You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Run the tests that need no Elasticsearch host in GitHub Actions #44
The repository has no .github/workflows, so nothing runs the test suite on a pull request. Two files, test_set_identifiers.py and test_normalized_nodes.py, are written against the su10:9200 Elasticsearch host and can't run anywhere at present: they construct NodeNormalizationAPINamespace first, which connects to the default localhost:9200, then set ES_HOST on configuration_namespace.elasticsearch, the client namespace rather than the config, which raises TypeError (confirmed on main with a local Elasticsearch on 9200), and they request /nodenorm/get_setid and /nodenorm/get_normalized_nodes where the routes are registered without the prefix. Their expected responses come from the nodenorm_20250507_4ibdxry7 index on su10, so rewriting them needs access to that cluster. Everything else, including test_cors.py, test_health.py, test_version.py and test_openapi.py (added in #42) run anywhere, and that is enough to catch the kind of drift that has gone unnoticed: before #42, the POST set-ID test posted a single object where the handler expects a list, and had evidently not been run in some time.
Suggested first step: a workflow that installs the package with the tests extra and runs those four files on pushes and pull requests. Running the Elasticsearch-backed tests would need either a fixture index in a service container or a reachable test cluster, and is a second step; #19 may be the place that gets decided.
Blocked on #43 if the workflow uses uv run; uv pip install works today.
Activity