Add hook to check SPDX lines

.pre-commit-hooks.yaml

@@ -61,6 +61,13 @@
     types: [text]
     stages: [pre-commit, pre-push, manual]
     minimum_pre_commit_version: 3.2.0
+-   id: check-spdx
+    name: check SPDX line
+    description: checks for the presence of an SPDX-License-Identifier in the comments of source files.
+    entry: check-spdx
+    language: python
+    files: \.(py|js|java|c|cpp|go|sh|ps1)$
+    types: [text]
 -   id: pretty-format-json
     name: pretty format json
     description: sets a standard for formatting json files.

README.md

@@ -64,6 +64,9 @@ Check for files that contain merge conflict strings.
 #### `check-shebang-scripts-are-executable`
 Checks that scripts with shebangs are executable.
 
+#### `check-spdx`
+Checks for the presence of an SPDX-License-Identifier in the comments of source files.
+
 #### `check-symlinks`
 Checks for symlinks which do not point to anything.
 

pre_commit_hooks/check_spdx.py

@@ -0,0 +1,62 @@
+"""This script checks for the presence of an SPDX-License-Identifier in the comments of source files."""
+from __future__ import annotations
+
+import argparse
+import re
+from collections.abc import Sequence
+
+
+def _load_file(file_path: str) -> str:
+    try:
+        with open(file_path, encoding='utf-8') as f:
+            return f.read()
+    except Exception as e:
+        print(f"Error loading file content of {file_path}: {e}")
+
+
+def _check_spdx(file_content: str) -> bool:
+    for line in file_content:
+        stripped_line = line.strip()
+        if stripped_line.startswith('#') or stripped_line.startswith('//') or re.match(r'^\s*/\*', stripped_line):
+            if 'SPDX-License-Identifier:' in stripped_line:
+                return True
+        else:
+            break
+
+
+def check_spdx(file_paths: list[str]) -> int:
+    """
+    Check if the given files contain an SPDX license identifier.
+
+    Args:
+        file_paths (list of str): List of file paths to check.
+
+    Returns:
+        int: Returns 0 if all files contain an SPDX license identifier,
+             otherwise returns 1 if any file is missing the SPDX line.
+    """
+    any_missing_spdx = False
+    for file_path in file_paths:
+        file_content = _load_file(file_path)
+        if not file_content:
+            return 1
+
+        if not _check_spdx(file_path):
+            print(f"Missing SPDX line in {file_path}")
+            any_missing_spdx = True
+
+    if any_missing_spdx:
+        return 1
+    return 0
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('filenames', nargs='*')
+    args = parser.parse_args(argv)
+
+    return check_spdx(args.filenames)
+
+
+if __name__ == '__main__':
+    raise SystemExit(main())

setup.cfg

@@ -40,6 +40,7 @@ console_scripts =
     check-json = pre_commit_hooks.check_json:main
     check-merge-conflict = pre_commit_hooks.check_merge_conflict:main
     check-shebang-scripts-are-executable = pre_commit_hooks.check_shebang_scripts_are_executable:main
+    check-spdx = pre_commit_hooks.check_spdx:main
     check-symlinks = pre_commit_hooks.check_symlinks:main
     check-toml = pre_commit_hooks.check_toml:main
     check-vcs-permalinks = pre_commit_hooks.check_vcs_permalinks:main