Skip to content

test: add default causal recovery accuracy check - #29

Open
chaewon-research wants to merge 1 commit into
Center-For-Complex-Systems-Science:mainfrom
chaewon-research:test/issue-15-default-accuracy-check
Open

chaewon-research wants to merge 1 commit into
Center-For-Complex-Systems-Science:mainfrom
chaewon-research:test/issue-15-default-accuracy-check

Conversation

@chaewon-research

@chaewon-research chaewon-research commented Sep 13, 2026

Copy link
Copy Markdown

Fixes #15

Summary

  • Adds a lightweight ground-truth causal-recovery accuracy check to the default test path (test_discover_network_basic_numpy).
  • Asserts recovery of the true X0 -> X1 edge at lag 1 and verifies no false edges are discovered.
  • Keeps the broader integration recovery suite behind the existing integration marker so CI stays fast.

Why

The existing integration recovery tests are excluded by default in pyproject.toml, allowing local test runs and CI to pass without checking whether causal recovery actually works.

Validation

  • Focused test passes in ~0.12s:
    python -m pytest causationentropy/tests/test_discovery.py::TestDiscoverNetwork::test_discover_network_basic_numpy -v
  • Full default test suite passes: 381 passed, 17 deselected in ~25s.
  • black, isort, and flake8 checks all pass cleanly.

@chaewon-research
chaewon-research marked this pull request as ready for review September 13, 2026 22:42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

No accuracy test ever runs — in CI or by default

1 participant