Install LightRAG Locally - Moduler RAG and Lightning Library for LLM Apps

This video is a step-by-step easy tutorial to install LightRAG which helps developers with both building and optimizing Retriever-Agent-Generator pipelines. It is light, modular, and robust, with a 100% readable codebase.

Code:

conda create -n lightrag python=3.11 -y && conda activate lightrag

git clone https://github.com/SylphAI-Inc/LightRAG.git && cd LightRAG

pip install lightrag

pip install openai==1.12.0

pip install faiss-cpu==1.8.0

pip install sqlalchemy==2.0.30

pip install pgvector==0.2.5

pip install groq==0.5.0

mv .env_example .env   #set your openai and groq api keys in .env file.

conda install jupyter -y

pip uninstall charset_normalizer -y

pip install charset_normalizer

jupyter notebook

from lightrag.utils import setup_env

setup_env()

from dataclasses import dataclass, field

from lightrag.core import Component, Generator, DataClass

from lightrag.components.model_client import GroqAPIClient

from lightrag.components.output_parsers import JsonOutputParser

@dataclass

class QAOutput(DataClass):

    explanation: str = field(

        metadata={"desc": "A brief explanation of the concept in one sentence."}

    )

    example: str = field(metadata={"desc": "An example of the concept in a sentence."})

qa_template = r"""<SYS>

You are a helpful assistant.

<OUTPUT_FORMAT>

{{output_format_str}}

</OUTPUT_FORMAT>

</SYS>

User: {{input_str}}

You:"""

class QA(Component):

    def __init__(self):

        super().__init__()

        parser = JsonOutputParser(data_class=QAOutput, return_data_class=True)

        self.generator = Generator(

            model_client=GroqAPIClient(),

            model_kwargs={"model": "llama3-8b-8192"},

            template=qa_template,

            prompt_kwargs={"output_format_str": parser.format_instructions()},

            output_processors=parser,

        )

    def call(self, query: str):

        return self.generator.call({"input_str": query})

    async def acall(self, query: str):

        return await self.generator.acall({"input_str": query})

       

qa = QA()

print(qa)

# call

output = qa("What is LLM?")

print(output)

qa.generator.print_prompt(

        output_format_str=qa.generator.output_processors.format_instructions(),

        input_str="What is LLM?",

)