A custom knowledge chatbot is a specialised artificial intelligence program designed to interact with users and provide accurate and tailored information based on a specific domain or topic. Unlike generic chatbots, custom knowledge chatbot is trained on a specific dataset which allows it to answer queries with higher precision and expertise in that particular area.
IPF Developer Documentation, similar to most vendor product manuals, is designed to be a longish book/tutorial type read. A chatbot style Q/A utility may be better suited for rapid learning on the job.
Link to IPF documentation:
Most of the topics from Core and Learn IPF are included as data for this chatbot.
- Integrate the IPF Developer Documentation into a Chatbot model, allowing it to access and utilize the information for answering specific queries related to IPF development.
- Enhance the model's natural language understanding capabilities to accurately interpret user queries, even when they are complex or contain technical terms related to IPF development.
- The model should be capable of understanding the intent behind the question and retrieving the most appropriate information from the knowledge base.
- Data collection - Web page scraping
- Data processing -
- Generating embeddings - vectors stored by using chromadb
- Saving embeddings to local DB - stored in a specific directory.
- Inference - Use embedding to answer queries
The files for web scraping is available in the data_collection folder and the files for the embedding and inference are available in the chatBot folder
If there is any change in the data that is used,make sure the correct data is present in the data directory. Also if there is any change in data,adjust the db directory as well.