README.md

metadata

title: Chatbot Sports Academies System
emoji: 🏆
colorFrom: blue
colorTo: green
sdk: docker
app_port: 7860
pinned: false

Chatbot Sports Academies System

Django-based chatbot for sports academy services, knowledge-base Q&A, service-intent detection, and backend workflow execution.

This Space runs as a Docker Space and exposes Django/Gunicorn on port 7860.

Runtime Flow

User message
  ↓
Input validation / rate limiting
  ↓
Prompt routing policy: live platform request, knowledge question, or follow-up
  ↓
Conservative service candidate ranking with conversation context
  ↓
If high confidence: return SERVICE_REQUEST to the authenticated backend
If ambiguous: return CLARIFICATION_REQUIRED
If knowledge-only: search the CSV knowledge base

The deterministic router is authoritative for service selection. The optional service-intent model enriches already-selected services with flexible parameters; it does not independently choose or execute a service. Explicit live-data requests never fall through to an unrelated knowledge-base answer.

Required CSV datasets

Copy these files into apps/chat/training_models/data/ before production/training:

• data.csv
• service_intents.csv
• service_intents_model_training.csv

Model skip/training logic

Startup runs:

python manage.py ensure_model --remote --wait --download-artifact

Logic:

1. Check SERVICE_INTENT_MODEL_DIR.
2. If Hugging Face model files already exist, skip training.
3. If missing, train remotely using service_intents_model_training.csv.
4. If training fails, app still starts with deterministic service detection.

Environment variables

Set these in Hugging Face Space settings:

DJANGO_SECRET_KEY=change-me
DJANGO_DEBUG=0
DJANGO_ALLOWED_HOSTS=.hf.space,localhost,127.0.0.1
DJANGO_SECURE_SSL_REDIRECT=1
SEMANTIC_SEARCH_ENABLED=False
SERVICE_INTENT_MODEL_DIR=/app/artifacts/service_intent_model
SERVICE_INTENT_TRAINING_SPACE_URL=https://mortadhabbb-train-model-chatbot.hf.space
N8N_BASE_URL=https://mortadhabbb-n8n-service-orchestration.hf.space

Optional secrets:

CHATBOT_API_KEY=
HF_TOKEN=
SERVICE_INTENT_TRAINING_SPACE_API_KEY=

Local run

python -m venv .venv
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

Docker run

docker build -t chatbot-sports-academies .
docker run --rm -p 7860:7860 -e DJANGO_SECRET_KEY=dev -e DJANGO_DEBUG=0 chatbot-sports-academies

Push to Hugging Face

git add .
git commit -m "Regenerate Django chatbot project"
git push space main

If replacing the whole Space repository:

git push --force space main