How to Build a RAG Chatbot in LangChain (Python)

python3 -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate

pip3 install langchain langchain-groq langchain-huggingface \
    langchain-community langchain-text-splitters \
    faiss-cpu pypdf sentence-transformers truststore

export GROQ_API_KEY="gsk_..."

import os
import sys
from langchain_community.document_loaders import PyPDFLoader
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS

INDEX_DIR = "faiss_index"
EMBED_MODEL = "all-MiniLM-L6-v2"


def ingest(pdf_path: str, index_dir: str = INDEX_DIR):
    if not os.path.exists(pdf_path):
        print(f"Error: file not found — {pdf_path}")
        sys.exit(1)

    print(f"Loading {pdf_path}...")
    loader = PyPDFLoader(pdf_path)
    docs = loader.load()
    print(f"  Loaded {len(docs)} pages")

    splitter = RecursiveCharacterTextSplitter(
        chunk_size=500,
        chunk_overlap=60,
    )
    chunks = splitter.split_documents(docs)
    print(f"  Split into {len(chunks)} chunks")

    print(f"  Embedding with {EMBED_MODEL} (local, no API key)...")
    embeddings = HuggingFaceEmbeddings(model_name=EMBED_MODEL)
    vectorstore = FAISS.from_documents(chunks, embeddings)
    vectorstore.save_local(index_dir)
    print(f"  Index saved to ./{index_dir}/")
    print("Done. Run chatbot.py to start chatting.")


if __name__ == "__main__":
    pdf = sys.argv[1] if len(sys.argv) > 1 else "document.pdf"
    ingest(pdf)

python3 ingest.py report.pdf

import os
import truststore
truststore.inject_into_ssl()  # fixes macOS TLS cert error

from langchain_groq import ChatGroq
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.output_parsers import StrOutputParser
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_community.chat_message_histories import ChatMessageHistory
from operator import itemgetter

INDEX_DIR = "faiss_index"
EMBED_MODEL = "all-MiniLM-L6-v2"
GROQ_MODEL  = "llama-3.3-70b-versatile"

if not os.path.exists(INDEX_DIR):
    print("Index not found. Run 'python3 ingest.py <your_document.pdf>' first.")
    raise SystemExit(1)

# ── Load vector store ──────────────────────────────────────
embeddings = HuggingFaceEmbeddings(model_name=EMBED_MODEL)
vectorstore = FAISS.load_local(
    INDEX_DIR, embeddings, allow_dangerous_deserialization=True
)
retriever = vectorstore.as_retriever(search_kwargs={"k": 4})

# ── LLM ───────────────────────────────────────────────────
llm = ChatGroq(model=GROQ_MODEL, temperature=0)

# ── Prompt ────────────────────────────────────────────────
system_prompt = """You are a helpful assistant that answers questions \
about a document. Use only the context provided below to answer. \
If the answer is not in the context, say "I don't know based on the document."

Context:
{context}"""

prompt = ChatPromptTemplate.from_messages([
    ("system", system_prompt),
    MessagesPlaceholder(variable_name="history"),
    ("human", "{question}"),
])

# ── Chain ─────────────────────────────────────────────────
def format_docs(docs):
    return "\n\n".join(doc.page_content for doc in docs)

chain = (
    {
        "context":  itemgetter("question") | retriever | format_docs,
        "question": itemgetter("question"),
        "history":  itemgetter("history"),
    }
    | prompt
    | llm
    | StrOutputParser()
)

# ── Memory ────────────────────────────────────────────────
store = {}

def get_history(session_id: str):
    if session_id not in store:
        store[session_id] = ChatMessageHistory()
    return store[session_id]

chat = RunnableWithMessageHistory(
    chain,
    get_history,
    input_messages_key="question",
    history_messages_key="history",
)

config = {"configurable": {"session_id": "default"}}

# ── Chat loop ─────────────────────────────────────────────
print(f"RAG Document Chatbot (Groq / {GROQ_MODEL}) — type 'quit' to exit\n")
while True:
    try:
        question = input("You: ").strip()
    except (EOFError, KeyboardInterrupt):
        print("\nBye!")
        break
    if question.lower() in ("quit", "exit", "q"):
        break
    if not question:
        continue
    answer = chat.invoke({"question": question}, config=config)
    print(f"Bot: {answer}\n")