Merge remote-tracking branch 'origin/master' into develop

2 files changed, 117 insertions, 0 deletions

diff --git a/lib/python/qmk/cli/__init__.py b/lib/python/qmk/cli/__init__.py
index b8bc99aa0d..695a180066 100644
--- a/lib/python/qmk/cli/__init__.py
+++ b/lib/python/qmk/cli/__init__.py
@@ -69,6 +69,7 @@ subcommands = [
     'qmk.cli.import.keymap',
     'qmk.cli.info',
     'qmk.cli.json2c',
+    'qmk.cli.license_check',
     'qmk.cli.lint',
     'qmk.cli.kle2json',
     'qmk.cli.list.keyboards',
diff --git a/lib/python/qmk/cli/license_check.py b/lib/python/qmk/cli/license_check.py
new file mode 100644
index 0000000000..4bda272ec9
--- /dev/null
+++ b/lib/python/qmk/cli/license_check.py
@@ -0,0 +1,116 @@
+# Copyright 2023 Nick Brassel (@tzarc)
+# SPDX-License-Identifier: GPL-2.0-or-later
+import re
+from pathlib import Path
+from milc import cli
+from qmk.constants import LICENSE_TEXTS
+
+L_PAREN = re.compile(r'\(\[\{\<')
+R_PAREN = re.compile(r'\)\]\}\>')
+PUNCTUATION = re.compile(r'[\.,;:]+')
+TRASH_PREFIX = re.compile(r'^(\s|/|\*|#)+')
+TRASH_SUFFIX = re.compile(r'(\s|/|\*|#|\\)+$')
+SPACE = re.compile(r'\s+')
+SUFFIXES = ['.c', '.h', '.cpp', '.cxx', '.hpp', '.hxx']
+
+
+def _simplify_text(input):
+    lines = input.lower().split('\n')
+    lines = [PUNCTUATION.sub('', line) for line in lines]
+    lines = [TRASH_PREFIX.sub('', line) for line in lines]
+    lines = [TRASH_SUFFIX.sub('', line) for line in lines]
+    lines = [SPACE.sub(' ', line) for line in lines]
+    lines = [L_PAREN.sub('(', line) for line in lines]
+    lines = [R_PAREN.sub(')', line) for line in lines]
+    lines = [line.strip() for line in lines]
+    lines = [line for line in lines if line is not None and line != '']
+    return ' '.join(lines)
+
+
+def _detect_license_from_file_contents(filename, absolute=False):
+    data = filename.read_text(encoding='utf-8', errors='ignore')
+    filename_out = str(filename.absolute()) if absolute else str(filename)
+
+    if 'SPDX-License-Identifier:' in data:
+        res = data.split('SPDX-License-Identifier:')
+        license = re.split(r'\s|//|\*', res[1].strip())[0].strip()
+        found = False
+        for short_license, _ in LICENSE_TEXTS:
+            if license.lower() == short_license.lower():
+                license = short_license
+                found = True
+                break
+
+        if not found:
+            if cli.args.short:
+                print(f'{filename_out} UNKNOWN')
+            else:
+                cli.log.error(f'{{fg_cyan}}{filename_out}{{fg_reset}} -- unknown license, or no license detected!')
+            return False
+
+        if cli.args.short:
+            print(f'{filename_out} {license}')
+        else:
+            cli.log.info(f'{{fg_cyan}}{filename_out}{{fg_reset}} -- license detected: {license} (SPDX License Identifier)')
+        return True
+
+    else:
+        simple_text = _simplify_text(data)
+        for short_license, long_licenses in LICENSE_TEXTS:
+            for long_license in long_licenses:
+                if long_license in simple_text:
+                    if cli.args.short:
+                        print(f'{filename_out} {short_license}')
+                    else:
+                        cli.log.info(f'{{fg_cyan}}{filename_out}{{fg_reset}} -- license detected: {short_license} (Full text)')
+                    return True
+
+        if cli.args.short:
+            print(f'{filename_out} UNKNOWN')
+        else:
+            cli.log.error(f'{{fg_cyan}}{filename_out}{{fg_reset}} -- unknown license, or no license detected!')
+
+    return False
+
+
+@cli.argument('inputs', nargs='*', arg_only=True, type=Path, help='List of input files or directories.')
+@cli.argument('-s', '--short', action='store_true', help='Short output.')
+@cli.argument('-a', '--absolute', action='store_true', help='Print absolute paths.')
+@cli.argument('-e', '--extension', arg_only=True, action='append', default=[], help='Override list of extensions. Can be specified multiple times for multiple extensions.')
+@cli.subcommand('File license check.', hidden=False if cli.config.user.developer else True)
+def license_check(cli):
+    def _default_suffix_condition(s):
+        return s in SUFFIXES
+
+    conditional = _default_suffix_condition
+
+    if len(cli.args.extension) > 0:
+        suffixes = [f'.{s}' if not s.startswith('.') else s for s in cli.args.extension]
+
+        def _specific_suffix_condition(s):
+            return s in suffixes
+
+        conditional = _specific_suffix_condition
+
+    # Pre-format all the licenses
+    for _, long_licenses in LICENSE_TEXTS:
+        for i in range(len(long_licenses)):
+            long_licenses[i] = _simplify_text(long_licenses[i])
+
+    check_list = set()
+    for filename in sorted(cli.args.inputs):
+        if filename.is_dir():
+            for file in sorted(filename.rglob('*')):
+                if file.is_file() and conditional(file.suffix):
+                    check_list.add(file)
+        elif filename.is_file():
+            if conditional(filename.suffix):
+                check_list.add(filename)
+
+    failed = False
+    for filename in sorted(check_list):
+        if not _detect_license_from_file_contents(filename, absolute=cli.args.absolute):
+            failed = True
+
+    if failed:
+        return False