log rag

gistfile1.txt

from langchain_community.llms import Ollama
from langchain_community.document_loaders import WebBaseLoader
from langchain_community.document_loaders import DirectoryLoader
from langchain_community.embeddings import OllamaEmbeddings
from langchain_community.vectorstores import FAISS
from langchain_core.prompts import ChatPromptTemplate
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain.chains import create_retrieval_chain
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
from langchain.callbacks.manager import CallbackManager

from rich.console import Console
from rich.progress import Progress
from rich.panel import Panel
from rich.text import Text
from rich.markdown import Markdown
from rich.prompt import Prompt
import time

def rag_chat_init(dir_path, filename):
    console = Console()

    callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])

    # Invoke chain with RAG context
    llm = Ollama(model="phi3", callbacks=callback_manager)

    # Invoke chain with RAG context
    loader = DirectoryLoader(
        dir_path, 
        glob=filename,
        use_multithreading=True
    )
    with console.status("[cyan]Loading log files..."):
        docs = loader.load()
    console.print(f"Total log files: {len(docs)}")

    with console.status("[cyan]Creating vector store..."):
        # Vector store things
        embeddings = OllamaEmbeddings(model="nomic-embed-text")
        text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
        split_documents = text_splitter.split_documents(docs)
        vector_store = FAISS.from_documents(split_documents, embeddings)

    # Construct prompt
    prompt = ChatPromptTemplate.from_template(
        """
        You are in an intelligent rag system named logrctx for log analysis and given some extracted parts from logs as context through rag system along with a question to answer.
        If you don't know the answer, just say "I don't know." Don't try to make up an answer.

        Don't provide any information that is not directly relevant to the question. Like debugging information, reasoning, recommendation or any extra context unless asked. 
        Just provide what's asked from the given context by summarizing the context.

        Prefer to use markdown format wherever possible for visually appealing output and use time from the logs in your response if needed for consise response.        
        Keep the response as short and to the point as possible while not leaving any important information out.

        Use only the following pieces of context to answer the question at the end.

        Context: {context}

        Question: {input}
        """
    )

    # Custom retrieval and generation process
    def custom_retrieval_chain(query):
        with console.status("[cyan]Retrieving relevant logs..."):
            docs = vector_store.similarity_search(query, k=5)
            console.print("[green]Context mapped successfully.")
            console.print("Retrieved docs 👇 ")
            for doc in docs:
                console.print(Panel.fit("[cyan]" + str(doc.metadata['source']) + "[/cyan]\n" + str(doc.page_content)))

        console.print("[cyan]Generating response...")
        response = docs_chain.invoke({"context": docs, "input": query})
        
        return response

    # Retrieve context from vector store
    docs_chain = create_stuff_documents_chain(llm, prompt)
    #retriever = vector_store.as_retriever()

    # Winner winner chicken dinner
    console.print("\n[bold cyan]Invoking chain...")
    while True:
        query = Prompt.ask("[bold green]Prompt[/bold green]")

        if query.lower() in ["exit", "quit"]:
            break

        response = custom_retrieval_chain(query)

        print("\n")
        console.print(Panel.fit("[bold green] logrctx ai 🧠 [/bold green]"))
        console.print(Panel.fit(Markdown(f"{response}")))

if __name__ == "__main__":
    rag_chat_init(
        dir_path="../logs",
        filename="reduced_sample.log"
    )

Can you post an example of output ?