How birdnet-team/birdnet-analyzer works
Project Type and Purpose
• BirdNET-Analyzer is a Python-based, open-source acoustic analysis tool.
• It offers both command-line (CLI) and graphical (GUI) interfaces for running deep-learning inference on audio, plus modules for segmentation, search, embeddings, evaluation, and training.
File Structure to Derive Components
• Top‐level Package: birdnet_analyzer/ – the Python package housing all functionality.
• User Interfaces:
• cli.py and subfolders (analyze/cli.py, search/cli.py, etc.)
• gui/ – HTML, CSS, JS assets plus Python wrappers.
• Core Processing Engines (for each domain):
• analyze/core.py
• segments/core.py
• search/core.py
• embeddings/core.py
• evaluation/assessment/... and evaluation/preprocessing/...
• train/core.py
• Model & Configuration:
• model.py – loads the neural net
• config.py – central config management
• eBird_taxonomy_codes_2024E.json – taxonomy data
• labels/ – multilingual label files
• Networking:
• network/server.py & network/client.py – optional remote-processing API.
• Utilities & Shared Logic:
• utils.py and */utils.py in each submodule.
• Packaging & Deployment:
• Dockerfile and .github/workflows/docker-build.yml
• pyproject.toml – build config.
• Tests: tests/ to validate each module.
README for Additional Insights
• Confirms dual‐interface (CLI + GUI), Docker support, and that models are fetched from Zenodo.
• Emphasizes extensibility for single files vs. bulk datasets.