Stdlib scripting glue

The reason Python is preinstalled in every ops runbook is its standard library: file paths, process control, JSON, CLI parsing, and HTTP all ship in the box. A useful script is one file that runs anywhere Python does — no npm install, no lockfile, no build.

pathlib treats paths as objects, folding in what Node splits across path and fs:

import json
from pathlib import Path

for f in sorted(Path("/var/log").glob("*.log")):
    print(f.name, f.stat().st_size)

config = json.loads(Path("config.json").read_text())

json.loads / json.dumps are JSON.parse / JSON.stringify — same mental model, different casing.

subprocess — shelling out, safely

subprocess.run takes the command as a list (no shell-injection string splicing), captures output, and can raise on a non-zero exit:

import subprocess

result = subprocess.run(
    ["git", "status", "--short"], capture_output=True, text=True, check=True
)
print(result.stdout)

argparse — a CLI in five lines

Where a Node script reaches for commander or util.parseArgs, argparse is built in and generates --help for free:

import argparse

parser = argparse.ArgumentParser(description="Tail a log file")
parser.add_argument("path")
parser.add_argument("--lines", type=int, default=10)
args = parser.parse_args()

For quick HTTP, urllib.request is in the box too; scripts that graduate into services usually step up to httpx. Node's builtins have caught up a lot, but the cultural default differs: Python glue stays stdlib-only, which is why a single .py file travels so well between machines and CI images.

The challenge is a JS model of a glue script end to end: an argparse-style parser, a JSON transform, and a faked subprocess.run.

Run it. This models a Python ops script: parse CLI args, run a subprocess, and emit JSON — all stdlib moves.

Loading editor…

That's the contrast tour: indentation and gradual types, uv-driven tooling, the container-and-dunder data model, an explicit event loop, and a stdlib built for glue. You can now read — and write — the everyday Python you'll meet in real codebases.