Baseline for ICASSP 2024 Cadenza Challenge

[groadabike]

Closes #339

This PR should result in pyclarity v0.4.0

• Generate Dataset
  • Generate train metadata scenes
  • Generate train audio dataset
• Baseline
  • Enhance
  • Evaluate
• Create source separation support in ustils
• Move ResultFile Class to utils
• Add Flac Encoder
• Test
  • generate train scene
  • generate train audio data
  • baseline enhance
  • baseline evaluate
  • source separation support
  • result file class
  • flac encoder

[groadabike]

@jonbarker68, @ns-rse
The tests in this Draft PR are failing in 2 unrelated tests. Are we having the same issue as before with different architectures?
😨

    def test_smear3():
        """Test smear3"""
        np.random.seed(0)
        input_signal = np.random.rand(10000)
        smear_mat = make_smear_mat3(rl=R_LOWER, ru=R_UPPER, sample_rate=SAMPLE_RATE)
        output_signal = smear3(smear_mat, input_signal)
        assert output_signal.shape == (10240,)
>       assert np.sum(np.abs(output_signal)) == pytest.approx(
            5066.986397433977, rel=pytest.rel_tolerance, abs=pytest.abs_tolerance
        )
E       assert 5047.791163861037 == 5066.986397433977 ± 5.1e-04
E         comparison failed
E         Obtained: 5047.79[116](https://github.com/claritychallenge/clarity/actions/runs/6150288476/job/16687922487?pr=340#step:7:117)3861037
E         Expected: 5066.986397433977 ± 5.1e-04

tests/evaluator/msbg/test_smearing.py:41: AssertionError
____________________________ test_smear_valid_input ____________________________

    def test_smear_valid_input():
        """Test smear with valid input"""
        np.random.seed(0)
        input_signal = np.random.rand(10000)
        output_signal = Smearer(rl=R_LOWER, ru=R_UPPER, sample_rate=SAMPLE_RATE).smear(
            input_signal
        )
        assert output_signal.shape == (10240,)
>       assert np.sum(np.abs(output_signal)) == pytest.approx(
            5066.986397433977, rel=pytest.rel_tolerance, abs=pytest.abs_tolerance
        )
E       assert 5047.791163861037 == 5066.986397433977 ± 5.1e-04
E         comparison failed
E         Obtained: 5047.791163861037
E         Expected: 5066.986397433977 ± 5.1e-04

tests/evaluator/msbg/test_smearing.py:67: AssertionError

[codecov]

Codecov Report

Patch coverage: 88.35% and project coverage change: -0.18% ⚠️

Comparison is base (1663f32) 92.95% compared to head (7dc037e) 92.78%.

❗ Current head 7dc037e differs from pull request most recent head 92b444f. Consider uploading reports for the commit 92b444f to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #340      +/-   ##
==========================================
- Coverage   92.95%   92.78%   -0.18%     
==========================================
  Files          40       43       +3     
  Lines        3778     3924     +146     
==========================================
+ Hits         3512     3641     +129     
- Misses        266      283      +17     

Files Changed	Coverage Δ
clarity/evaluator/haaqi/haaqi.py	90.47% <ø> (ø)
clarity/utils/flac_encoder.py	84.05% <84.05%> (ø)
clarity/utils/source_separation_support.py	85.71% <85.71%> (ø)
clarity/utils/results_support.py	100.00% <100.00%> (ø)
clarity/utils/signal_processing.py	97.43% <100.00%> (+0.88%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jonbarker68]

Happy to merge this once tests are passing.