Context: Develop a Chatbot using LUIS API from Azure Repository of OpenClassrooms project 10' AI Engineer path
Our role is to develop a chatbot for flight booking purpose:
- BOT Framework and Azure Language understanding (LUIS)
- We expect to achieve a first MVP for FlyMe employees.
The project is using below dataset : Frames dataset.
For Setup:
- Python and VSCode with Jupyter extension to read notebooks.
- Environment settings available below (availaible in requirements.txt)
This repository is part of a 3-repos project :
- Main repo : Notebook for project exploration : this repo
- ChatBot repo : CoreBot with application insights
- Luis repo : CoreBot with application insights
- Notebook : Exploratory Data Analysis
- Data Analysis: Visualisation, Modification
- Generation of LUIS compatible JSON files
- Intent and Entity settings
- Output Json Files
- Luis performance results analysis
- Library for BOT modified compared to official SDK pip (4.15 requested instead of 4.14)
- requirements version for emoji==1.7 and aiohttp==3.6.2 and pytest==6.2.3
- requirements library asyncio removed due to compatibility issues
- publish-profile is unique to any Webapp resource, it should never be communicated outside of the project workflow from GitHub to Azure
- GitHub secrets and Azure environment variables are key sensitive (Pay attention to scripts calling them)
- PORT in config.py set to 8000, to reach URL as deployed
- init_func in app.py should match with Azure starting command : python -m aiohttp.web -H 0.0.0.0 -P 8000 app:init_func
- scripts using resolve method to reach endpoint (instead of get_slot_production recommended in azure doc)