sakthi535 / private-gpt

Upload pdfs and query them with help of almighty OpenAI Models...

Private-gpt : Flask Server

The flask server handles api endpoint for frontend application and connects to OpenAI Models for chat completion with the help of langchain framework.

Deployment

Before deployment ensure to create a config file with required connection details for both elasticsearch database and OpenAI key.

Install requirements with,

  pip install requirements.txt  

Then start the server with

  python server.py

API Reference

Get all items

  POST /addFiles

Parameter	Type	Description
data	FileList	Files to upload in bytes

Get item

  POST /eraseIndex

Parameter	Type	Description
name	string	Name of index to be removed from alias

  GET /get_indices

Response	Type	Description
	List	List of indices with alias

  GET /createResponse

Parameter	Type	Description
question	string	User generated query
index	string	Index associated with the query

  GET /add_index

Parameter	Type	Description
name	string	Name of index to be added

Environment Variables

To run this project, you will need to add the following environment variables to your config.py file

config = {
    "cloud_id" : <Usage of cloud id is not supported currently by codebase>,
    "host" : <elastic-search address>,
    "auth" : [
        {
            "username" : <elasticsearch username>,
            "password" : <elasticsearch password>
        }   
    ],
    "application_name" : "romTalk",
    "Open-AI Key": <OpenAI Key>,
    "mapping" : {
    "properties": {
      "metadata": {
        "properties": {
          "name": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          },
          "type": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          }
        }
      },
      "text": {
        "type": "text"
      },
      "vector": {
        "type": "dense_vector",
        "dims": 1536
      }
    }
  }
}

The above mapping is necessery to store dense vectors generated from OpenAI embedding model in es database.

Todo Next

• Add a relational sql database to track changes and monitor token usage.
• Integrate IAM and security policies

Authors

• @sakthi535