🚀 LangChain Practice

A modular FastAPI project built to practice and demonstrate core LangChain concepts through multiple real-world APIs.

This project follows a clean architecture where each task is organized into its own module while sharing a common configuration and LLM setup.

---

📚 Concepts Covered

• FastAPI
• LangChain Expression Language (LCEL)
• PromptTemplate
• RunnableBranch
• RunnableLambda
• RunnableParallel
• RunnablePassthrough
• Structured Output
• Pydantic Models
• Google Gemini Integration
• Environment Configuration
• Modular API Design

---

---

⚙️ Features

✅ Health Check API

Simple API to verify the application is running.

Endpoint

GET /health

---

✅ Task 1 — Smart Text Toolkit

Transforms text using RunnableBranch.

Supported actions:

• Summarize
• Translate
• Tone Shift

Endpoint

POST /transform

Concepts

• PromptTemplate
• RunnableBranch
• StrOutputParser

---

✅ Task 2 — Resume Parser

Extracts structured information from resume text.

Returns

• Name
• Skills
• Years of Experience
• Last Role

Endpoint

POST /parse-resume

Concepts

• Structured Output
• Pydantic Models
• PromptTemplate

---

✅ Task 3 — Query Router

Classifies user questions into

• Coding
• Math
• General

Routes them to domain-specific prompts.

Endpoint

POST /ask

Concepts

• RunnableLambda
• RunnableBranch
• LCEL

---

✅ Task 4 — Parallel Content Generator

Generates multiple social media contents simultaneously.

Returns

• Tweet
• LinkedIn Caption
• Hashtags

Endpoint

POST /generate-post

Concepts

• RunnableParallel
• RunnablePassthrough

---

✅ Task 5 — Self Correcting Answer API

Creates an answer, reviews it, and improves it if required.

**Endpoint**

POST /answer

### Concepts

- Multi-step Chains
- Prompt Chaining
- Self-Correcting Workflow

---

# 🏗 Architecture

Every module follows the same architecture.

Request │ ▼ Router │ ▼ Workflow │ ▼ Prompt │ ▼ Gemini │ ▼ Response

This separation keeps the project clean, reusable, and easy to maintain.

---

# 🛠 Installation

Clone the repository

```bash
git clone <repository-url>

Navigate to the project

cd langchain-practice

Create virtual environment

uv venv

Activate environment

Windows

.venv\Scripts\activate

Install dependencies

uv sync

---

🔑 Environment Variables

Create a .env file.

GOOGLE_API_KEY=YOUR_API_KEY
MODEL_NAME=gemini-2.5-flash

---

▶️ Run the Application

uv run uvicorn main:app --reload

Swagger UI

http://127.0.0.1:8000/docs

---

🧠 LangChain Components Used

• PromptTemplate
• RunnableBranch
• RunnableLambda
• RunnableParallel
• RunnablePassthrough
• StrOutputParser
• Structured Output
• LCEL Pipelines

---

🧪 APIs

Method	Endpoint	Description
GET	/health	Health Check
POST	/transform	Smart Text Toolkit
POST	/parse-resume	Resume Parser
POST	/ask	Query Router
POST	/generate-post	Parallel Content Generator
POST	/answer	Self Correcting Answer

---

📖 Learning Outcomes

This project demonstrates practical implementation of:

• Modular FastAPI Architecture
• LangChain LCEL
• Prompt Engineering
• Structured Output
• Branching Workflows
• Parallel Execution
• Self-Correcting LLM Pipelines
• Reusable Prompt Management
• Environment-based Configuration
• Clean Project Organization

---

👨‍💻 Author

Rishu Raj

Built as a LangChain Practice Assessment to gain hands-on experience with modern LLM application development using FastAPI and LangChain.