Tool for seeing which Python stdlib modules have dependencies on others

try-remove.py

import ast
import json
import sys
import sysconfig

from pathlib import Path


def remove(data, key):
    removed = {key}
    removed.update(k for k in data if k.startswith(key + "."))

    report = {}

    any_removed = bool(removed)
    while any_removed:
        for k in removed:
            data.pop(k, None)

        any_removed = False
        for k, v in data.items():
            cause = removed & set(v)
            if not cause:
                continue

            any_removed = True
            report[k] = cause
            removed.add(k)
            removed.update(k2 for k2 in data if k2.startswith(f"{k}."))

    return report


def find_deps(file):
    try:
        with open(file, "rb") as f:
            tree = ast.parse(f.read(), str(file))
    except SyntaxError:
        return

    # for n in ast.walk(tree):
    for n in ast.iter_child_nodes(tree):
        if isinstance(n, ast.ImportFrom):
            yield n.module
        elif isinstance(n, ast.Import):
            yield from (nom.name for nom in n.names)


def module_name(root, file):
    p = file.with_suffix("").relative_to(root)
    parts = p.parts[:-1] if p.parts[-1] == "__init__" else p.parts
    return ".".join(parts)


SKIP_DIRS = {"test", "tests", "site-packages", "idle_test"}


def search(root):
    found = {}
    for p in Path(root).resolve().rglob("**/*.py"):
        if SKIP_DIRS & set(p.parts):
            continue
        deps = set(find_deps(p))
        deps.discard(None)
        found[module_name(root, p)] = sorted(deps)
    return found


if __name__ == "__main__":
    if len(sys.argv) < 2:
        print("Usage: try-remove MODULE ...")
        print()
        sys.exit(1)

    try:
        with open("try-remove.json", "r", encoding="utf-8") as f:
            data = json.load(f)
    except FileNotFoundError:
        print("Scanning sysconfig.get_path('stdlib')")
        print(" -", sysconfig.get_path("stdlib"))
        print(" - caching to to", Path("try-remove.json").resolve())
        data = search(sysconfig.get_path("stdlib"))
        print()
        with open("try-remove.json", "w", encoding="utf-8") as f:
            json.dump(data, f)

    report = {}
    for m in sys.argv[1:]:
        report.update(remove(data, m))

    print("Removed:")
    for k in sorted(report):
        v = report[k]
        print(f" {k} because {sorted(v)}")