Skip to content

Python/React chatbot I made for a great French insurance company

Notifications You must be signed in to change notification settings

jear/MNT-chatbot

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MNTBot

MNTBot is a chatbot I made for the Mutuelle Nationale Territoriale. It mainly focuses on managing professional trips and buying train & hotel tickets.

screen app

How to install

You need to have at least python 3.6. You can use pip to install the python requirements.

sudo apt-get install python3 pip
pip install -r requirements.txt

Create the database and the classifier

Run these two scripts to initialize the database and create the classifier :

python train_bot_script.py  # to train the classifier  
python db_create.py  # to initialize the database

Run Duckling

The date and time parsing uses Facebook's Duckling library. It is written in Haskell, so you need to run a Haskell environment : we recommand using stack. When you have downloaded or cloned the repo, cd into it and :

stack setup
stack build
stack exec duckling-example-exe

This will launch a local Duckling server on the port 8000.

Running the back

Then you have to launch the back :

python run.py

The back will run on the port 5000.

Running the chat interface

Here we have two options to run the chat interface :

With Slack

If you want to run the bot on Slack, you need to create a bot account with a valid token, and then set this token as an environment variable. You can then run the interface.

 export SLACK_BOT_TOKEN=xoxb-24767...
 python slackinterface.py

That's it ! You can talk to the bot on Slack.

With the custom front

You can also run the special chat interface which is built in React, and uses react-simple-chatbot. You need to have node and npm installed, and then you can launch the front :

sudo apt-get install nodejs npm
cd reactFront
npm install
npm start

The React app will run on the port 3000. You finally have to launch the middleware, which will connect your front with your back :

python runMiddleware.py

The middleware will run on the port 5001. That's it ! You just have to point your browser to http://localhost:3000.

You can read the doc file for more info (in French, sorry !)

About

Python/React chatbot I made for a great French insurance company

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 93.6%
  • JavaScript 4.9%
  • HTML 1.2%
  • CSS 0.3%