Intermediate 22 min Prerequisite: RAG

LangChain Framework

AI applications ko build, chain, aur manage karna seekho LangChain ke saath

Course Progress 5/8 Lessons (62%)

Why Learn LangChain?

LangChain se AI applications build karte hain � chains, agents, tools sab LangChain se manage hote hain. Production AI apps ke liye zaroori hai. Ye framework LLMs ke saath complex workflows ko simple banata hai.

LangChain ke Core Concepts

CHAINS

Sequential operations ka pipeline jo multiple steps ko ek saath connect karta hai

AGENTS

AI with tools � decision making aur dynamic tool usage

PROMPT TEMPLATES

Reusable prompts jo different inputs ke saath kaam karte hain

MEMORY

Conversation history store aur manage karna for context retention

LangChain Basics

LangChain ek Python framework hai jo LLM-powered applications banane ke liye use hota hai. Ye provide karta hai:

Chains - Multiple components ko connect karna
Agents - Dynamic tool selection aur execution
Prompt Templates - Reusable prompt structures
Memory - Conversation context maintain karna
Document Loaders - Different data sources se data load karna
Vector Stores - Embeddings store aur retrieve karna

Code Example: LangChain Basics

# LangChain basics
# pip install langchain openai

# Simple chain
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain

template = """You are a helpful assistant. Answer in Hinglish:
Question: {question}
Answer:"""

prompt = PromptTemplate(template=template, input_variables=["question"])

# RAG with LangChain
from langchain.vectorstores import FAISS
from langchain.embeddings import OpenAIEmbeddings

# documents = ["doc1", "doc2"]
# vectorstore = FAISS.from_documents(documents, OpenAIEmbeddings())
# relevant_docs = vectorstore.similarity_search("query", k=2)

print("LangChain components ready!")
print("Chains, Agents, Memory sab available hain")

Output:

LangChain components ready! Chains, Agents, Memory sab available hain

Prompt Templates Deep Dive

Prompt templates ko different use cases ke liye customize kar sakte ho:

Variable placeholders use karo ({question}, {context})
Few-shot examples add karo for better responses
System instructions define karo for role-playing
Output format specify karo (JSON, list, etc.)

✓ Interactive Editor: Create Your LangChain Chain

Memory Management

LangChain mein different memory types available hain:

ConversationBufferMemory - Poori conversation store karta hai
ConversationSummaryMemory - Summary store karta hai (token save)
ConversationBufferWindowMemory - Sliding window approach
ConversationTokenBufferMemory - Token limit ke saath

Agents Overview

Agents LangChain ka sabse powerful feature hain:

Dynamic Tool Selection - Kaunsa tool use karna hai decide karta hai
Reasoning - Step-by-step thinking karta hai
Loop Execution - Jab tak task complete nahi hota, tools use karta hai
Human-in-the-loop - Zarurat par human se input leta hai

Code Example: Memory Integration

# Memory integration with LangChain
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationChain

# Memory create karo
memory = ConversationBufferMemory()

# Conversation chain with memory
# chain = ConversationChain(
# llm=llm,
# memory=memory,
# verbose=True
# )

# Memory types:
# 1. ConversationBufferMemory - Full history
# 2. ConversationSummaryMemory - Summary
# 3. ConversationBufferWindowMemory - Sliding window

print("Memory types available:")
print("1. Buffer Memory")
print("2. Summary Memory")
print("3. Window Memory")

Output:

Memory types available: 1. Buffer Memory 2. Summary Memory 3. Window Memory

Exercise Time!

LangChain mein Chain kya hai?

LLM ka naam hai

Sequential operations ka pipeline

Ek database type hai

Prompt ka format hai

📌

Key Takeaways

LangChain AI applications ke liye ek complete framework hai
Chains, Agents, Memory - ye three pillars hain LangChain ke
Prompt Templates se reusable prompts bana sakte ho
Memory se conversation context maintain hota hai
Agents dynamic tool usage enable karte hain
Production mein LangChain bahut useful hai