php-rag - An enterprise-level internal RAG system supporting DeepSeek and PHP for enhanced information processing and decision support.

## Introduction to php-rag php-rag is an enterprise-level internal Retrieval Augmented Generation (RAG) system that supports DeepSeek and PHP. It is designed for internal network applications, combining information retrieval and text generation to enhance information processing efficiency, accuracy, and personalized service experiences. ## Supported LLMs in php-rag php-rag supports a variety of large language models (LLMs), including GPT-4o, Claude-3.5, Llama3.2, Mixtral, Bielik, Gemini2, DeepSeek, DeepSeek-R1-7B, and DeepSeek-Coder-v2. ## Key Features of php-rag The key features of php-rag include: - Support for multiple LLMs such as GPT-4o, Claude-3.5, and Llama3.2. - A database containing over 1,000 websites. - Access via web interface, API, and Command Line Interface (CLI). - Installation options using Docker or local Ollama setup. - Use of PostgreSQL with pg_vector extension for vector storage. - Embedding model: OpenAI's text-embedding-ada-002. - Evaluation metrics: ROUGE, BLEU, METEOR, and criteria-based evaluation (correctness, helpfulness, relevance, etc.). ## Information Processing in php-rag php-rag improves information processing by combining information retrieval and text generation. This approach enhances efficiency and accuracy, provides personalized service experiences, and optimizes service processes and decision support. ## Installation Options for php-rag php-rag can be installed using Docker with docker-compose or locally via Ollama. The local setup requires more CPU/RAM and preferably GPU support. Ollama can be downloaded from [Ollama](https://ollama.com/). ## Configuration of php-rag After installation, php-rag is configured by copying the .env-sample file to .env and setting up the MODEL and other configurations. This includes providing API keys for cloud LLMs like OpenAI, DeepSeek, Claude, and Gemini. ## Primary Functions of php-rag The primary functions of php-rag are: - Text Generation: Generating coherent and contextually appropriate responses based on retrieved information. - Entity Distinction: Differentiating between similar entities, such as individuals with the same name, to handle ambiguous queries with precision. ## Accessing the php-rag Project The php-rag project can be accessed via its GitHub repository at [php-rag](https://github.com/mzarnecki/php-rag). The provided URL [http://wwwGITPP.com/laipang/php-rag](http://wwwGITPP.com/laipang/php-rag) is invalid and likely contains a typo or is outdated. ## Evaluation Metrics in php-rag php-rag uses evaluation metrics such as ROUGE, BLEU, METEOR, and criteria-based evaluation (correctness, helpfulness, relevance, etc., scored 1-5) to assess the quality of generated responses. ## Embedding Model in php-rag php-rag uses OpenAI's text-embedding-ada-002 as the embedding model for document embedding. ### Citation sources: - [php-rag](https://github.com/mzarnecki/php-rag) - Official URL Updated: 2025-03-31