AlterNLU is an open source tool for building chatbot and AI assistant. It gives the developer a minimal approach to build an assistant or bot in production environment.
It converts natural language to structured data. Eg:
Input:
{"text": "i want few covers for my mobile and pillows. and two bed sheet"}
Output:
{
"intent": "search_product",
"confidence": "0.9999974",
"entities": [
{
"value": "bedsheet",
"category": "product_type"
},
{
"value": "Mobile cover",
"category": "product_type"
},
{
"value": "pillow covers",
"category": "product_type"
}
]
}
Once AlterNLU is set up see Setting Up, you can try it out in 4 simple steps.
-
Get Training Data. Recommended : Use Kontiki Console
-
Run AlterNLU Engine.
python3 server.py
-
Rest API training.
http://<ip_address>:5001/train Method : POST Accept / Content_type : application/json Data : training_data file from Kontiki Platform.
curl -H "Content-Type: application/json" --data @<file path> http://localhost:5001/train
-
Rest API parse query.
http://<ip_address>:5001/parse Method : POST Accept / Content_type : application/json Data : {"text": "<your_query>"}
Latest Model will be deployed automatically once training get completed and ready for serving.
- Install Python3 & pip
apt-get update apt-get install python3.6 apt-get install python3-pip
- Setup Virtual Environment
pip3 install virtualenv virtualenv nlu cd nlu source bin/activate
- Install AlterNLU
git clone https://github.com/anuragmishra1/alter-nlu.git cd alter-nlu python3 setup.py install
- Run Server
python3 server.py