Reject inputs without CLR compositional variation

[RonitBStudent]

Summary

• reject inputs with no compositional variation after the CLR transform
• add a regression test using samples that differ only by total library size
• run the unit-test suite in the existing verification workflow

Rows with identical feature proportions have no meaningful circular projection after CLR. Previously, floating-point noise could produce arbitrary near-zero coordinates and a misleading finite rho value instead of surfacing that condition.

Validation

• python3 -m unittest discover -s tests -p 'test_*.py' -v
• PYTHONPATH=. python3 -m clrcycle tests/data/toy_cycle.csv --output-dir <tmp> --all-features --n-swaps 20

[Copilot]

Pull request overview

This PR adds a guard in fit() to reject inputs whose samples have no meaningful compositional variation after the CLR transform (i.e., samples differ only by total library size), preventing misleading near-zero projections and finite rho values caused by floating-point noise.

Changes:

• Add a numerical-tolerance check after CLR-centering to detect and reject “no compositional variation” inputs.
• Add a regression unit test covering scale-only sample matrices.
• Update the GitHub Actions verification workflow to run the unit test suite.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
clrcycle/__init__.py	Adds tolerance-based detection and raises a clear ValueError when CLR-centered variation is effectively zero.
tests/test_clrcycle.py	Adds a regression test ensuring scale-only samples are rejected.
.github/workflows/verify.yml	Runs unittest discovery as part of CI verification.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.