How ecoconvert/ecoconvert-raspi works
Project type and purpose
• This is an edge-device Python application for real-time object detection/vision processing using TensorFlow Lite on a Raspberry Pi (or similar). It orchestrates camera capture, model inference, LED feedback, and logging, with helper scripts for testing and deployment.
File structure
• Top-level files and dirs:
• main.py – entry point orchestrating workflow
• logging_config.py – configures file‐based logging
• requirements.txt – Python deps
• test.sh / upd.sh – deployment & update scripts
• Models/ – two pre‐trained TFLite models + label maps + SavedModel dirs
• process/ – “business logic” subpackage
• blink.py – blink‐detection logic
• inserts.py – insertion or record logic
• modules/Camera.py – camera hardware interface
• modules/Led.py – LED hardware interface
• tester/webcam_tflite.py – standalone TFLite webcam tester
• logs/ – output logs
README
• Details on branch strategy (deployment vs. playground)
• Virtual‐env naming strictness (“env”)
• .env file for MODEL_PATH and LABEL_PATH
• Emphasis on activating env before install
Detail
• Configuration as an injected dependency (.env → main.py)
• Test.sh/upd.sh flows from version control to runtime