How rishav822/care-connect works
Care Connect is a Flask-based health-information chatbot. Its main workflow accepts a chat message, classifies intent using NLP and a saved neural network, applies conversation-state logic, and returns a response. The app also supports a profile API backed by MongoDB and a guided BMI flow. The README documents a browser chat interface and model-training workflow, though frontend files are absent from the tree. Training and frontend links are based on README statements; runtime links are grounded in the app excerpt.
Main entry points: app.py Conversation flow · train_model.py Model training · intents.json Intent patterns · chatbot_model.h5 Neural model
Components 14
Chat interface
- Web chat
Flask application
- Flask routes
app.py - Conversation flow
app.py - BMI calculation
app.py - Profile API
app.py
NLP and intent
- Text preprocessing
app.py - Intent prediction
app.py - Response selection
app.py - Model training
train_model.py
Model and profile data
- Intent patterns
intents.json - Neural model
chatbot_model.h5 - Vocabulary classes
words.pkl - MongoDB profiles
Other components
- Chat user
Connections 15
- Chat user → Web chat: enters message
README.md - Web chat → Flask routes: posts JSON
README.md - Flask routes → Conversation flow: dispatches message
app.py - Conversation flow → Intent prediction: predicts intent
app.py - Intent prediction → Text preprocessing: vectorizes text
app.py - Intent prediction → Neural model: runs inference
app.py - Intent prediction → Vocabulary classes: uses classes
app.py - Conversation flow → Response selection: selects reply
app.py - Response selection → Intent patterns: matches response
app.py - Conversation flow → BMI calculation: calculates BMI
app.py - Flask routes → Profile API: routes profile
app.py - Profile API → MongoDB profiles: reads and writes
app.py - Model training → Intent patterns: trains from
train_model.py - Model training → Neural model: saves model
train_model.py - Model training → Vocabulary classes: saves vocabulary
train_model.py