{ "run_id": "gridpin--gridpin-claude-haiku-5-5-1792779867077218910", "checker": { "claims": [ { "claim": "The CLI lists the country sheets GridPin offers, including Serbia (rs).", "command": " --list", "evidence": "Serbia 32 MB downloaded\nsheets directory: \n", "expected": "a list of countries that includes Serbia", "input": "gridpin-fetch --list", "kind": "claim", "outcome": "Real input produced the expected result: Serbia", "returncode": 0, "source": "Countries: `br` `de` `es` `fr` `it` `nl` `rs` (`gridpin-fetch --list`). For AI agents there is an", "source_checked": true, "source_file": "README.md", "verdict": "yes" }, { "claim": "The gridpin CLI forward-geocodes a French address against the France sheet and returns a house-level match.", "command": " query '1 rue de la Paix Paris' -k 3", "evidence": "Rue de la Paix\",\"housenumber\":\"1\",\"commune\":\"Paris 2e Arrondissement\",\"postcode\":\"75002\",\"flags\":[\"street_exact\",\"commune_prefix\",\"house_rep\"]}\n", "expected": "a house-precision candidate on Rue de la Paix", "input": "1 rue de la Paix Paris, against france.bin", "kind": "claim", "outcome": "Real input produced the expected result: Rue de la Paix", "returncode": 0, "source": "$ gridpin query france.bin \"1 rue de la Paix Paris\" -k 3", "source_checked": true, "source_file": "README.md", "verdict": "yes" }, { "claim": "After a sheet is downloaded, geocoding runs from the local file: the MCP server returns a house-level Serbian address match.", "command": " tools/call {\"tool\": \"geocode\", \"arguments\": {\"query\": \"Knez Mihailova 5, Beograd\", \"country\": \"rs\"}}", "evidence": "Kneza Mihaila\\\",\\n \\\"housenumber\\\": \\\"5\\\",\\n \\\"commune\\\": \\\"Stari Grad\\\",\\n \\\"postcode\\\": \\\"\\\",\\n \\\"confidence\\\": 0.9399999976158142,\\n \\\"flags\\\": [\\n \\\"street_exact\\\",\\n \\\"commune_exact\\\",\\n \\\"house_rep\\\"\\n ]\\n }\\n ]\\n}\", \"type\": \"text\"}], \"isError\": false, \"structuredContent\": {\"country\": \"rs\", \"query\": \"Knez Mihailova 5, Beograd\", \"results\": [{\"lat\": 44.8154522, \"lon\": 20.4593897, \"precision\": \"house\", \"score\": 5.5504302978515625, \"confidence\": 0.9399999976158142, \"street\": \"Kneza Mihaila\", \"commune\": \"Stari Grad\", \"postcode\": \"\", \"hous", "expected": "a house-precision result on Kneza Mihaila", "input": "{\"tool\": \"geocode\", \"arguments\": {\"query\": \"Knez Mihailova 5, Beograd\", \"country\": \"rs\"}}", "kind": "claim", "outcome": "Real input produced the expected result: Kneza Mihaila", "source": "Everything after the download runs offline.", "source_checked": true, "source_file": "README.md", "verdict": "yes" } ], "cycle_checks": [ { "claim": "Scheduled or looping main behavior", "command": "", "evidence": "Entry points inspected: the gridpin CLI subcommands (build, reverse, query, batch, meta, repack), gridpin-fetch, and the gridpin-mcp stdio server are all request-driven and have no cron, scheduler, --once/--now flag, webhook or loop. The while-True loops are in the download read loop (mcp/src/gridpin_mcp/sheets.py) and in the internal prep/ data pipeline, not a product schedule. Nothing is triggered on a timer.", "expected": "identify the product trigger", "input": "source inspection", "kind": "cycle", "outcome": "Entry points inspected: the gridpin CLI subcommands (build, reverse, query, batch, meta, repack), gridpin-fetch, and the gridpin-mcp stdio server are all request-driven and have no cron, scheduler, --once/--now flag, webhook or loop. The while-True loops are in the download read loop (mcp/src/gridpin_mcp/sheets.py) and in the internal prep/ data pipeline, not a product schedule. Nothing is triggered on a timer.", "source": "repository entry points", "verdict": "yes" } ], "details": { "argv": [ "" ], "kind": "mcp", "scope": "product operation", "serverInfo": { "name": "gridpin", "title": "GridPin offline geocoder", "version": "0.1.0", "websiteUrl": "https://gridpin.dev" }, "tool_result": { "content": [ { "text": "{\n \"country\": \"fr\",\n \"query\": \"1 rue de la Paix, Paris\",\n \"results\": [\n {\n \"lat\": 48.868546,\n \"lon\": 2.33031,\n \"precision\": \"house\",\n \"score\": 0.08887004852294922,\n \"street\": \"Rue de la Paix\",\n \"housenumber\": \"1\",\n \"commune\": \"Paris 2e Arrondissement\",\n \"postcode\": \"75002\",\n \"confidence\": 0.8399999737739563,\n \"flags\": [\n \"street_exact\",\n \"commune_prefix\",\n \"house_rep\"\n ]\n }\n ]\n}", "type": "text" } ], "isError": false, "structuredContent": { "country": "fr", "query": "1 rue de la Paix, Paris", "results": [ { "commune": "Paris 2e Arrondissement", "confidence": 0.8399999737739563, "flags": [ "street_exact", "commune_prefix", "house_rep" ], "housenumber": "1", "lat": 48.868546, "lon": 2.33031, "postcode": "75002", "precision": "house", "score": 0.08887004852294922, "street": "Rue de la Paix" } ] } }, "tool_result_checked": true, "tool_schema_bytes": 6758, "tools": [ "list_countries", "download_sheet", "geocode", "reverse", "batch" ], "working": true }, "evidence_strength": "readme_checked", "example_checks": [ { "claim": "Batch-geocode five French addresses with the Python binding against the France sheet", "command": " examples/batch.py ", "deferred": false, "evidence": " \"commune_exact\",\n \"house_rep\",\n \"pc_exact\"\n ]\n }\n ]\n },\n {\n \"query\": \"16 rue de la Monnaie, 59800 Lille\",\n \"results\": [\n {\n \"lat\": 50.640827,\n \"lon\": 3.063621,\n \"precision\": \"house\",\n \"score\": 7.2687788009643555,\n \"street\": \"Rue de la Monnaie\",\n \"housenumber\": \"16\",\n \"commune\": \"Lille\",\n \"postcode\": \"59800\",\n \"confidence\": 0.949999988079071,\n \"flags\": [\n \"street_exact\",\n \"commune_exact\",\n \"house_rep\",\n \"pc_exact\"\n ]\n }\n ]\n }\n]", "expected": "JSON with one result per address, e.g. 10 rue de Rivoli resolves to Rue de Rivoli", "input": "examples/batch.py with the downloaded france.bin", "kind": "example", "outcome": "command completed successfully", "returncode": 0, "source": "examples/README.md (batch.py: python batch.py /path/to/france.bin)", "verdict": "yes" }, { "claim": "The MCP package's offline test suite passes", "command": " -m pytest mcp/tests -q", "deferred": false, "evidence": "sss.............sssssssssssss.. [100%]\n15 passed, 16 skipped in 3.10s", "expected": "15 passed; 16 skipped because the Monaco test sheet (make mc) is not built", "input": "python -m pytest mcp/tests -q", "kind": "example", "outcome": "command completed successfully", "returncode": 0, "source": "mcp/tests (CI runs python -m pytest in this layout)", "verdict": "yes" }, { "claim": "Photon-compatible loopback HTTP facade for a sheet", "command": "", "deferred": true, "evidence": "Left out on purpose: it is a loopback interop adapter for geocoder-tester runs and needs the CLI binary plus a long-running server; core geocoding was exercised through the MCP probe, CLI and batch example instead.", "expected": "documented command completes successfully", "input": "", "kind": "example", "outcome": "Left out on purpose: it is a loopback interop adapter for geocoder-tester runs and needs the CLI binary plus a long-running server; core geocoding was exercised through the MCP probe, CLI and batch example instead.", "source": "examples/README.md (gridpin_http.py)", "verdict": "deferred" }, { "claim": "Streamlit live demo and Colab quickstart", "command": "", "deferred": true, "evidence": "Left out on purpose: a UI demo needing Streamlit and a browser or Colab; no browser was installed or used.", "expected": "documented command completes successfully", "input": "", "kind": "example", "outcome": "Left out on purpose: a UI demo needing Streamlit and a browser or Colab; no browser was installed or used.", "source": "examples/README.md (streamlit/, colab_quickstart.ipynb)", "verdict": "deferred" }, { "claim": "Data-build and benchmark pipeline (prep/, make mc, make smoke)", "command": "", "deferred": false, "evidence": "Not run: building the Monaco and country sheets needs the OSM/Overture source downloads and the DuckDB pipeline, which is outside the install-and-query path; the Monaco sheet is missing so 16 mcp tests skip.", "expected": "documented command completes successfully", "input": "", "kind": "example", "outcome": "Not run: building the Monaco and country sheets needs the OSM/Overture source downloads and the DuckDB pipeline, which is outside the install-and-query path; the Monaco sheet is missing so 16 mcp tests skip.", "source": "Makefile smoke and mc targets", "verdict": "blocked" } ], "status": "PASS" }, "diagnosis": null }