0.4 The Python labs: what you build for each guide in the series

All labs live in one uv project:

uv init --package py-labs && cd py-labs
mkdir -p labs                       # each lab is labs/<name>.py
uv run labs/wordfreq.py README.md   # run one lab

Series guide	Python labs in this guide
OS & Linux	`logstat` (Ch 12), `supervisor` with graceful SIGTERM (34), `atomicwrite` (35), `pyps` from `/proc` (36), `shm` shared memory (37), `minibox` namespace container (39)
Networking (TCP/IP)	four `echo` servers (40), `dnsq` raw-bytes DNS client (41), `tinyhttp` (42), `healthcheck` (46)
Security from Zero	`seal` AES-256-GCM (48), Argon2id password store (48), HMAC webhook verify (48), `authsvc` JWT service (51)
Security in Depth	`pickle` exploit and safe loading (47), `safefetch` SSRF guard (43), `pip-audit` in CI (50), `detect` log detector (52)
HTTPS Lifecycle	`client` with retries and idempotency keys (43), `api` production FastAPI service (44), TLS/mTLS servers (45), `requests` trace (64)
AI from Zero to LLMs	ticket classifier (56), `tinygrad` autograd (57), PyTorch MLP and char-LM (58), JSONL corpus builder (60), RAG + agent + MCP server (61)
Go / Rust	`lru` (24, compare Rust §27), `hasher` (32), the capstone `lifecycle` (72, compare Rust §64)

The packages used across the guide (versions as written):

[project]
name = "py-labs"
version = "0.1.0"
requires-python = ">=3.13"
dependencies = [
    "httpx>=0.28",
    "requests>=2.32",
    "fastapi>=0.120",
    "uvicorn>=0.37",
    "pydantic>=2.12",
    "cryptography>=46",
    "argon2-cffi>=25.1",
    "pyjwt>=2.10",
    "numpy>=2.3",
    "pandas>=2.3",
    "polars>=1.33",
    "matplotlib>=3.10",
    "scikit-learn>=1.7",
    "rich>=14",
]

[project.optional-dependencies]
ml = ["torch>=2.8", "transformers>=5", "datasets>=4.1", "tokenizers>=0.22", "safetensors>=0.6", "mcp[cli]>=2"]

[dependency-groups]
dev = ["pytest>=8.4", "hypothesis>=6.140", "ruff>=0.13", "mypy>=1.18", "pip-audit>=2.9"]