A contextual Chatbot build with TensorflowJS and NodeJS
Install dependencies and run.
$ npm cd ChatBotNodeJS
$ npm install
$ npm start
Install dependencies and run with sample train.
$ npm cd ChatBotNodeJS
$ npm install
$ npm run seed
$ npm start
$ npm cd ChatBotNodeJS
$ docker network create backend-bot
$ docker network create frontend-bot
$ docker-compose up --build
$ docker pull docker.pkg.github.com/ran-j/chatbotnodejs/chatbotnodejs:latest
$ docker run --publish 3000:3000 --detach --name chatbotnodejs chatbotnodejs:latest
Change mongoURI in ./Libs/BotConfig.js
- To interact with bot: http://localhost:3007/
- To train bot intents: http://localhost:3007/admin/intents
- To train bot synonyms: http://localhost:3007/admin/synonyms
You can register your intents and the build agent.
Register | Train | Anwser |
---|---|---|
- Auth
- User CRUD
- Fallback dashboard
- Token for requests
- Metricis (Count request numbers ...)
- Docker
- Update Tensorflow JS
- Log conversation and fallbacks
- Data table for intents list
- Optimizations on train and answer