{ "checker": { "claims": [ { "claim": "schema document prints the JSON Schema your transform must produce", "command": " schema document", "evidence": "\u2026\"model\": \"DocumentDefinition\",\n \"schema\": {\n \"$defs\": {\n \"CommentDefinition\": {\n \"description\": \"Describes a comment on a document\",\n \"properties\": {\n \"id\": {\n \"title\": \"Id\",\n \"type\": \"string\"\n },\n \"author\": {\n \"anyOf\": [\n {\n \"$ref\": \"#/$defs/UserReferenceDefinition\"\n },\n {\n \"type\": \"null\"\n }\n ],\n \"default\": null\n },\n \"content\": {\n \"anyOf", "expected": "DocumentDefinition JSON schema", "input": "schema document", "kind": "claim", "outcome": "Real input produced the expected result: \"model\": \"DocumentDefinition\"", "returncode": 0, "source": "`glean-idx schema document` prints the JSON Schema your `transform()` has to", "source_checked": true, "source_file": "README.md", "verdict": "yes" }, { "claim": "mock test runs the connector without touching the network", "command": " test --phase mock --connector complete:create_connector", "evidence": "\u2026\"documents\": 1\n }\n }\n ]\n }\n}\n", "expected": "1 document posted to mocked Glean", "input": "README snippet connector with static data, mock phase", "kind": "claim", "outcome": "Real input produced the expected result: \"documents\": 1", "returncode": 0, "source": "Test it without touching the network:", "source_checked": true, "source_file": "README.md", "verdict": "yes" }, { "claim": "every command takes --output json for a stable envelope", "command": " schema document --output json", "evidence": "\u2026\"ok\": true,\n \"data\": {\n \"name\": \"document\",\n \"model\": \"DocumentDefinition\",\n \"schema\": {\n \"$defs\": {\n \"CommentDefinition\": {\n \"description\": \"Describes a comment on a document\",\n \"properties\": {\n \"id\": {\n \"title\": \"Id\",\n \"type\": \"string\"\n },\n \"author\": {\n \"anyOf\": [\n {\n \"$ref\": \"#/$defs/UserReferenceDefinition\"\n },\n {\n \"type\": \"null\"\n }\n ],\n \"default\": null\n ", "expected": "JSON envelope with ok true", "input": "schema document --output json", "kind": "claim", "outcome": "Real input produced the expected result: \"ok\": true", "returncode": 0, "source": "Every command takes `--output json` for a stable envelope, `--yes` to skip", "source_checked": true, "source_file": "README.md", "verdict": "yes" } ], "cycle_checks": [ { "claim": "the connector indexing job runs on a recurring schedule", "command": " test --phase mock --connector complete:create_connector", "evidence": "{\n \"ok\": true,\n \"data\": {\n \"connector\": \"create_connector\",\n \"mode\": \"full\",\n \"clients\": [\n \"data_client\"\n ],\n \"phases\": [\n {\n \"phase\": \"mock\",\n \"status\": \"ran\",\n \"fidelity\": \"mocked Glean, the connector's own data clients\",\n \"posted\": {\n \"documents\": 1\n }\n }\n ]\n }\n}", "expected": "one full-mode cycle posts 1 document to mocked Glean", "input": "snippet connector project, WIKI_API_TOKEN placeholder", "kind": "cycle", "method": "run-once substitute: glean-idx test --phase mock runs the same connector full-mode fetch/transform/upload path once against mocked Glean", "outcome": "Triggered one cycle by hand (run-once substitute: glean-idx test --phase mock runs the same connector full-mode fetch/transform/upload path once against mocked Glean; normal trigger: Kubernetes CronJob daily at 02:00 UTC (not triggered here; needs GKE/GCP secrets)): {\n \"ok\": true,\n \"data\": {\n \"connector\": \"create_connector\",\n \"mode\": \"full\",\n \"clients\": [\n \"data_client\"\n ],\n \"phases\": [\n {\n \"phase\": \"mock\",\n \"status\": \"ran\",\n \"fidelity\": \"mocked Glean, the connector's own data clients\",\n \"posted\": {\n \"documents\": 1\n }\n }\n ]\n }\n}", "returncode": 0, "schedule": "Kubernetes CronJob daily at 02:00 UTC (not triggered here; needs GKE/GCP secrets)", "source": "src/glean/indexing/deployment/config.py", "verdict": "yes" } ], "details": { "argv": [ "", "test", "--phase", "mock", "--connector", "complete:create_connector" ], "kind": "cli", "needs_browser": false, "needs_credentials": false, "returncode": 0, "scope": "product operation", "stderr": "", "stdout": "{\n \"ok\": true,\n \"data\": {\n \"connector\": \"create_connector\",\n \"mode\": \"full\",\n \"clients\": [\n \"data_client\"\n ],\n \"phases\": [\n {\n \"phase\": \"mock\",\n \"status\": \"ran\",\n[excerpt truncated]", "working": true }, "evidence_strength": "readme_checked", "example_checks": [ { "claim": "deploy init generates Docker and Terraform files for a connector project", "command": " deploy init --cloud gcp --connector-factory create_connector --output-dir -y", "deferred": false, "evidence": "les.tf\",\n \"glean_deployment.yaml\",\n \".env.example\",\n \".dockerignore\"\n ],\n \"next_steps\": [\n \"Edit glean_deployment.yaml \\u2014 set cluster_name, region, and registry.\",\n \"GCP GKE docs: https://cloud.google.com/kubernetes-engine/docs\",\n \"GCP Artifact Registry: https://cloud.google.com/artifact-registry/docs\",\n \"cp .env.example .env # fill in connector credentials\",\n \"glean-idx datasource configure\",\n \"glean-idx deploy build --push\",\n \"glean-idx deploy secrets upload\",\n \"glean-idx deploy apply\",\n \"glean-idx deploy run\"\n ]\n }\n}", "expected": "generated_files list including glean_deployment.yaml", "input": "deploy init --cloud gcp --connector-factory create_connector into a fresh directory", "kind": "example", "outcome": "command completed successfully", "returncode": 0, "source": "README.md", "verdict": "yes" }, { "claim": "repository unit tests (documented mise run test task, expanded to uv run pytest)", "command": "uv run pytest -q -p no:logging tests/", "deferred": false, "evidence": "gs.py:7: could not import 'watchtower': No module named 'watchtower'\nSKIPPED [1] tests/unit_tests/observability/plugins/aws/test_cloudwatch_metrics.py:5: could not import 'boto3': No module named 'boto3'\nSKIPPED [1] tests/unit_tests/observability/plugins/gcp/test_cloud_logging.py:7: could not import 'google.cloud.logging_v2': No module named 'google'\nSKIPPED [1] tests/unit_tests/observability/plugins/gcp/test_cloud_monitoring.py:5: could not import 'google.cloud.monitoring_v3': No module named 'google'\n1017 passed, 4 skipped in 36.87s\nCannot read termcap database;\nusing dumb terminal settings.", "expected": "all unit tests pass; 4 optional cloud-plugin tests skipped", "input": "tests/", "kind": "example", "outcome": "command completed successfully", "returncode": 0, "source": "mise.toml [tasks.test]", "verdict": "yes" } ], "status": "PASS" }, "deviations": [ "No mise installed; used the equivalent uv commands from mise.toml tasks setup and test directly.", "glean-idx test requires a project folder containing glean_deployment.yaml; the README snippet was run from a project made with deploy init in , not in the snippet folder.", "The README snippet's connector was used unmodified; WIKI_API_TOKEN was set to a placeholder because the snippet reads it at import time, and the mock phase never calls the source.", "Cycle replay: the real trigger is a Kubernetes CronJob (default cron 0 2 * * *) running generated run.py, which needs GCP Secret Manager and a GKE cluster. It was not triggered. Substitute: one run of glean-idx test --phase mock, which runs the same full-mode fetch/transform/upload path against mocked Glean. This does not test the schedule, secrets, or real upload.", "glean-idx doctor was run without credentials: it exits 3 with missing_credentials (GLEAN_SERVER_URL and GLEAN_INDEXING_API_TOKEN unset), matching the README's exit-code contract. No real credentials were available, so the live phases (integration/live) were not run.", "Claim README 'Every command takes --output json' was tested on the schema command only.", "The validator's example for deploy init uses a fresh output directory () that the checker creates." ], "fault_owner": null, "main_blocker": null, "operation": "Build and run the glean-indexing-sdk CLI from source, generate a connector project from the README quickstart snippet, and run its offline mock test (glean-idx test --phase mock), which fetches, transforms and posts one document to mocked Glean; also glean-idx schema document.", "operation_scope": "Offline only. No live Glean instance, no Glean credentials, no GCP/Kubernetes. Python 3.12.3 with uv 0.13.5 (README requires Python >=3.10). Pinned commit 8b6a809814922ca77bf21f80d9b5bbbc6e1f457f (version 1.2.2). CDP endpoint http://:9226/json/version returned HTTP 200 (Chrome/155), but no browser step was needed.", "outcome": "working", "plain_failure": null, "steps": [ "git init; git remote add origin https://github.com/gleanwork/glean-indexing-sdk; git fetch --depth 1 origin 8b6a809814922ca77bf21f80d9b5bbbc6e1f457f; git reset --mixed FETCH_HEAD", "uv sync --locked --extra dev --extra test --extra lint --extra typing (mise.toml [tasks.setup])", "uv run glean-idx --help; uv run glean-idx schema document", "uv run pytest -q -p no:logging tests/ (mise.toml [tasks.test]): 1017 passed, 4 skipped", "glean-idx deploy init --cloud gcp --connector-factory create_connector (README CLI section) to create a connector project", "copy snippets/non_streaming/complete.py (README quickstart code) into the project as complete.py", "glean-idx test --phase mock --connector complete:create_connector (README: test without touching the network) -> \"documents\": 1" ], "stuck_step": null, "suggested_fix": null }