LangChain — Aasan Chatbot Tutorial

Preparation: Zaruri Tools 🛠️

Sabse pehle kuch packages install karne honge:

pip install langchain-huggingface langchain-ollama

Ye packages aapko Hugging Face embeddings aur Ollama chat models ko LangChain ke andar use karne me madad karenge.

Code Explanation — Short Me 🧐

• langchain-huggingface: Hugging Face ke models ko LangChain me integrate karne ke liye package.
• langchain-ollama: Ollama open-source models ko LangChain me use karne ka support.

Model Selection: Open-Source Ke Sitare 🌟

Hum do models use karenge:

• Hugging Face se ek embedding model (text ko vector me badalne ke liye)
• Ollama se ek chat model (actual conversational responses ke liye)

Hugging Face Embedding Model 📊

Example code — embedding model initialize karna:

from langchain_huggingface import HuggingFaceEmbeddings

embeddings = HuggingFaceEmbeddings(
    model_name="BAAI/bge-base-en-v1.5"
)

Code ka matlab: HuggingFaceEmbeddings text ko vectors me convert karta hai. model_name="BAAI/bge-base-en-v1.5" ek strong semantic embedding model hai jo text ka meaning acche se capture karta hai.

Ollama Chat Model 💬

Example code — chat model initialize karna:

from langchain_ollama import ChatOllama

model = ChatOllama(
    model="llama3",
    temperature=0.7
)

Code ka explanation: ChatOllama Ollama ke open-source language model ko LangChain me integrate karta hai. temperature creativity control karta hai (0–1). 0.7 matlab thodi creativity, par zyada precise jawab bhi milenge.

Chatbot Logic: Magic Shuru! 🪄

Niche ka code dikhata hai kaise prompt template aur chain banakar model ko call karte hain:

from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough

# Prompt template banate hain
prompt = ChatPromptTemplate.from_template(
    "Aap ek friendly chatbot hain. {question} ka jawab dijiye."
)

# Chain create karte hain
chain = (
    {"question": RunnablePassthrough()} 
    | prompt 
    | model
)

# Bot se baat karte hain
response = chain.invoke("Aap kaun hain?")
print(response)

Code ka gehrayi se explanation:

• ChatPromptTemplate: Custom prompt template banane ka tool — {question} jagah par actual user question jayega.
• RunnablePassthrough(): Input ko chain me seedha aage pass karta hai — flexibility badhata hai.
• Chain Creation (| operator): Ye operator components ko connect karta hai: Question → Prompt → Model.
• chain.invoke(): Actual conversation start karta hai — model se response nikalta hai.

Pro Tips 💡