Skip to content

UPLOAD_TO_RELEASE.md

Latest commit

 

History

History
executable file
·
101 lines (77 loc) · 3.09 KB

UPLOAD_TO_RELEASE.md

File metadata and controls

executable file
·
101 lines (77 loc) · 3.09 KB
Raw

Files to Upload to GitHub Release v0.2.0

Current Status

✅ Release created: https://github.com/prossm/chirpkit/releases/tag/v0.2.0 ✅ birdnet-models.zip uploaded (22.4 MB) ⚠️ chirpkit-ensemble.zip NEEDS UPDATE (18 MB)

Action Required

Update chirpkit-ensemble.zip

The existing chirpkit-ensemble.zip on the release needs to be replaced with the updated version that includes:

  • ✅ Label encoder with scikit-learn 1.7.2 (eliminates version warning)

Steps:

  1. Go to https://github.com/prossm/chirpkit/releases/tag/v0.2.0
  2. Click "Edit release"
  3. Delete the old chirpkit-ensemble.zip (if present)
  4. Upload the new chirpkit-ensemble.zip from project root (18 MB)
  5. Click "Update release"

File location:

/Volumes/PortableSSD/dev/chirpkit/chirpkit-ensemble.zip

Verify the file:

ls -lh /Volumes/PortableSSD/dev/chirpkit/chirpkit-ensemble.zip
# Should show: 18M Nov  6 19:58 chirpkit-ensemble.zip

# Check contents:
unzip -l /Volumes/PortableSSD/dev/chirpkit/chirpkit-ensemble.zip
# Should include:
#   - chirpkit-ensemble/ensemble_info.json
#   - chirpkit-ensemble/ensemble_model_1.pth through ensemble_model_7.pth
#   - chirpkit-ensemble/label_encoder.joblib (UPDATED)

What Changed

Code Fixes (Already in Repository)

  1. ✅ Fixed import paths in classifier.py and simple_ui.py
  2. ✅ Fixed BirdNET path resolution in birdnet_embeddings.py
  3. ✅ Fixed GPU/MPS detection in dependencies.py
  4. ✅ Added is_available() method to InsectClassifier
  5. ✅ Updated requirements.txt with secure package versions
  6. ✅ Improved model_downloader.py with automatic fallbacks

Model Files (Need Release Upload)

  1. ✅ birdnet-models.zip - Already uploaded, no changes needed
  2. ⚠️ chirpkit-ensemble.zip - Needs update with new label_encoder.joblib

Testing After Upload

After uploading the updated chirpkit-ensemble.zip, test the download:

# Test download URL
curl -I https://github.com/prossm/chirpkit/releases/download/v0.2.0/chirpkit-ensemble.zip

# Should return: HTTP/2 200 or 302 (redirect)

# Test full download and installation
python3 -m venv test_env
source test_env/bin/activate
pip install git+https://github.com/prossm/chirpkit.git@main

# Test model download
python -c "
from chirpkit.model_downloader import ModelDownloader
ModelDownloader.download_all_models()
"

# Verify no scikit-learn warning
python -c "
from chirpkit import InsectClassifier
c = InsectClassifier()
c.load_model()
print('✅ All models loaded successfully')
"

Expected Results

After updating the release, users should see:

  • ✅ No import errors
  • ✅ No scikit-learn version warnings
  • ✅ No misleading GPU warnings on macOS with MPS
  • ✅ Fast model downloads from GitHub Release
  • ✅ Automatic Git LFS fallback if needed

SoundCurious Integration

Once the updated chirpkit-ensemble.zip is uploaded, notify SoundCurious that they can:

  1. Update their Docker build to latest ChirpKit
  2. Models will auto-download with all fixes
  3. No more corruption or version mismatch errors

See SOUNDCURIOUS_FIX.md for their integration instructions.