flux loss or detected-star loss. - Moderate coherent spatial structure plus substantial global flux loss, regional forced-flux loss or detected-star loss.
- Coherent structure at least 8× reference scatter, at least 8% affected overlap, and a connected anomalous area of at least 3.5%, corroborated by positive excess-sky coverage of at least 8% and a positive connected area of at least 3.5%. At least 20 reference apertures in two regions inside that excess sky must show below 90% of nearby same-aperture flux, with at least three times the larger of reference scatter, a 3% floor, and propagated local photometry uncertainty. Affected-aperture comparisons use the nearest 40 reference members, excluding the tested exposure, and require at least 12 valid comparison frames.
- Severe global flux loss (below 60% of clear-frame baseline) with substantial detected-star loss. - Severe global flux loss (below 60% of clear-frame baseline) with substantial detected-star loss.
- Severe combined stellar width and ellipticity degradation, or almost complete star disappearance relative to a rich reference field. - Severe combined stellar width and ellipticity degradation, or almost complete star disappearance relative to a rich reference field.
Automatic rejection requires enough comparable frames, valid time and adequate aligned overlap. Ordinary rejection also requires reliable matching/aperture measurements. Ambiguous structure, moderate flux loss, substantial single-metric morphology changes and uncertain measurements route to REVIEW. Minor ellipticity fluctuations and a high global background alone do not trigger review. Automatic rejection requires enough comparable frames, valid time and adequate aligned overlap. Ordinary spatial rejection also requires verified geometry: at least 20 tight correspondences, at least 80% of correspondences within the width-scaled tight tolerance, three regions with at least three tight matches each, and a spread of at least 15% of the image width and height. The median residual must be at most the larger of 0.5 original pixels and 0.2× reference width; the 95th percentile must be at most the larger of one original pixel and 0.5× reference width. Transform scale must remain between 0.975 and 1.025. Missing stars alone no longer veto a cloud rejection when geometry and aperture coverage are verified; low star survival still cannot certify KEEP. Catalog-only classification retains its previous match-fraction safeguard.
An existing elongation flag can be dismissed only when at least 50 matched bright reference stars and 12 nearby reference frames provide a precise, small shape change. Its median change plus three estimated standard errors must be below 0.10, and the uncertainty must not exceed 0.025. Missing pairs cannot dismiss a warning. Matched stars use the same reference identities rather than independently selected detection populations. This remains moment-based photometry, not a new PSF-fitting implementation. Ambiguous structure, moderate flux loss, substantial single-metric morphology changes and uncertain measurements route to REVIEW. Minor ellipticity fluctuations and a high global background alone do not trigger review.
These significance and effect-size limits are **engineering policy, not calibrated probabilities or guarantees**. Reference scatter, noise, star population, flux, seeing and background are measured separately per session. No fixed acceptable star count or absolute sky brightness is used. The score exists only for sorting. These significance and effect-size limits are **engineering policy, not calibrated probabilities or guarantees**. Reference scatter, noise, star population, flux, seeing and background are measured separately per session. No fixed acceptable star count or absolute sky brightness is used. The score exists only for sorting.

View file

@ -0,0 +1,360 @@
{
"date": "2026-10-04",
"baselineCommit": "a8e7a87b94d0f8046209e9c937bd83bf4b811f10",
"baselineEngine": "0.2.0-spatial-reference-1",
"newEngine": "0.2.1-spatial-reference-2",
"sessions": [
{
"session": "NGC 1432 \u00b7 2026-09-20 06:50 \u00b7 1 \u00b7 IRCUT",
"frames": 490,
"old": {
"Keep": 358,
"Review": 55,
"Reject": 77
},
"new": {
"Keep": 358,
"Review": 32,
"Reject": 100
}
},
{
"session": "NGC 1432 \u00b7 2026-10-02 06:00 \u00b7 2 \u00b7 IRCUT",
"frames": 563,
"old": {
"Keep": 547,
"Review": 16,
"Reject": 0
},
"new": {
"Keep": 547,
"Review": 16,
"Reject": 0
}
},
{
"session": "NGC 7000 \u00b7 2026-08-17 04:33 \u00b7 1 \u00b7 LP",
"frames": 147,
"old": {
"Keep": 141,
"Review": 6,
"Reject": 0
},
"new": {
"Keep": 141,
"Review": 6,
"Reject": 0
}
},
{
"session": "IC 1805 \u00b7 2026-08-26 08:00 \u00b7 1 \u00b7 LP",
"frames": 309,
"old": {
"Keep": 301,
"Review": 3,
"Reject": 5
},
"new": {
"Keep": 301,
"Review": 2,
"Reject": 6
}
}
],
"changes": [
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260919-235506.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 21.93785621095097,
"affectedFluxRatio": 0.8636848120909967,
"affectedFluxDeviation": 4.543839596966776,
"alignmentMedianPixels": 0.2631228145939164,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260919-235537.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 14.205363402591354,
"affectedFluxRatio": 0.8845403044429598,
"affectedFluxDeviation": 3.848656518568007,
"alignmentMedianPixels": 0.25334731221934753,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260919-235608.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 19.304613846485182,
"affectedFluxRatio": 0.8423382721761337,
"affectedFluxDeviation": 5.255390927462213,
"alignmentMedianPixels": 0.25116491416396974,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-000300.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 21.99244300903803,
"affectedFluxRatio": 0.83036711072136,
"affectedFluxDeviation": 5.6544296426213325,
"alignmentMedianPixels": 0.23723050615616165,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-000331.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 18.008937347417586,
"affectedFluxRatio": 0.8425858408547843,
"affectedFluxDeviation": 5.247138638173857,
"alignmentMedianPixels": 0.23176112388221165,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-001125.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 8.238866042072468,
"affectedFluxRatio": 0.8453711916213303,
"affectedFluxDeviation": 5.154293612622321,
"alignmentMedianPixels": 0.2004272687660229,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-001156.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 14.060493454157665,
"affectedFluxRatio": 0.873622460980021,
"affectedFluxDeviation": 4.2125846339993,
"alignmentMedianPixels": 0.21057367839555907,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-002817.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 19.741168661451677,
"affectedFluxRatio": 0.8912902562183401,
"affectedFluxDeviation": 3.6236581260553278,
"alignmentMedianPixels": 0.18335484479463354,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-003713.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 231.33479320777224,
"affectedFluxRatio": 0.039192529060025114,
"affectedFluxDeviation": 32.026915697999165,
"alignmentMedianPixels": 0.4229262918099691,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-004304.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 15.49719600393806,
"affectedFluxRatio": 0.8461550650098837,
"affectedFluxDeviation": 5.128164499670544,
"alignmentMedianPixels": 0.16358180745346368,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-004624.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 15.391777248244273,
"affectedFluxRatio": 0.8897937985360267,
"affectedFluxDeviation": 3.6735400487991123,
"alignmentMedianPixels": 0.17600984075984275,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-004726.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 18.57963411527099,
"affectedFluxRatio": 0.8705488825072726,
"affectedFluxDeviation": 4.315037249757581,
"alignmentMedianPixels": 0.17524893834549105,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-005235.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 24.15637588387764,
"affectedFluxRatio": 0.8709361884911153,
"affectedFluxDeviation": 4.302127050296157,
"alignmentMedianPixels": 0.14926813111727083,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-005421.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 8.17568631153736,
"affectedFluxRatio": 0.8956752750757109,
"affectedFluxDeviation": 3.477490830809639,
"alignmentMedianPixels": 0.14261613075981966,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-010318.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 13.187685772318256,
"affectedFluxRatio": 0.8887345385372606,
"affectedFluxDeviation": 3.708848715424648,
"alignmentMedianPixels": 0.16306293375502223,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-011113.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 18.169921576052825,
"affectedFluxRatio": 0.8396524330778192,
"affectedFluxDeviation": 5.344918897406026,
"alignmentMedianPixels": 0.16379873442258716,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-012632.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 19.52391631691044,
"affectedFluxRatio": 0.8490009027116143,
"affectedFluxDeviation": 5.033303242946189,
"alignmentMedianPixels": 0.12881799374449404,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-012818.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 17.723635511036186,
"affectedFluxRatio": 0.8455906425280958,
"affectedFluxDeviation": 5.146978582396807,
"alignmentMedianPixels": 0.15677678144061136,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-013137.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 10.824700198303587,
"affectedFluxRatio": 0.8297659241686153,
"affectedFluxDeviation": 5.674469194379491,
"alignmentMedianPixels": 0.13213426121083371,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-013310.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 9.551977498073267,
"affectedFluxRatio": 0.8809177425716295,
"affectedFluxDeviation": 3.969408580945685,
"alignmentMedianPixels": 0.137763541386481,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-013746.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 19.91122796256937,
"affectedFluxRatio": 0.7445523676711074,
"affectedFluxDeviation": 8.51492107762975,
"alignmentMedianPixels": 0.13331108959050958,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-013817.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 21.69705495844935,
"affectedFluxRatio": 0.8457382705969594,
"affectedFluxDeviation": 5.14205764676802,
"alignmentMedianPixels": 0.16633085057987979,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_NGC 1432_30.0s_IRCUT_20260920-023920.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 13.414930894720293,
"affectedFluxRatio": 0.867893857149813,
"affectedFluxDeviation": 4.403538095006232,
"alignmentMedianPixels": 0.1296821468757567,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
},
{
"file": "Light_IC 1805_30.0s_LP_20260826-012414.fit",
"before": "Review",
"after": "Reject",
"skySignificance": 9.482428788618922,
"affectedFluxRatio": 0.7321572057985136,
"affectedFluxDeviation": 7.7859449008662605,
"alignmentMedianPixels": 0.34200350430379745,
"visualAssessment": "Transient cloud structure corroborated by affected-aperture attenuation"
}
],
"ownerExamples": {
"passes": 9,
"total": 9
},
"previouslyScreenedExamples": {
"passes": 40,
"total": 40,
"independent": false
},
"newKeeps": 0,
"keepToReject": 0,
"rejectToKeep": 0,
"sourceFingerprints": 1509,
"limitation": "Assistant inspection of all differences; no independently owner-labeled full session or controlled stack comparison.",
"checks": {
"engine": 77,
"updater": 23,
"mockedPublisher": 10
},
"additionalExperiments": {
"productionRulesAdded": false,
"spatial": {
"frames": 1509,
"cellScales": [
1,
3,
7
],
"newKeepCloudCandidates": 0,
"ngc1432SkyReviewCandidates": [
15,
14,
11
],
"ic1805SkyReviewCandidates": [
2,
2,
1
],
"ic1805ExistingRejectSkyCandidatesLostAtLargestScale": 2,
"decision": "Retain existing connected-cell evidence; smoothing did not establish additional safe verdict improvements."
},
"directional": {
"frames": 104,
"coherenceMedian": {
"ngc1432Review": 0.987,
"ngc1432Keep": 0.874,
"ngc7000Review": 0.963,
"ngc7000Keep": 0.653,
"ic1805Keep": 0.93
},
"decision": "Retain morphology review and matched-star corroboration; high coherence also exists in normal controls and trail acceptance labels are unavailable."
}
}
}

View file

@ -0,0 +1,12 @@
# AstroFrame 1.1.1
Improves conservative cloud rejection without promoting uncertain exposures to Keep.
- Verifies stellar alignment using positional residuals and distributed tight matches, so severe cloud cover does not automatically defeat rejection by hiding reference stars.
- Corroborates broad cloud structure with brightness loss measured at the same stars in the affected area, using nearby references and photometry uncertainty.
- Cross-checks elongation against matched reference stars; measurable or uncertain shape changes continue to require Review.
- Preserves human decisions when reanalysis refreshes the new measurements.
In the supplied 490-frame NGC 1432 session, Review decreases from 55 to 32 and 23 visibly clouded exposures become Reject. All previous Keeps and Rejects remain unchanged. NGC 7000's 147-frame nebula session is unchanged; IC 1805 changes one cloud-contaminated Review to Reject while preserving every Keep. A second 563-frame NGC 1432 night is unchanged. Every changed decision was inspected; these results support the tested improvement rather than a guarantee for every future dataset.
Engine, updater and release automation checks pass on the versioned source. Old automatic measurements require reanalysis; manual labels are retained. Windows and Linux update channels are included.

View file

@ -48,6 +48,15 @@ public static class Classification
bool moderateStructure=spatial&&f.CloudSignificance>=5&&f.CloudCoverage>=.02&&connected>=.01; bool moderateStructure=spatial&&f.CloudSignificance>=5&&f.CloudCoverage>=.02&&connected>=.01;
bool fluxLoss=fluxZ>=4&&fluxRatio<.78,starLoss=countZ>=3&&countRatio<.75; bool fluxLoss=fluxZ>=4&&fluxRatio<.78,starLoss=countZ>=3&&countRatio<.75;
bool strongFlux=fluxZ>=5&&fluxRatio<.60&&starLoss; bool strongFlux=fluxZ>=5&&fluxRatio<.60&&starLoss;
bool localCloudLoss=spatial&&f.CloudSignificance>=8&&f.CloudCoverage>=.08&&connected>=.035&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud positive coverage")>=.08&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud positive connected coverage")>=.035&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud aperture samples")>=20&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud aperture regions")>=2&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud aperture flux ratio")<.90&&
f.AdditionalMeasurements.GetValueOrDefault("Cloud aperture flux deviation")>=3;
if(localCloudLoss)f.Evidence.Add(new("Cloud aperture flux",f.AdditionalMeasurements["Cloud aperture flux ratio"],1,f.AdditionalMeasurements["Cloud aperture flux deviation"],
$"Coherent excess sky coincides with stellar flux {f.AdditionalMeasurements["Cloud aperture flux ratio"]:P0} of the same-aperture reference ({f.AdditionalMeasurements["Cloud aperture samples"]:0} apertures); uncertainty {f.AdditionalMeasurements.GetValueOrDefault("Cloud aperture flux uncertainty"):P1}"));
bool regionLoss=false; bool regionLoss=false;
for(int r=0;r<12;r++) for(int r=0;r<12;r++)
{ {
@ -68,19 +77,32 @@ public static class Classification
if(fluxLoss)f.Evidence.Add(new("Stellar flux",f.Transparency,f.Baselines["Stellar flux"],fluxZ,$"Forced stellar flux reduced to {fluxRatio:P0} of nearby clear-frame baseline; lost detections remain in the measurement")); if(fluxLoss)f.Evidence.Add(new("Stellar flux",f.Transparency,f.Baselines["Stellar flux"],fluxZ,$"Forced stellar flux reduced to {fluxRatio:P0} of nearby clear-frame baseline; lost detections remain in the measurement"));
if(starLoss)f.Evidence.Add(new("Star count",f.StarCount,f.Baselines["Star count"],countZ,$"Detected stars reduced to {countRatio:P0} of clear-frame baseline")); if(starLoss)f.Evidence.Add(new("Star count",f.StarCount,f.Baselines["Star count"],countZ,$"Detected stars reduced to {countRatio:P0} of clear-frame baseline"));
bool widthBad=widthZ>4&&widthRatio>1.25,shapeBad=shapeZ>4&&f.Ellipticity>f.Baselines["Ellipticity"]+.15; bool widthBad=widthZ>4&&widthRatio>1.25,shapeBad=shapeZ>4&&f.Ellipticity>f.Baselines["Ellipticity"]+.15;
if(shapeBad&&f.AdditionalMeasurements.GetValueOrDefault("Matched shape samples")>=50&&
f.AdditionalMeasurements.GetValueOrDefault("Matched shape reference frames")>=12&&
f.AdditionalMeasurements.GetValueOrDefault("Matched shape uncertainty")>0&&
f.AdditionalMeasurements["Matched shape uncertainty"]<=.025&&
f.AdditionalMeasurements.GetValueOrDefault("Matched shape change")+3*f.AdditionalMeasurements["Matched shape uncertainty"]<.10)
{
shapeBad=false;
f.Evidence.Add(new("Matched stellar shape",f.AdditionalMeasurements["Matched shape change"],.10,0,"Matched reference stars do not corroborate the detection-cohort elongation flag"));
}
if(widthBad)f.Evidence.Add(new("FWHM",f.Fwhm,f.Baselines["FWHM"],widthZ,$"Stellar width {widthRatio:0.00} × clear-frame baseline")); if(widthBad)f.Evidence.Add(new("FWHM",f.Fwhm,f.Baselines["FWHM"],widthZ,$"Stellar width {widthRatio:0.00} × clear-frame baseline"));
if(shapeBad)f.Evidence.Add(new("Ellipticity",f.Ellipticity,f.Baselines["Ellipticity"],shapeZ,$"Substantial stellar elongation ({f.Ellipticity:0.00} vs {f.Baselines["Ellipticity"]:0.00})")); if(shapeBad)f.Evidence.Add(new("Ellipticity",f.Ellipticity,f.Baselines["Ellipticity"],shapeZ,$"Substantial stellar elongation ({f.Ellipticity:0.00} vs {f.Baselines["Ellipticity"]:0.00})"));
bool optics=widthBad&&shapeBad&&widthRatio>1.65; bool optics=widthBad&&shapeBad&&widthRatio>1.65;
bool catastrophic=f.StarCount<3&&f.Baselines["Star count"]>50; bool catastrophic=f.StarCount<3&&f.Baselines["Star count"]>50;
bool cloud=strongStructure||moderateStructure&&(fluxLoss||regionLoss||starLoss)||strongFlux; bool cloud=strongStructure||moderateStructure&&(fluxLoss||regionLoss||starLoss)||strongFlux||localCloudLoss;
bool weak=f.MatchFraction<.35||f.Transparency<=0||f.Fwhm<=0||(spatial&&f.AdditionalMeasurements.GetValueOrDefault("Forced flux samples")<20); bool lowMatches=f.MatchFraction<.35;
bool geometryWeak=spatial?f.AdditionalMeasurements.GetValueOrDefault("Alignment verified")<=0:lowMatches;
bool weak=geometryWeak||f.Transparency<=0||f.Fwhm<=0||(spatial&&f.AdditionalMeasurements.GetValueOrDefault("Forced flux samples")<20);
bool skyNoise=bgZ>5&&noiseZ>4&&Ratio("Background",f.Background)>1.4; bool skyNoise=bgZ>5&&noiseZ>4&&Ratio("Background",f.Background)>1.4;
if(skyNoise)f.Evidence.Add(new("Sky noise",f.Noise,f.Baselines["Noise"],noiseZ,"Background and sky noise are substantially above the clear-frame baseline")); if(skyNoise)f.Evidence.Add(new("Sky noise",f.Noise,f.Baselines["Noise"],noiseZ,"Background and sky noise are substantially above the clear-frame baseline"));
bool suspicious=moderateStructure||fluxLoss||regionLoss||widthBad||shapeBad||skyNoise||starLoss; bool suspicious=moderateStructure||fluxLoss||regionLoss||widthBad||shapeBad||skyNoise||starLoss||lowMatches;
f.Automatic=reliable&&(catastrophic||!weak&&(cloud||optics))?Verdict.Reject:!reliable||weak||suspicious||f.Warning!=""?Verdict.Review:Verdict.Keep; f.Automatic=reliable&&(catastrophic||!weak&&(cloud||optics))?Verdict.Reject:!reliable||weak||suspicious||f.Warning!=""?Verdict.Review:Verdict.Keep;
f.Cause=!reliable?"Insufficient reliable session reference":catastrophic?"Severe disappearance of reference stars":weak?"Uncertain alignment / reference photometry":cloud?"Likely cloud / transparency loss":optics?"Likely tracking or focus failure":suspicious?"Quality change needs review":"Consistent with the clean session reference"; f.Cause=!reliable?"Insufficient reliable session reference":catastrophic?"Severe disappearance of reference stars":weak?"Uncertain alignment / reference photometry":cloud?"Likely cloud / transparency loss":optics?"Likely tracking or focus failure":suspicious?"Quality change needs review":"Consistent with the clean session reference";
if(!reliable)f.Evidence.Add(new("Reference",local.Length,12,0,"Automatic rejection disabled: insufficient clear comparison frames, timestamp or aligned overlap")); if(!reliable)f.Evidence.Add(new("Reference",local.Length,12,0,"Automatic rejection disabled: insufficient clear comparison frames, timestamp or aligned overlap"));
if(weak)f.Evidence.Add(new("Matching",f.MatchFraction,.35,0,"Uncertain star alignment or reference aperture coverage; inspect manually")); if(weak)f.Evidence.Add(new("Matching",f.MatchFraction,.35,0,"Unverified geometric alignment or insufficient reference photometry; inspect manually"));
else if(lowMatches)f.Evidence.Add(new("Alignment",f.AdditionalMeasurements.GetValueOrDefault("Alignment tight matches"),20,0,
$"Star survival is low, but {f.AdditionalMeasurements.GetValueOrDefault("Alignment tight matches"):0} tight matches verify geometry; median residual {f.AdditionalMeasurements.GetValueOrDefault("Alignment median residual pixels"):0.00} source pixels"));
if(f.Warning!="")f.Evidence.Add(new("Metadata",0,0,0,f.Warning)); if(f.Warning!="")f.Evidence.Add(new("Metadata",0,0,0,f.Warning));
if(!widthBad)f.Evidence.Add(new("FWHM",f.Fwhm,f.Baselines["FWHM"],0,"Stellar width is within the practical tolerance of nearby clear frames")); if(!widthBad)f.Evidence.Add(new("FWHM",f.Fwhm,f.Baselines["FWHM"],0,"Stellar width is within the practical tolerance of nearby clear frames"));
if(f.Evidence.Count==1)f.Evidence.Add(new("Reference",f.Transparency,f.Baselines["Stellar flux"],0,"Aligned sky structure, stellar flux and morphology show no substantial degradation")); if(f.Evidence.Count==1)f.Evidence.Add(new("Reference",f.Transparency,f.Baselines["Stellar flux"],0,"Aligned sky structure, stellar flux and morphology show no substantial degradation"));

View file

@ -81,19 +81,40 @@ public static class Matching
{ {
frame.Transparency=0;frame.MatchFraction=0;frame.RegionFlux=new double[12];frame.RegionLoss=new double[12];frame.RegionMatches=new int[12]; frame.Transparency=0;frame.MatchFraction=0;frame.RegionFlux=new double[12];frame.RegionLoss=new double[12];frame.RegionMatches=new int[12];
frame.Alignment=[];frame.ReferenceId=reference.Id; frame.Alignment=[];frame.ReferenceId=reference.Id;
frame.MatchedShape=Enumerable.Repeat(double.NaN,Math.Min(600,reference.Stars.Count)).ToArray();
foreach(string key in new[]{"Alignment verified","Alignment matches","Alignment tight matches","Alignment median residual pixels","Alignment p95 residual pixels","Alignment occupied regions"})frame.AdditionalMeasurements[key]=0;
var tr=Align(reference,frame);if(tr==null)return; var tr=Align(reference,frame);if(tr==null)return;
frame.Alignment=[tr.A,tr.B,tr.Tx,tr.Ty]; frame.Alignment=[tr.A,tr.B,tr.Tx,tr.Ty];
double aspect=(double)reference.Height/reference.Width;var grid=new Grid(frame.Stars,aspect); double aspect=(double)reference.Height/reference.Width;var grid=new Grid(frame.Stars,aspect);
List<double> ratios=[];List<double>[] regional=Enumerable.Range(0,12).Select(_=>new List<double>()).ToArray();int[] expected=new int[12],found=new int[12];HashSet<Star> used=[]; List<double> ratios=[];List<double>[] regional=Enumerable.Range(0,12).Select(_=>new List<double>()).ToArray();int[] expected=new int[12],found=new int[12];HashSet<Star> used=[];
List<double> residuals=[];List<Star> tight=[];int[] tightRegions=new int[12];int index=-1;
foreach(var s in reference.Stars.Take(600)) foreach(var s in reference.Stars.Take(600))
{ {
index++;
var(x,y)=tr.Apply(s.X,s.Y*aspect);y/=aspect; var(x,y)=tr.Apply(s.X,s.Y*aspect);y/=aspect;
if(x<.025||x>.975||y<.025||y>.975)continue; if(x<.025||x>.975||y<.025||y>.975)continue;
int region=Math.Clamp((int)(s.Y*4),0,3)*3+Math.Clamp((int)(s.X*3),0,2);expected[region]++; int region=Math.Clamp((int)(s.Y*4),0,3)*3+Math.Clamp((int)(s.X*3),0,2);expected[region]++;
var other=grid.Nearest(x,y*aspect);if(other==null||!used.Add(other))continue;found[region]++; var other=grid.Nearest(x,y*aspect);if(other==null||!used.Add(other))continue;found[region]++;
double residual=Math.Sqrt(Math.Pow(other.X-x,2)+Math.Pow((other.Y-y)*aspect,2))*reference.Width;
residuals.Add(residual);
if(residual<=Math.Max(1,reference.Fwhm*.4))
{tight.Add(s);tightRegions[region]++;frame.MatchedShape[index]=other.Ellipticity;}
double ratio=other.Flux/Math.Max(s.Flux,.001);if(double.IsFinite(ratio)&&ratio>0){ratios.Add(ratio);regional[region].Add(ratio);} double ratio=other.Flux/Math.Max(s.Flux,.001);if(double.IsFinite(ratio)&&ratio>0){ratios.Add(ratio);regional[region].Add(ratio);}
} }
frame.MatchFraction=(double)found.Sum()/Math.Max(expected.Sum(),1); frame.MatchFraction=(double)found.Sum()/Math.Max(expected.Sum(),1);
double median=Robust.Median(residuals),p95=Robust.Quantile(residuals,.95),scale=Math.Sqrt(tr.A*tr.A+tr.B*tr.B);
int occupied=tightRegions.Count(n=>n>=3);
// Star survival is attenuation evidence, not a geometric fit-quality measure.
// Require tight, distributed correspondences before trusting any rejection.
bool verified=tight.Count>=20&&tight.Count>=residuals.Count*.8&&occupied>=3&&
tight.Max(s=>s.X)-tight.Min(s=>s.X)>=.15&&tight.Max(s=>s.Y)-tight.Min(s=>s.Y)>=.15&&
median<=Math.Max(.5,reference.Fwhm*.2)&&p95<=Math.Max(1,reference.Fwhm*.5)&&scale>=.975&&scale<=1.025;
frame.AdditionalMeasurements["Alignment verified"]=verified?1:0;
frame.AdditionalMeasurements["Alignment matches"]=residuals.Count;
frame.AdditionalMeasurements["Alignment tight matches"]=tight.Count;
frame.AdditionalMeasurements["Alignment median residual pixels"]=median;
frame.AdditionalMeasurements["Alignment p95 residual pixels"]=p95;
frame.AdditionalMeasurements["Alignment occupied regions"]=occupied;
if(ratios.Count>=15)frame.Transparency=Robust.Median(ratios); if(ratios.Count>=15)frame.Transparency=Robust.Median(ratios);
for(int i=0;i<12;i++){frame.RegionMatches[i]=regional[i].Count;frame.RegionLoss[i]=expected[i]>=8?1-(double)found[i]/expected[i]:0;frame.RegionFlux[i]=regional[i].Count>=5?Robust.Median(regional[i]):0;} for(int i=0;i<12;i++){frame.RegionMatches[i]=regional[i].Count;frame.RegionLoss[i]=expected[i]>=8?1-(double)found[i]/expected[i]:0;frame.RegionFlux[i]=regional[i].Count>=5?Robust.Median(regional[i]):0;}
frame.SpatialLoss=frame.RegionLoss.Max(); frame.SpatialLoss=frame.RegionLoss.Max();

View file

@ -40,6 +40,7 @@ public sealed class Frame
[System.Text.Json.Serialization.JsonIgnore] public float[] SkyMap {get;set;}=[]; [System.Text.Json.Serialization.JsonIgnore] public float[] SkyMap {get;set;}=[];
[System.Text.Json.Serialization.JsonIgnore] public float[] SkyNoise {get;set;}=[]; [System.Text.Json.Serialization.JsonIgnore] public float[] SkyNoise {get;set;}=[];
[System.Text.Json.Serialization.JsonIgnore] public double[] ForcedFlux {get;set;}=[]; [System.Text.Json.Serialization.JsonIgnore] public double[] ForcedFlux {get;set;}=[];
[System.Text.Json.Serialization.JsonIgnore] public double[] MatchedShape {get;set;}=[];
[System.Text.Json.Serialization.JsonIgnore] public string ForcedReferenceId {get;set;}=""; [System.Text.Json.Serialization.JsonIgnore] public string ForcedReferenceId {get;set;}="";
[System.Text.Json.Serialization.JsonIgnore] public bool CacheDirty {get;set;} [System.Text.Json.Serialization.JsonIgnore] public bool CacheDirty {get;set;}
public double Score {get;set;} public double Score {get;set;}
@ -69,7 +70,7 @@ public sealed class Frame
} }
public sealed class Project public sealed class Project
{ {
public const string EngineVersion="0.2.0-spatial-reference-1"; public const string EngineVersion="0.2.1-spatial-reference-2";
public string Version {get;set;}=EngineVersion; public string Version {get;set;}=EngineVersion;
public string Name {get;set;}="Untitled"; public string Name {get;set;}="Untitled";
public string Source {get;set;}=""; public string Source {get;set;}="";

View file

@ -55,6 +55,18 @@ public static class SpatialReference
if(ratios.Count>=20)f.Transparency=Math.Max(0,Robust.Median(ratios)); if(ratios.Count>=20)f.Transparency=Math.Max(0,Robust.Median(ratios));
for(int r=0;r<12;r++){f.RegionMatches[r]=regional[r].Count;if(regional[r].Count>=5)f.RegionFlux[r]=Robust.Median(regional[r]);} for(int r=0;r<12;r++){f.RegionMatches[r]=regional[r].Count;if(regional[r].Count>=5)f.RegionFlux[r]=Robust.Median(regional[r]);}
f.AdditionalMeasurements["Forced flux samples"]=ratios.Count; f.AdditionalMeasurements["Forced flux samples"]=ratios.Count;
// Compare the same bright reference stars, avoiding changes in the
// independently selected detection cohort. Missing/uncertain pairs
// cannot dismiss an existing morphology warning.
var shapePeers=Nearby(a,clean,f).Where(x=>x.MatchedShape.Length==stars).ToArray();
double[] normalShape=Enumerable.Range(0,Math.Min(250,stars)).Select(j=>Robust.Median(shapePeers.Select(x=>x.MatchedShape[j]))).ToArray();
var changes=Enumerable.Range(0,Math.Min(250,stars)).Where(j=>
f.MatchedShape.Length==stars&&double.IsFinite(f.MatchedShape[j])&&normalShape[j]>0&&normal[j]*anchor.Exposure>=30*anchor.Noise)
.Select(j=>f.MatchedShape[j]-normalShape[j]).ToArray();
f.AdditionalMeasurements["Matched shape samples"]=changes.Length;
f.AdditionalMeasurements["Matched shape reference frames"]=shapePeers.Length;
f.AdditionalMeasurements["Matched shape change"]=Robust.Median(changes);
f.AdditionalMeasurements["Matched shape uncertainty"]=Math.Max(.005,1.253*Robust.Mad(changes)/Math.Sqrt(Math.Max(1,changes.Length)));
} }
var residuals=clean.Select(i=>Difference(maps[i],reference)).ToArray(); var residuals=clean.Select(i=>Difference(maps[i],reference)).ToArray();
double referenceNoise=Robust.Median(clean.Select(i=>a[i].Noise)); double referenceNoise=Robust.Median(clean.Select(i=>a[i].Noise));
@ -70,6 +82,10 @@ public static class SpatialReference
// The two-sided residual remains classification evidence. Pink display // The two-sided residual remains classification evidence. Pink display
// cells identify excess sky brightness, not the compensating dark residual. // cells identify excess sky brightness, not the compensating dark residual.
f.ExcessSkyMask=overlay?.Measure(f)??[]; f.ExcessSkyMask=overlay?.Measure(f)??[];
bool[] positive=f.ExcessSkyMask.Select(v=>v>=128).ToArray();
f.AdditionalMeasurements["Cloud positive coverage"]=positive.Length==SkyModel.Count?(double)positive.Count(v=>v)/Math.Max(valid,1):0;
f.AdditionalMeasurements["Cloud positive connected coverage"]=positive.Length==SkyModel.Count?(double)SkyModel.LargestComponent(positive)/Math.Max(valid,1):0;
MeasureCloudApertures(f,anchor,a,clean,normal,positive);
f.CloudMask=mask.Select(v=>(byte)(v?1:0)).ToArray();f.CloudCoverage=(double)mask.Count(v=>v)/Math.Max(valid,1);f.CloudSignificance=Robust.Quantile(z,.95); f.CloudMask=mask.Select(v=>(byte)(v?1:0)).ToArray();f.CloudCoverage=(double)mask.Count(v=>v)/Math.Max(valid,1);f.CloudSignificance=Robust.Quantile(z,.95);
f.ReferenceConfidence=stability*Math.Min(1,clean.Length/20.0)*(double)valid/SkyModel.Count; f.ReferenceConfidence=stability*Math.Min(1,clean.Length/20.0)*(double)valid/SkyModel.Count;
f.AdditionalMeasurements["Cloud connected coverage"]=(double)SkyModel.LargestComponent(mask)/Math.Max(valid,1); f.AdditionalMeasurements["Cloud connected coverage"]=(double)SkyModel.LargestComponent(mask)/Math.Max(valid,1);
@ -81,6 +97,38 @@ public static class SpatialReference
} }
} }
} }
static void MeasureCloudApertures(Frame frame,Frame anchor,Frame[] frames,int[] clean,double[] normal,bool[] positive)
{
var indices=Enumerable.Range(0,normal.Length).Where(j=>positive.Length==SkyModel.Count&&
positive[Math.Clamp((int)(anchor.Stars[j].Y*SkyModel.Rows),0,SkyModel.Rows-1)*SkyModel.Columns+Math.Clamp((int)(anchor.Stars[j].X*SkyModel.Columns),0,SkyModel.Columns-1)]&&
normal[j]*anchor.Exposure>=30*anchor.Noise&&frame.ForcedFlux.Length==normal.Length&&double.IsFinite(frame.ForcedFlux[j])).ToArray();
frame.AdditionalMeasurements["Cloud aperture samples"]=indices.Length;
frame.AdditionalMeasurements["Cloud aperture regions"]=indices.Select(j=>Math.Clamp((int)(anchor.Stars[j].Y*4),0,3)*3+Math.Clamp((int)(anchor.Stars[j].X*3),0,2)).Distinct().Count();
frame.AdditionalMeasurements["Cloud aperture flux ratio"]=0;frame.AdditionalMeasurements["Cloud aperture flux deviation"]=0;
if(indices.Length<20)return;
double Flux(Frame f)=>Robust.Median(indices.Where(j=>f.ForcedFlux.Length==normal.Length&&double.IsFinite(f.ForcedFlux[j])).Select(j=>f.ForcedFlux[j]/normal[j]));
var comparisons=Nearby(frames,clean,frame).Where(f=>f.ForcedFlux.Length==normal.Length&&indices.Count(j=>double.IsFinite(f.ForcedFlux[j]))>=indices.Length*.8).Select(Flux).ToArray();
double baseline=Robust.Median(comparisons);if(comparisons.Length<12||baseline<=0)return;
// Radius-4 aperture (49 pixels) and radius-6..8 median annulus.
// Include background-estimation uncertainty, rather than interpreting a
// noisy local flux estimate as confident extinction.
const double varianceFactor=49+49*49*Math.PI/(2*88);
double LocalNoise(int j)
{
var s=anchor.Stars[j];var a=frame.Alignment;double aspect=(double)anchor.Height/anchor.Width;
if(frame.SkyNoise.Length!=SkyModel.Count||a.Length!=4)return frame.Noise;
double x=a[0]*s.X-a[1]*s.Y*aspect+a[2],y=(a[1]*s.X+a[0]*s.Y*aspect+a[3])/aspect;
return Math.Max(frame.Noise,SkyModel.Sample(frame.SkyNoise,x,y));
}
double uncertainty=1.253*Math.Sqrt(indices.Sum(j=>Math.Pow(LocalNoise(j)*Math.Sqrt(varianceFactor)/Math.Max(1,frame.Exposure)/normal[j],2)))/indices.Length;
double spread=Math.Max(baseline*.03,Math.Max(Robust.Mad(comparisons),uncertainty));
double measured=Flux(frame);
frame.AdditionalMeasurements["Cloud aperture flux ratio"]=measured/baseline;
frame.AdditionalMeasurements["Cloud aperture flux deviation"]=(baseline-measured)/spread;
frame.AdditionalMeasurements["Cloud aperture flux uncertainty"]=uncertainty/baseline;
}
static IEnumerable<Frame> Nearby(Frame[] frames,int[] clean,Frame selected)=>clean.Select(i=>frames[i]).Where(f=>f.Id!=selected.Id)
.OrderBy(f=>Math.Abs((f.Timestamp.GetValueOrDefault()-selected.Timestamp.GetValueOrDefault()).TotalSeconds)).Take(40);
static double Amplitude(double[] map)=>Robust.Quantile(map.Select(Math.Abs),.95); static double Amplitude(double[] map)=>Robust.Quantile(map.Select(Math.Abs),.95);
static double[] Difference(double[] map,double[] reference)=>SkyModel.Detrend(map.Select((v,j)=>v-reference[j]).ToArray()); static double[] Difference(double[] map,double[] reference)=>SkyModel.Detrend(map.Select((v,j)=>v-reference[j]).ToArray());
static double[] MedianMap(double[][] maps)=>Enumerable.Range(0,SkyModel.Count).Select(j=>maps.Any(m=>double.IsFinite(m[j]))?Robust.Median(maps.Select(m=>m[j])):double.NaN).ToArray(); static double[] MedianMap(double[][] maps)=>Enumerable.Range(0,SkyModel.Count).Select(j=>maps.Any(m=>double.IsFinite(m[j]))?Robust.Median(maps.Select(m=>m[j])):double.NaN).ToArray();

File diff suppressed because one or more lines are too long

View file

@ -5,6 +5,11 @@ using System.Text;
using System.Text.Json; using System.Text.Json;
ResourcePolicy.InstallWindowsLimit(); ResourcePolicy.InstallWindowsLimit();
if(args.Length>0&&args[0]=="quality-regression")
{
int failed=0;QualityRegression.Run((ok,name)=>{Console.WriteLine($"{(ok?"PASS":"FAIL")} {name}");if(!ok)failed++;});
if(failed>0)throw new Exception($"{failed} quality regressions failed");return;
}
if(args.Length>0&&args[0]=="preview") if(args.Length>0&&args[0]=="preview")
{ {
var plane=PreviewRenderer.Reduce(Fits.Read(Path.GetFullPath(args[1])),3);var pixels=PreviewRenderer.Render(plane);using var writer=new BinaryWriter(File.Create(args[2]));writer.Write(plane.Width);writer.Write(plane.Height);writer.Write(pixels);return; var plane=PreviewRenderer.Reduce(Fits.Read(Path.GetFullPath(args[1])),3);var pixels=PreviewRenderer.Render(plane);using var writer=new BinaryWriter(File.Create(args[2]));writer.Write(plane.Width);writer.Write(plane.Height);writer.Write(pixels);return;
@ -121,6 +126,7 @@ try
} }
finally{Directory.Delete(temp,true);} finally{Directory.Delete(temp,true);}
await ImageRegression.Run(Check); await ImageRegression.Run(Check);
QualityRegression.Run(Check);
OverlayRegression.Run(Check); OverlayRegression.Run(Check);
PreviewRegression.Run(Check); PreviewRegression.Run(Check);
Console.WriteLine($"{passed} checks passed. Synthetic tests validate mechanics, not astrophotography stack quality."); Console.WriteLine($"{passed} checks passed. Synthetic tests validate mechanics, not astrophotography stack quality.");

View file

@ -0,0 +1,63 @@
using AstroFrame.Engine;
using System.Text.Json;
using System.Text.Json.Serialization;
static class QualityRegression
{
public static void Run(Action<bool,string> check)
{
var stars=Enumerable.Range(0,100).Select(i=>new Star(.08+(i%10)*.09,.08+(i/10)*.09,3000-i*10,4,.1)).ToList();
Frame Make(int n)=>new(){Id=n.ToString(),Path="quality"+n,Target="Quality",Exposure=30,Width=2160,Height=3840,
Timestamp=new DateTime(2026,9,20,8,0,0,DateTimeKind.Utc).AddSeconds(n*30),Measured=true,
StarCount=100,Stars=stars.ToList(),Fwhm=4,Ellipticity=.1,Background=1000,Noise=20,
CloudMask=new byte[SkyModel.Count],ReferenceConfidence=1,
AdditionalMeasurements=new(){{"Clear reference member",1},{"Forced flux samples",100},{"Matched shape reference frames",20}}};
var frames=Enumerable.Range(0,40).Select(Make).ToList();
var opaque=Make(40);opaque.Stars=stars.Take(25).ToList();opaque.StarCount=25;
var broad=Make(41);var gradient=Make(42);var uncertain=Make(43);var cohort=Make(44);var trailing=Make(45);
var subtleShape=Make(46);var insufficient=Make(47);var noisyFlux=Make(48);
frames.AddRange([opaque,broad,gradient,uncertain,cohort,trailing,subtleShape,insufficient,noisyFlux]);
foreach(var f in frames)
{
Matching.Measure(frames[0],f);f.Transparency=1;f.RegionFlux=Enumerable.Repeat(1.0,12).ToArray();f.RegionMatches=Enumerable.Repeat(8,12).ToArray();
}
void Cloud(Frame f,double significance,double coverage,double ratio)
{
f.AdditionalMeasurements["Clear reference member"]=0;f.CloudSignificance=significance;f.CloudCoverage=coverage;
f.AdditionalMeasurements["Cloud connected coverage"]=coverage*.6;
f.AdditionalMeasurements["Cloud positive coverage"]=coverage*.6;
f.AdditionalMeasurements["Cloud positive connected coverage"]=coverage*.4;
f.AdditionalMeasurements["Cloud aperture samples"]=80;f.AdditionalMeasurements["Cloud aperture regions"]=6;
f.AdditionalMeasurements["Cloud aperture flux ratio"]=ratio;f.AdditionalMeasurements["Cloud aperture flux deviation"]=(1-ratio)/.03;
}
Cloud(opaque,100,.9,.05);opaque.Transparency=.05;
Cloud(broad,15,.5,.85);broad.Transparency=.90;
Cloud(gradient,15,.5,1);
Cloud(insufficient,15,.5,.85);insufficient.AdditionalMeasurements["Cloud aperture samples"]=19;
Cloud(noisyFlux,15,.5,.85);noisyFlux.AdditionalMeasurements["Cloud aperture flux deviation"]=2.9;
Cloud(uncertain,100,.9,.05);uncertain.Transparency=.05;uncertain.MatchFraction=.1;
uncertain.AdditionalMeasurements["Alignment verified"]=0;
cohort.Ellipticity=.3;cohort.AdditionalMeasurements["Matched shape samples"]=100;
cohort.AdditionalMeasurements["Matched shape change"]=.04;cohort.AdditionalMeasurements["Matched shape uncertainty"]=.01;
trailing.Ellipticity=.3;trailing.AdditionalMeasurements["Matched shape samples"]=100;
trailing.AdditionalMeasurements["Matched shape change"]=.2;trailing.AdditionalMeasurements["Matched shape uncertainty"]=.01;
subtleShape.Ellipticity=.3;subtleShape.AdditionalMeasurements["Matched shape samples"]=100;
subtleShape.AdditionalMeasurements["Matched shape change"]=.12;subtleShape.AdditionalMeasurements["Matched shape uncertainty"]=.005;
Classification.Run(frames);
check(opaque.Automatic==Verdict.Reject,"severe cloud with accurate geometry rejects despite missing reference stars");
check(broad.Automatic==Verdict.Reject,"broad coherent sky structure with local aperture attenuation rejects");
check(gradient.Automatic==Verdict.Review,"uncorroborated sky structure stays REVIEW rather than rejecting a possible gradient");
check(uncertain.Automatic==Verdict.Review,"severe apparent cloud with unverified geometry stays REVIEW");
check(cohort.Automatic==Verdict.Keep,"matched-star evidence can resolve an elongation flag caused by a changed detection cohort");
check(trailing.Automatic==Verdict.Review,"matched-star elongation remains REVIEW when other measurements are normal");
check(subtleShape.Automatic==Verdict.Review,"measurable sub-threshold matched elongation is not promoted to KEEP");
check(insufficient.Automatic==Verdict.Review&&noisyFlux.Automatic==Verdict.Review,"too few or noisy affected apertures cannot authorize cloud rejection");
check(frames.Take(40).All(f=>f.Automatic==Verdict.Keep),"quality policy changes preserve unchanged clear comparisons");
var options=new JsonSerializerOptions{PropertyNameCaseInsensitive=true,Converters={new JsonStringEnumConverter()}};
var fixture=JsonSerializer.Deserialize<Fixture>(File.ReadAllText(Path.Combine(AppContext.BaseDirectory,"Fixtures","quality-real-frames.json")),options)!;
var real=fixture.Frames.ToList();Classification.Run(real);
foreach(var label in fixture.Expected)
check(real.Single(f=>f.Name==label.Key).Automatic==label.Value,$"real measured regression {label.Key}: {label.Value}");
}
sealed record Fixture(string Source,Frame[] Frames,Dictionary<string,Verdict> Expected);
}

View file

@ -1,10 +1,14 @@
{ {
"source": "User-reviewed screenshots supplied 2026-10-02; these are regression examples, not a representative accuracy sample", "source": "User-reviewed screenshots supplied 2026-10-02 and 2026-10-04; these are regression examples, not a representative accuracy sample",
"frames": [ "frames": [
{"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235146.fit","expected":"Reject","reason":"Extensive upper-frame cloud cover"}, {"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235146.fit","expected":"Reject","reason":"Extensive upper-frame cloud cover"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235218.fit","expected":"Reject","reason":"Extensive upper-frame cloud cover incorrectly kept"}, {"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235218.fit","expected":"Reject","reason":"Extensive upper-frame cloud cover incorrectly kept"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235652.fit","expected":"Keep","reason":"User-confirmed clear exposure"}, {"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235652.fit","expected":"Keep","reason":"User-confirmed clear exposure"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235856.fit","expected":"Keep","reason":"User-confirmed clear exposure falsely reviewed"}, {"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235856.fit","expected":"Keep","reason":"User-confirmed clear exposure falsely reviewed"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260920-001734.fit","expected":"Reject","reason":"Extensive bands of cloud incorrectly kept"} {"file":"Light_NGC 1432_30.0s_IRCUT_20260920-001734.fit","expected":"Reject","reason":"Extensive bands of cloud incorrectly kept"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260919-235506.fit","expected":"Reject","reason":"User-supplied 2026-10-04 example with extensive cloud bands previously REVIEW"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260920-000331.fit","expected":"Reject","reason":"User-supplied 2026-10-04 example with extensive cloud bands previously REVIEW"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260920-003713.fit","expected":"Reject","reason":"User-supplied nearly opaque frame previously REVIEW due to lost detections"},
{"file":"Light_NGC 1432_30.0s_IRCUT_20260920-004304.fit","expected":"Reject","reason":"User-supplied 2026-10-04 example with extensive cloud bands previously REVIEW"}
] ]
} }