Auto reply suggestions to chat messages/emails built using rasa_nlu framework.
- Make sure you have git, python, virtualenv and npm installed
- Clone the repository
git clone https://github.com/varunon9/chat-reply-suggestions.git
- Go to project directory
cd chat-reply-suggestions/
- Create a virtual environment
virtualenv venv
- Activate virtual environment
source venv/bin/activate
- Install Rasa NLU as well as spacy
pip install rasa_nlu[spacy] --default-timeout=100
- Install language model for the english language
python -m spacy download en_core_web_md
- Link to model data
python -m spacy link en_core_web_md en
- Install some additional dependencies
pip install -r requirements.txt
- Place your nlu_data.json file inside data directory (a sample file is provided).
- Train the intent models using below command
python -m rasa_nlu.train --config config_spacy.yml --data data/nlu_data.json --path projects
- Install node dependencies
npm install
- Start nlu server in current terminal
python -m rasa_nlu.server --path projects
- Open another terminal in same project directory (no need to activate virtual environment)
- Start the app
node index.js
- The above app uses sample data (collected by me) for demo. Bot can be made more efficient by training with more real world data.
- Currently suggestions are bunch of hardcoded arrays based on intent. Entity recognition can be done and suggestions can be generated on the fly.