Bird Call Pi Email Alerts and Cloud Uploads

Practical Summary

Field-tested Raspberry Pi scripts for BirdNET-Go users who want a customisable bird-call e-mail alert, with confidence thresholds, attached audio clips, call isolation in bird call choruses, and safer cloud uploads that verify files before deleting local WAVs to ensure memory cards do not fill.

What This Is Useful For

• Passive acoustic monitoring with a Raspberry Pi and BirdNET-Go.
• Getting e-mail alerts for chosen priority species.
• Attaching original WAV clips and optional filtered review FLACs to alerts.
• Creating review packages with copied original audio, filtered listening aids, spectrograms, metadata, and a review log.
• Uploading audio and detections to an rclone remote while protecting newly-created files during long uploads.

Safety Defaults

• Secrets stay outside Git in an env file.
• Raw audio, logs, databases, archives, keys, and data folders are ignored.
• Weekly upload uses a start-of-run WAV manifest, verifies the uploaded files, then deletes only verified manifest files.
• Location columns are excluded from weekly CSV exports by default. Set INCLUDE_LOCATION=1 only if your destination is appropriate for location data.

Main Scripts

File	Purpose
scripts/alert_poller.py	Polls a BirdNET-Go SQLite database for new high-confidence priority species and triggers e-mail alerts.
scripts/alert_email.py	Sends an e-mail with the original WAV and, when available, a filtered review FLAC.
scripts/review_artifacts.py	Builds review artefacts for selected detections. Requires ffmpeg and sox for audio/spectrogram outputs.
scripts/upload_weekly.sh	Uploads audio and detections through rclone with verify-before-delete behaviour for audio.

Quick Start

1. Install BirdNET-Go on your Raspberry Pi and confirm it is saving detections and WAV clips.
2. Copy examples/email.env.example to an env file outside the repository, for example /opt/bird-call-alerts/secrets/email.env.
3. Fill in your SMTP details and PRIORITY_SPECIES.
4. Run the poller manually once before adding timers:

python3 scripts/alert_poller.py

5. Dry-run the upload script before enabling it:

bash scripts/upload_weekly.sh --dry-run

See docs/INSTALL.md for setup notes and docs/SECURITY_AND_PRIVACY.md before making a public deployment.

Contributing And Project Status

• Please read CONTRIBUTING.md before opening issues or pull requests.
• Report security or privacy concerns using SECURITY.md.
• See ROADMAP.md for planned maintainer-friendly improvements.
• Release notes are tracked in CHANGELOG.md.

Configuration

Useful environment variables:

Variable	Default	Purpose
BIRD_CALL_EMAIL_ENV	/opt/bird-call-alerts/secrets/email.env	Path to the e-mail env file.
BIRD_CALL_STATION_LABEL	Bird monitoring station	Generic label used in e-mail subjects/bodies. Can also be set as STATION_LABEL inside the env file.
BIRD_CALL_DB_PATH	/data/detections/birdnet.db	BirdNET-Go SQLite database.
BIRD_CALL_AUDIO_DIR	/data/audio	BirdNET-Go WAV clip root.
BIRD_CALL_REVIEW_DIR	/data/review	Review artefact root.
REMOTE	gdrive:bird-call-monitoring	rclone remote/path used by upload_weekly.sh.
AUDIO_DIR	/data/audio	Upload script audio root.
DETECTIONS_DIR	/data/detections	Upload script detection/database root.
LOG_DIR	/data/logs	Upload script log root.
INCLUDE_LOCATION	0	Set to 1 to include latitude/longitude in weekly CSV exports.

Tests

python3 -m unittest discover -s tests -p "test_*.py"
bash tests/test_upload_weekly.sh
bash scripts/check-public-repo.sh

Licence

MIT. Use it, adapt it, and improve it, but review the safety/privacy notes first.