https://github.com/kolman-freecss/kf-rag-wowinfo

AI RAG | FastAPI | Python | Vector DB | docker
https://github.com/kolman-freecss/kf-rag-wowinfo

ai docker fastapi python rag

Last synced: 3 months ago
JSON representation

AI RAG | FastAPI | Python | Vector DB | docker

• Host: GitHub
• URL: https://github.com/kolman-freecss/kf-rag-wowinfo
• Owner: Kolman-Freecss
• Created: 2025-03-13T18:51:24.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-03-13T22:46:24.000Z (over 1 year ago)
• Last Synced: 2025-03-24T21:40:31.365Z (over 1 year ago)
• Topics: ai, docker, fastapi, python, rag
• Language: Python
• Homepage:
• Size: 18.6 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# KF RAG WoW Information

A Retrieval-Augmented Generation (RAG) application for World of Warcraft information. This project enhances large language model responses with accurate and up-to-date WoW game information.

## Overview

This application uses RAG architecture to:

- Index and store World of Warcraft information in a vector database (ChromaDB)

- Retrieve relevant context based on user queries

- Generate accurate responses by augmenting LLM (Gemini) with retrieved information

- Provide specific, factual answers about WoW lore, mechanics, classes, raids, etc.

## Technology Stack

- Python

- ChromaDB for vector storage

- Google Gemini API

- Docker & Docker Compose

- LangChain for RAG implementation

## Setup

1.  Install Docker and Docker Compose.

2.  Clone this repository.

3.  Create a `.env` file in the root directory with the `GEMINI_API_KEY` environment variable. You can use the `.env.template` file as a template.

4.  Run `docker-compose up --build` to start the ChromaDB and RAG application.

## Usage

1.  Run `docker-compose up --build` to start the ChromaDB and RAG application.

2.  Access the API endpoints:

    *   `/query`:  Send a GET request to `http://localhost:8000/query?query=&num_results=&creativity=` to ask a question about World of Warcraft.

        *   `num_results`: (optional) The number of search results to retrieve (default: 5).

        *   `creativity`: (optional) The creativity of the response (0.0-1.0) (default: 0.5).

    *   `/feedback`: Send a POST request to `http://localhost:8000/feedback` with `query_id` and `feedback` in the request body to provide feedback on the answers.

    *   `/load_data`: Send a POST request to `http://localhost:8000/load_data` to reload the data from the CSV file into the ChromaDB collection.

## API Endpoints

* `/query`: Send a GET request to `http://localhost:8000/query?query=&num_results=&creativity=&max_length=&response_format=&additional_context=` to ask a question about World of Warcraft.

    * `query`: (required) The question to ask.

    * `num_results`: (optional) The number of search results to retrieve (default: 5).

    * `creativity`: (optional) The creativity of the response (0.0-1.0) (default: 0.5).

    * `max_length`: (optional) The maximum length of the response.

    * `response_format`: (optional) The format of the response.

    * `additional_context`: (optional) Additional context to provide to the model.

* `/feedback`: Send a POST request to `http://localhost:8000/feedback` with `query_id` and `feedback` in the request body to provide feedback on the answers.

* `/summarize`: Send a POST request to `http://localhost:8000/summarize` with `document_id`, `document_text`, `urls`, `summary_length`, and `summary_style` in the request body to summarize a document.

    * `document_id`: (optional) The ID of the document to summarize.

    * `document_text`: (optional) The text of the document to summarize.

    * `urls`: (optional) A list of URLs to summarize.

    * `summary_length`: (optional) The length of the summary (default: medium).

    * `summary_style`: (optional) The style of the summary (default: general).

* `/compare`: Send a POST request to `http://localhost:8000/compare` with `document1_id`, `document1_text`, `document2_id`, and `document2_text` in the request body to compare two documents.

    * `document1_id`: (optional) The ID of the first document to compare.

    * `document1_text`: (optional) The text of the first document to compare.

    * `document2_id`: (optional) The ID of the second document to compare.

    * `document2_text`: (optional) The text of the second document to compare.

* `/translate`: Send a POST request to `http://localhost:8000/translate` with `text` and `target_language` in the request body to translate a text.

    * `text`: (required) The text to translate.

    * `target_language`: (required) The target language.

* `/multi_turn`: Send a POST request to `http://localhost:8000/multi_turn` with `query` and `session_id` in the request body to handle multi-turn conversations.

    * `query`: (required) The question to ask.

    * `session_id`: (required) The session ID.

* `/new_session`: Send a GET request to `http://localhost:8000/new_session` to create a new session ID for multi-turn conversations.

* `/generate_questions`: Send a POST request to `http://localhost:8000/generate_questions` with `document_id`, `document_text`, and `num_questions` in the request body to generate questions based on a document.

    * `document_id`: (optional) The ID of the document to generate questions from.

    * `document_text`: (optional) The text of the document to generate questions from.

    * `num_questions`: (optional) The number of questions to generate (default: 5).

* `/paraphrase`: Send a POST request to `http://localhost:8000/paraphrase` with `text` in the request body to paraphrase a text.

    * `text`: (required) The text to paraphrase.

* `/extract_entities`: Send a POST request to `http://localhost:8000/extract_entities` with `text` in the request body to extract entities from a text.

    * `text`: (required) The text to extract entities from.

* `/admin/add_document`: Send a POST request to `http://localhost:8000/admin/add_document` with `document`, `metadata`, and `doc_id` in the request body to add a document to the knowledge base. Requires authentication.

    * `document`: (required) The document to add.

    * `metadata`: (required) The metadata of the document.

    * `doc_id`: (required) The ID of the document.

* `/admin/update_document`: Send a POST request to `http://localhost:8000/admin/update_document` with `doc_id`, `document`, and `metadata` in the request body to update a document in the knowledge base. Requires authentication.

    * `doc_id`: (required) The ID of the document to update.

    * `document`: (optional) The updated document.

    * `metadata`: (optional) The updated metadata.

* `/admin/delete_document`: Send a DELETE request to `http://localhost:8000/admin/delete_document?doc_id=` to delete a document from the knowledge base. Requires authentication.

    * `doc_id`: (required) The ID of the document to delete.

* `/admin/reload_data`: Send a POST request to `http://localhost:8000/admin/reload_data` to reload the data from the CSV file into the ChromaDB collection. Requires authentication.

## Docs

- https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models?hl=es-419