How azizos1235/projet_mangeoire-intelligente works
This repository appears to combine livestock-management apps, cloud and ESP32 code, and feeding-behavior AI. The clearest documented workflow is feeding telemetry sent by Node-RED to the Forest AI service: it retains per-cow history, computes model features and verdicts, optionally publishes the result, and responds. The app and cloud modules are present, but their product behavior and connections are not documented in the supplied excerpts; those links are therefore omitted. Native app nodes represent platform shells, not inspected Dart features.
Main entry points: service.py Forest AI Service · main.c ESP32 Firmware · train_rf.py Feature Builder
Components 14
Livestock Apps
- Farmer App
MI_application/eleveur_app - Admin App
MI_application/admin_app
Device and Cloud
- ESP32 Firmware
MI_esp32/main/main.c - Cloud API
MI_cloud_server/app/main.py - Profile Module
MI_cloud_server/app/profile.py
Feeding AI
- Forest AI Service
MI_module_AI_Forest/service.py - Feeding History
MI_module_AI_Forest/service.py - Feature Builder
MI_module_AI_Forest/train_rf.py - Forest Classifier
MI_module_AI_Forest/train_rf.py - Regression Data Generator
MI_module_AI_Regression/generate_data.py - Regression Trainer
MI_module_AI_Regression/train_AI.py - Forest Data Generator
MI_module_AI_Forest/generate_data.py
Other components
- Node-RED
- AMQP Broker
Connections 6
- ESP32 Firmware → Node-RED: sends telemetry
- Node-RED → Forest AI Service: POSTs readings
MI_module_AI_Forest/service.py - Forest AI Service → Feeding History: stores readings
MI_module_AI_Forest/service.py - Forest AI Service → Feature Builder: builds features
MI_module_AI_Forest/service.py - Forest AI Service → Forest Classifier: requests verdict
MI_module_AI_Forest/service.py - Forest AI Service → AMQP Broker: publishes results
MI_module_AI_Forest/service.py