scripts/update_copyright.py

#!/usr/bin/env python

import glob
import os
from pathlib import Path

SCRIPT_DIR = Path(os.path.realpath(__file__)).parent


class Copyright:
    # Canonical license body, shared with every generator — see scripts/license_header.txt.
    NOTICE = (SCRIPT_DIR / "license_header.txt").read_text(encoding="utf-8").rstrip("\n")

    def __init__(self, comment_characters="//", prefix=None):
        self._comment_characters = comment_characters
        self._prefix = prefix or []

    def update(self, files):
        for file in files:
            with open(file, encoding="utf-8-sig") as f:
                lines = f.readlines()

            index = -1
            for i, line in enumerate(lines):
                if line.startswith(self._comment_characters) or self.valid_copyright_notice_line(line, index, file):
                    index += 1
                else:
                    break

            if index == -1:
                self.write_update_notice(file, lines)
            else:
                current = "".join(lines[: index + 1])
                if current != self.copyright_notice(file):
                    self.write_update_notice(file, lines[index + 1 :])

    def valid_copyright_notice_line(self, line, index, file):
        return index + 1 < len(self.copyright_notice_lines(file)) and line.startswith(
            self.copyright_notice_lines(file)[index + 1]
        )

    def copyright_notice(self, file):
        return "".join(self.copyright_notice_lines(file))

    def copyright_notice_lines(self, file):
        return self.dotnet(file) if file.endswith("cs") else self._prefix + self.commented_notice_lines

    def dotnet(self, file):
        file_name = os.path.basename(file)
        first = f'{self._comment_characters} <copyright file="{file_name}" company="Selenium Committers">\n'
        last = f"{self._comment_characters} </copyright>"
        return [first] + self.commented_notice_lines + [last]

    @property
    def commented_notice_lines(self):
        return [f"{self._comment_characters} {line}".rstrip() + "\n" for line in self.NOTICE.split("\n")]

    def write_update_notice(self, file, lines):
        # Build new content
        new_content = self.copyright_notice(file) + "\n"
        if lines and lines[0] != "\n":
            new_content += "\n"
        new_content += "".join(line.rstrip() + "\n" for line in lines)

        # Only write if different
        with open(file, encoding="utf-8-sig") as f:
            old_content = f.read()
        if new_content == old_content:
            return

        print(f"Adding notice to {file}")
        with open(file, "w") as f:
            f.write(new_content)


ROOT = Path(os.path.realpath(__file__)).parent.parent

JS_EXCLUSIONS = [
    f"{ROOT}/javascript/**/node_modules/**",
    f"{ROOT}/javascript/atoms/test/jquery.min.js",
]

PY_EXCLUSIONS = [
    f"{ROOT}/py/generate.py",
    f"{ROOT}/py/selenium/webdriver/common/devtools/**/*",
    f"{ROOT}/py/venv/**/*",
]


def update_files(file_pattern, exclusions, comment_characters="//", prefix=None):
    included = set(glob.glob(file_pattern, recursive=True))
    excluded = set()
    for pattern in exclusions:
        excluded.update(glob.glob(pattern, recursive=True))
    files = included - excluded

    copyright = Copyright(comment_characters, prefix)
    copyright.update(files)


if __name__ == "__main__":
    update_files(f"{ROOT}/javascript/**/*.js", JS_EXCLUSIONS)
    update_files(f"{ROOT}/javascript/**/*.mjs", JS_EXCLUSIONS)
    update_files(f"{ROOT}/javascript/**/*.cjs", JS_EXCLUSIONS)
    update_files(f"{ROOT}/javascript/**/*.tsx", JS_EXCLUSIONS)
    update_files(f"{ROOT}/javascript/**/*.ts", JS_EXCLUSIONS + [f"{ROOT}/javascript/**/*.d.ts"])
    update_files(f"{ROOT}/py/**/*.py", PY_EXCLUSIONS, comment_characters="#")
    update_files(f"{ROOT}/py/**/*.pyi", PY_EXCLUSIONS, comment_characters="#")
    update_files(
        f"{ROOT}/rb/**/*.rb",
        [],
        comment_characters="#",
        prefix=["# frozen_string_literal: true\n", "\n"],
    )
    update_files(
        f"{ROOT}/rb/**/*.rb.erb",
        [],
        comment_characters="#",
        prefix=["# frozen_string_literal: true\n", "\n"],
    )
    update_files(
        f"{ROOT}/rb/**/*.rbs",
        [f"{ROOT}/rb/sig/gems/**/*.rbs"],
        comment_characters="#",
    )
    # *.rbs.erb matches neither the *.rb.erb nor the *.rbs pattern, so it needs its own pass.
    update_files(
        f"{ROOT}/rb/**/*.rbs.erb",
        [],
        comment_characters="#",
    )
    update_files(f"{ROOT}/java/**/*.java", [])
    update_files(f"{ROOT}/rust/**/*.rs", [])
    update_files(f"{ROOT}/dotnet/**/*.cs", [])