Jens/geointel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Prepare GeoIntel for public release
Managed validation / Managed repository validation (pull_request) Successful in 1m46s
Details
GeoIntel release gates / Compile, test, contracts and builds (pull_request) Successful in 1m51s
Details
GeoIntel release gates / Python and npm vulnerability policy (pull_request) Successful in 20s
Details
GeoIntel release gates / Production AI image, SBOM and container scan (pull_request) Successful in 15m3s
Details
GeoIntel release gates / Deploy exact gated revision to Unraid (pull_request) Skipped
Details

Browse Source
This commit is contained in:
Jens
2026-08-31 21:33:10 +02:00
parent dcd67b11d1
commit 2cdf9c99c6
195 changed files with 595 additions and 207595 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/build_accuracy_phase1_evidence_manifest.py
+5 -9
View File
@@ -14,14 +14,11 @@ DEFAULT_EVIDENCE_ROOT = REPOSITORY_ROOT / "artifacts" / "evidence" / "accuracy"
DEFAULT_OUTPUT = DEFAULT_EVIDENCE_ROOT / "evidence-manifest.json"
PROGRAM_PATHS = (
"docs/accuracy-program/00-execution-contract.md",
"docs/accuracy-program/01-system-inventory.md",
"docs/accuracy-program/02-data-lineage.md",
"docs/accuracy-program/03-baseline-and-gaps.md",
"docs/accuracy-program/04-risk-register.md",
"docs/accuracy-program/05-metric-framework.md",
"docs/accuracy-program/06-implementation-roadmap.md",
"docs/accuracy-program/status.json",
"docs/ACCURACY.md",
"docs/DATA_SOURCES.md",
"docs/KNOWN_LIMITATIONS.md",
"SECURITY.md",
"fixtures/accuracy/readiness/status.json",
"scripts/build_accuracy_phase1_evidence_manifest.py",
"scripts/collect_accuracy_phase1_inference_smoke.py",
"scripts/collect_accuracy_phase1_ml_lineage.py",
@@ -29,7 +26,6 @@ PROGRAM_PATHS = (
"scripts/reproduce_accuracy_phase1_findings.py",
"scripts/run_accuracy_phase1_baseline.py",
"scripts/verify_accuracy_phase1_evidence.py",
"tests/test_accuracy_phase1_baseline.py",
)
scripts/check_repository_hygiene.py
+7 -2
View File
@@ -10,7 +10,13 @@ from pathlib import Path
ROOT = Path(__file__).resolve().parents[1]
BLOCKED_PATHS = {".mcp.json"}
BLOCKED_PREFIXES = (".codex-input/", ".codex-artifacts/", ".playwright-mcp/", "data/")
BLOCKED_PREFIXES = (
".codex-input/",
".codex-artifacts/",
".playwright-mcp/",
"artifacts/",
"data/",
)
BLOCKED_SUFFIXES = (".db", ".db-shm", ".db-wal")
BLOCKED_CONTENT = (
b"192.168." + b"10.150",
@@ -21,7 +27,6 @@ BLOCKED_CONTENT = (
)
LARGE_FILE_LIMIT = 20 * 1024 * 1024
LARGE_FILE_ALLOWLIST_PREFIXES = (
"artifacts/evidence/accuracy/",
"docs/assets/",
"frontend/public/portfolio/",
"output/pdf/",
scripts/run_accuracy_phase3_full_data_scan.py
+1 -1
View File
@@ -464,7 +464,7 @@ def detect_leakage(items: list[dict[str, Any]], repo_root: Path | None = None) -
prior_quality: dict[str, Any] | None = None
prior_attention = False
if repo_root is not None:
status_path = repo_root / "docs/accuracy-program/status.json"
status_path = repo_root / "fixtures/accuracy/readiness/status.json"
try:
status = json.loads(status_path.read_text(encoding="utf-8"))
v56 = ((status.get("ml_data") or {}).get("v56") or {})
scripts/run_accuracy_phase4_benchmark.py
+15 -15
View File
@@ -2148,9 +2148,9 @@ def product_baseline_manifest_gate(
def readiness_snapshot(repo_root: Path) -> dict[str, Any]:
status_path = repo_root / "docs/accuracy-program/status.json"
p3_path = repo_root / "artifacts/evidence/accuracy/P3/full-scan-manifest.json"
leakage_path = repo_root / "artifacts/evidence/accuracy/P3/leakage-report.json"
status_path = repo_root / "fixtures/accuracy/readiness/status.json"
p3_path = repo_root / "fixtures/accuracy/readiness/full-scan-manifest.json"
leakage_path = repo_root / "fixtures/accuracy/readiness/leakage-report.json"
status = json.loads(status_path.read_text(encoding="utf-8"))
p3 = json.loads(p3_path.read_text(encoding="utf-8"))
leakage = json.loads(leakage_path.read_text(encoding="utf-8"))
@@ -2162,15 +2162,15 @@ def readiness_snapshot(repo_root: Path) -> dict[str, Any]:
"schema_version": 1,
"source_paths": {
"accuracy_status_projection": {
"path": "docs/accuracy-program/status.json",
"path": "fixtures/accuracy/readiness/status.json",
"selected_json_pointers": ["/runtime/active_model"],
"sha256": canonical_hash(status_projection),
},
"phase3_full_scan": repository_file(
repo_root, "artifacts/evidence/accuracy/P3/full-scan-manifest.json"
repo_root, "fixtures/accuracy/readiness/full-scan-manifest.json"
),
"phase3_leakage": repository_file(
repo_root, "artifacts/evidence/accuracy/P3/leakage-report.json"
repo_root, "fixtures/accuracy/readiness/leakage-report.json"
),
},
"active_model": status_projection["runtime"]["active_model"],
@@ -2610,10 +2610,10 @@ def build_input_manifest(
"fixtures/accuracy/p4/split-source-manifest.json",
"fixtures/accuracy/p4/protected-baseline-cases.json",
"fixtures/golden/golden_qa_benchmarks.json",
"docs/accuracy-program/05-metric-framework.md",
"docs/accuracy-program/07-source-authority-matrix.md",
"artifacts/evidence/accuracy/P3/full-scan-manifest.json",
"artifacts/evidence/accuracy/P3/leakage-report.json",
"docs/ACCURACY.md",
"docs/DATA_SOURCES.md",
"fixtures/accuracy/readiness/full-scan-manifest.json",
"fixtures/accuracy/readiness/leakage-report.json",
]
baseline_gate = product_gates.get("executed_product_incumbent_baseline") or {}
manifest_binding: dict[str, Any] | None = None
@@ -2722,7 +2722,7 @@ def run_workflow(
"evaluator_families": sorted(evaluation["evaluated_task_families"]),
"implemented_capabilities": evaluation["task_inventory"],
"metric_contract": repository_file(
repo_root, "docs/accuracy-program/05-metric-framework.md"
repo_root, "docs/ACCURACY.md"
),
"gate_states": sorted(GATE_STATES),
"protected_policy": evaluation["protected_policy"],
@@ -2756,7 +2756,7 @@ def run_workflow(
]
failure_gallery = {
"schema_version": 2,
"taxonomy": "docs/accuracy-program/05-metric-framework.md section 4",
"taxonomy": "docs/ACCURACY.md",
"failure_count": len(evaluation["failures"]),
"items": evaluation["failures"],
"items_canonical_json_sha256": canonical_hash(evaluation["failures"]),
@@ -2770,7 +2770,7 @@ def run_workflow(
)
error_taxonomy = {
"schema_version": 2,
"source": "docs/accuracy-program/05-metric-framework.md section 4",
"source": "docs/ACCURACY.md",
"observed_codes": [
{"error_code": code, "kind": kind} for code, kind in taxonomy_entries
],
@@ -2893,10 +2893,10 @@ def run_workflow(
repo_root, "fixtures/golden/golden_qa_benchmarks.json"
),
"phase3_full_scan": repository_file(
repo_root, "artifacts/evidence/accuracy/P3/full-scan-manifest.json"
repo_root, "fixtures/accuracy/readiness/full-scan-manifest.json"
),
"phase3_leakage": repository_file(
repo_root, "artifacts/evidence/accuracy/P3/leakage-report.json"
repo_root, "fixtures/accuracy/readiness/leakage-report.json"
),
},
"code": input_manifest["code"],
scripts/train_building_proposal_classifier.py
+2 -2
View File
@@ -319,7 +319,7 @@ def main() -> int:
state_path = args.output_dir / "best-state.pt"
if not state_path.is_file():
raise RuntimeError(f"missing existing best state: {state_path}")
model.load_state_dict(torch.load(state_path, map_location=device))
model.load_state_dict(torch.load(state_path, map_location=device, weights_only=True))
model.eval()
torch.jit.script(model).save(str(args.output_dir / "proposal-classifier.torchscript.pt"))
print(
@@ -369,7 +369,7 @@ def main() -> int:
if float(metric["f1"]) > best_f1:
best_f1 = float(metric["f1"])
torch.save(model.state_dict(), args.output_dir / "best-state.pt")
model.load_state_dict(torch.load(args.output_dir / "best-state.pt", map_location=device))
model.load_state_dict(torch.load(args.output_dir / "best-state.pt", map_location=device, weights_only=True))
model.eval()
scripted = torch.jit.script(model)
model_path = args.output_dir / "proposal-classifier.torchscript.pt"