Updated flat sigmoid to match birdnet_analyzer

[Josef-Haupt]

• See implementation here
• PR on main because dev is outdated (?)

[Copilot]

Pull request overview

This PR updates the flat_sigmoid_logaddexp_fast function in src/birdnet/utils/helper.py to match the BirdNET-Analyzer reference implementation. The change adds a bias parameter that allows adjustment of the sigmoid transformation, while maintaining backward compatibility through a default value.

Changes:

• Added bias parameter with default value of 1.0 to flat_sigmoid_logaddexp_fast function
• Implemented bias transformation: transformed_bias = (bias - 1.0) * 10.0
• Modified clipping to operate on x + transformed_bias instead of just x

---

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

[Copilot]

The new bias parameter lacks test coverage. Consider adding unit tests to verify:

1. Backward compatibility when bias=1.0 (default)
2. Correct behavior for various bias values
3. The transformed_bias calculation matches the BirdNET-Analyzer reference implementation

This is especially important since there are comprehensive tests for other functions in the repository (e.g., in src/birdnet_tests/), and this function is used in the prediction pipeline.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/birdnet/utils/helper.py	0.00%	2 Missing ⚠️

Files with missing lines	Coverage Δ
src/birdnet/utils/helper.py	62.43% <0.00%> (-0.35%)	⬇️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.