Python for Gen AI

A hands-on learning workspace for Python fundamentals and the building blocks commonly used in Gen AI apps — HTTP clients, FastAPI, Pydantic, LangChain, LangGraph, CrewAI, and MCP (Model Context Protocol).

Repository map

• 1_python_basics/: Core Python (types, strings, collections, control flow, functions, files, exceptions, OOP, decorators).
• 2_packages/: A tiny package + unit tests (good for learning imports + tests).
• 3_fastapi/: A small FastAPI app demonstrating common API patterns.
• 4_requests_basics/: Runnable requests examples (auth, retries, sessions).
• 5_basic_pydantic/: Pydantic basics + a few advanced patterns.
• 6_httpx_basics/: httpx sync/async/streaming patterns, plus LLM-style examples.
• 7_jupyter_notebook/: Notebook(s) for interactive exploration.
• 8_langchain/: LangChain v0.3 components — chains, prompts, memory, tools, agents, and LCEL.
• 9_langserve/: Serving LangChain chains as REST APIs with LangServe.
• 10_langgraph/: LangGraph — stateful multi-agent workflows with graph-based control flow, HITL, and multi-agent patterns.
• 11_crewai/: CrewAI — multi-agent orchestration from basics to advanced (tools, flows, hierarchical crews, guardrails, callbacks).
• 12_mcp/: MCP (Model Context Protocol) — full coverage from basics to production patterns: tools, resources, prompts, transports, server context, lifespan, caching, and LangChain integration.
• projects/: Larger "projects" sample projects (complete mini-apps you can run end-to-end).

Projects

Smart Study Assistant (projects/smart-study-assistant/)

Multi-role CLI assistant that routes tasks between OpenAI / Gemini / Claude based on the query type/complexity.

cd "projects/smart-study-assistant"
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Set at least one API key (recommended: use a local .env file)
# OPENAI_API_KEY=...
# GEMINI_API_KEY=...
# ANTHROPIC_API_KEY=...

python main.py chat

Smart Study Assistant Simple (projects/smart-study-assistant-simple/)

A streamlined version of the Smart Study Assistant, focusing on core logic.

cd "projects/smart-study-assistant-simple"
# ... setup steps similar to above ...

Function Calling: Weather (projects/function_calling/)

Small demo showing OpenAI tool/function calling that fetches current weather from OpenWeatherMap (city name or ZIP+country).

cd "projects/function_calling"
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Create a .env file (see env.example)
# OPENAI_API_KEY=...
# WEATHER_API_KEY=...

python main.py

FAQ RAG System (projects/faq-rag-system/)

A Retrieval-Augmented Generation (RAG) system for answering FAQs interactive CLI, supporting OpenAI and local embeddings (Sentence Transformers).

cd "projects/faq-rag-system"
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Create .env with OPENAI_API_KEY if using OpenAI
# Or use local embeddings (see README)

python main.py

LLM API Wrapper (projects/llm-api-wrapper/)

Unified wrapper for multiple LLM providers.

cd "projects/llm-api-wrapper"
# ... setup steps ...

Setup (recommended)

Create a virtual environment per module (each folder may have its own requirements.txt).

# From this repo root
cd "3_fastapi"  # or 4_requests_basics / 6_httpx_basics / ...

python3 -m venv .venv
source .venv/bin/activate

pip install -r requirements.txt  # if this module has one

Run examples

Python scripts

python3 1_python_basics/1_print_functions.py
python3 4_requests_basics/01_get_json.py
python3 6_httpx_basics/01_httpx-basics.py

MCP (Model Context Protocol)

cd "12_mcp"
python3 -m venv .venv
source .venv/bin/activate
pip install fastmcp mcp python-dotenv

python 1_mcp_basics.py          # protocol basics, server & client
python 2_mcp_tools.py           # all tool definition patterns
python 3_mcp_resources.py       # static, dynamic & template resources
python 4_mcp_prompts.py         # reusable prompt templates
python 5_mcp_transports.py      # stdio, SSE & in-process transports
python 6_mcp_with_langchain.py  # MCP tools inside LangChain agents
python 7_mcp_context_and_state.py  # lifespan, caching, progress reporting

FastAPI

cd "3_fastapi"
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

uvicorn main:app --reload

Then open:

• Docs: http://127.0.0.1:8000/docs
• Health: http://127.0.0.1:8000/health

Notes

• Internet required for many HTTP examples (requests/httpx).
• API keys: if you run any LLM/OpenAI-style examples, prefer environment variables (e.g. OPENAI_API_KEY) and keep secrets out of git.