docs/CHANGELOG.md

# Changelog

All notable changes to SatQuery AI. Dates are ISO-8601. This changelog records **what changed and
what was verified**, and marks unresolved items honestly rather than omitting them.

Status tags: `RESOLVED` · `OPEN` · `CLOSED` · `DEFERRED` · `REJECTED`.

---

## [1.0.0] — 2026-09-25 — Final documentation, model release and archival freeze

The release that makes the project legible, citable and reproducible: repository documentation, a
deep research/engineering documentation set, model documentation, a Hugging Face model release, a
curated public repository, and a verified evidence archive.

### Added

**Documentation**

- **Repository README** — motivation through citation, with the deployed topology, the six supported
  tasks, routing, the eight execution events, measured results, live validation, and known
  limitations.
- **Deep architecture reference** (hub + 10 chapters):
  | Chapter | Topic |
  |---|---|
  | `docs/ARCHITECTURE.md` | hub: thesis, sub-document index, cross-cutting principles |
  | `docs/architecture/01-system-overview.md` | thesis, component inventory, frozen-backbone strategy |
  | `docs/architecture/02-deployment-topology.md` | four tiers, gateway, tunnel, wake flow, cold start |
  | `docs/architecture/03-request-lifecycle.md` | the nine-state controller, validation, tiling |
  | `docs/architecture/04-router.md` | frozen MiniLM, the 5-head adapter, `interpret()` vs `chooseTask()`, lexical fallback |
  | `docs/architecture/05-specialists.md` | all six tasks in depth |
  | `docs/architecture/06-evidence-and-confidence.md` | evidence aggregation, temperature scaling, the eight events |
  | `docs/architecture/07-configuration-freeze.md` | the registry, enforced invariants, the config hash |
  | `docs/architecture/08-api-contract.md` | the four endpoints, envelopes, error codes |
  | `docs/architecture/09-frontend.md` | the static pages, real-vs-preview, platform traps |
  | `docs/architecture/10-observability-and-ops.md` | health, counters, traces, what is not observed |
- **Research documentation** — `MODELS.md`, `BENCHMARKS.md`, `DATASETS.md`, `TRAINING.md`,
  `EVALUATION.md`, `DEPLOYMENT.md`, `REPRODUCIBILITY.md`, `LIMITATIONS.md`, `RESEARCH_NOTES.md`,
  `CHANGELOG.md`.
- **Model documentation** — `MODEL_CARD.md`, plus **generated** `models/manifest.json` and
  `models/checksums.sha256` (computed by reading the files, never typed).
- **Metric verification tooling** — `release/tools/verify_readme_metrics.py`, which walks every quoted
  metric to its source artifact. Result: **`ALL CLAIMS VERIFIED`** (20/20).

**Release**

- **Hugging Face release** of the six trained artifacts to `thundercode/SatQuery`, with a verification
  record (`HF_RELEASE_VERIFICATION.md`). Independent re-download confirmed **6/6 artifacts
  byte-identical**.
- **Curated public repository** at `Anish-lab-blip/SatQuery-AI` with `RELEASE_MANIFEST.md`.
- **Evidence archive** (ZIP64, `artifacts/` verbatim per owner decision G3) with an integrity-verified
  extraction and a documented remote location.

### Changed

- The monorepo `README.md` was **materially stale** and is superseded by this release's README. The
  stale version called the frontend "hermetic — no backend calls", put Render/Codespace as
  "in progress", described a `/v1/*` contract, and pointed at the untracked `deploy/` as the
  deployment source. All four were wrong.
- `hf/` documentation was **stale** — it asserted the project "does not own any model weights … ships
  no weights, no binaries, and no large artifacts" and "has no Hugging Face credentials". Both were
  false at release time and are corrected.

### Fixed

- **Router defect** — *"Where are the built-up areas in this image?"* collapsed to `vqa` and answered
  "River". Fixed in the dispatch path and validated across three independent live passes (8/8 each).
  `RESOLVED`.
- **Live-validation harness false positive** — synthetic CDP key events were silently dropped when the
  window lacked OS focus, so the harness dispatched the page's default query and recorded a false
  pass. Fixed with pre-dispatch form-state assertions. `RESOLVED`.
- **Bundled EO change pair** — the demo pair differed in shape (726² vs 736²) and errored. Replaced
  with a same-shape pair (both 720×720). `RESOLVED` (supersedes `FINAL_DELIVERY_REPORT.md` §6).
- **B-01 (Hugging Face header link)** — previously `BLOCKED`; the link is now live on all 11 pages.
  `CLOSED`.
- **Metric-verifier dotted-path bug** — the resolver split the artifact key `"0.50"` on its own dot,
  producing a false `NOKEY` for the two grounding recall claims. Fixed to prefer the **longest**
  matching key at each step.

### Known issues at this release

- **B-07** — transient tunnel-agent gaps; a request can hang or return 504. Patch prepared, **NOT
  deployed**. `OPEN`.
- **B-02** — `/api/health` `codespace_name` carries a trailing `\n`. Cosmetic. `OPEN`.
- **No `LICENSE` file** exists. `OPEN`.
- **No system-level end-to-end benchmark** exists. `NOT RUN`.
- **Router test split** never scored. `NOT RUN`.
- **Calibration** worsened ECE (0.013755 → 0.014929). Retained as part of the frozen config only.
- **VLM adapter** is `ACCEPTANCE-REJECTED` despite usable metrics; the deployed path uses the
  unadapted model.
- **optical-SAR** and **change-VQA** rulings remain `OPEN`.

### Deployment revisions at this release

| Component | Revision |
|---|---|
| Frontend | `2d7ae53b482d` |
| Backend | `89d80eaddec5` |
| Inference | `5a0936ace491` |
| Public umbrella | `3dcabd32da41` |

---

## [0.9.0] — 2026-09-25 — Router-defect sprint

- Diagnosed and fixed the router misroute (region queries → `vqa`): the temporal slot in
  `interpret()` now derives from `changeStem` / `pairWording` / `newAsChange`; `built` was dropped as a
  change marker; the quantifier `area` was tightened to `\barea\b`.
- Deployed three commits to the frontend.
- Ran **3 independent live validation passes**, 8/8 each — **24 live runs, 24 correct dispatches,
  0 mock nodes, trace fill 94.4444 %**.
- Wrote the delivery report and the handoff.
- Prepared (but did not deploy) the B-07 backend patch.

---

## [0.8.0] — 2026-09-24/25 — Deployment hardening

- Live topology settled on **Cloudflare Pages → Render → outbound tunnel → Codespace** (CPU).
- Closed the five historical backend blockers (ASGI stack in requirements; `$PORT` read; `OPTIONS`
  preflight; fail-fast app construction; adapter integrity compared against an expected digest).
- Documented cold start; recorded the `transport_mode: auto` fallthrough.

---

## [0.7.0] — Grounding and change heads

- **Grounding**: trainable head over frozen RemoteCLIP ViT-B/32. Resolution **224 chosen, 448
  REJECTED** by a pre-registered paired test (mean best IoU −0.0147, t = −22.63).
- **Change**: STANet-style Siamese detector on LEVIR-CD-256 — the only `VERIFIED` headline metric
  (pooled IoU 0.8122 / macro IoU 0.8457 / pooled F1 0.8964).
- **Optical-SAR**: CROMA-base fusion head, armA/armB × seeds 100–104 sweep, pre-registered 115-class
  metric (accuracy 0.931 / macro-F1 0.434161, ruling `OPEN`).

---

## [0.6.0] — VLM adaptation

- SmolVLM-500M LoRA adapter trained (PEFT 0.19.1, r = 16, α = 32, dropout 0.05, fp16).
- Metrics usable (exact_match 0.963, F1 0.96432) but **ACCEPTANCE-REJECTED** for promotion.
  `REJECTED`.
- Findings F5-1/F5-2/F5-3 recorded (loader class, ~17× processor overrun, chat-template requirement).

---

## [0.5.0] — Router

- 50,822-parameter adapter over frozen MiniLM; cached embeddings; CPU-only training (0.28 s for 20
  epochs / 4,096 vectors).
- Group-based splits with hard negatives placed in the test split.
- Deterministic lexical fallback with ordered precedence and a `self_check()` against curated hard
  negatives.

---

## [0.1.0] — Foundation

- Monorepo, frozen configuration registry (`configs/base.yaml`), config-hash freeze
  (`78f1e3700da15aa1`), nine-state controller, four-endpoint inference contract, evidence schema
  (`Region` / `ChangeRegion` / `Evidence`), eight frontend execution events.

---

## Unreleased / open

| Item | State |
|---|---|
| Deploy the B-07 tunnel-fallthrough fix | prepared, **NOT deployed** |
| Score the router test split | **NOT RUN** |
| Build a system-level end-to-end benchmark | **NOT RUN** |
| Select a licence | **OPEN** |
| Resolve the optical-SAR and change-VQA rulings | **OPEN** |
| Plot the calibrated reliability curve | **NOT RUN** |
| Replace the illustrative benchmark adapter numbers with real runs | **NOT RUN** |

---

## Documentation inventory shipped in 1.0.0

| File | Purpose |
|---|---|
| `README.md` | repository front page |
| `MODEL_CARD.md` | model card for the six trained artifacts |
| `docs/ARCHITECTURE.md` + `docs/architecture/01…10` | the deep architecture reference |
| `docs/MODELS.md` | model detail, hyperparameters, acceptance status |
| `docs/BENCHMARKS.md` | every measured metric with its artifact and protocol |
| `docs/DATASETS.md` | dataset provenance, splits, leakage controls |
| `docs/TRAINING.md` | training procedures and hyperparameters |
| `docs/EVALUATION.md` | evaluation protocols and honesty rules |
| `docs/DEPLOYMENT.md` | live topology, env vars, mechanics, traps |
| `docs/REPRODUCIBILITY.md` | what can and cannot be reproduced |
| `docs/LIMITATIONS.md` | the full limitation catalogue |
| `docs/RESEARCH_NOTES.md` | engineering findings and negative results |
| `docs/CHANGELOG.md` | this file |
| `models/manifest.json`, `models/checksums.sha256` | generated artifact identities |