kevinveenbirkenbach/directory-content-scanner
1
0
Fork 0
mirror of https://github.com/kevinveenbirkenbach/directory-content-scanner.git synced 2024-11-22 13:01:02 +01:00
Code Issues Packages Projects Releases Wiki Activity

Compare commits

base: kevinveenbirkenbach:0a315ad6ea09f23fe034e6d89795eaf4fb487f69
Branches Tags
kevinveenbirkenbach:main
..
compare: kevinveenbirkenbach:e78834da0859adea65653ca531ca4b7aa8a1a301
Branches Tags
kevinveenbirkenbach:main

No commits in common. "0a315ad6ea09f23fe034e6d89795eaf4fb487f69" and "e78834da0859adea65653ca531ca4b7aa8a1a301" have entirely different histories.
0a315ad6ea ... e78834da08

1 changed files with 25 additions and 56 deletions
Show Stats Expand all files Collapse all files

81
scan.py
View File

@ -51,61 +51,31 @@ class CodeProcessor:
class DirectoryHandler: class DirectoryHandler:
@staticmethod @staticmethod
def filter_directories(dirs, ignore_file_strings, ignore_hidden): def filter_directories(dirs, ignore_strings, ignore_hidden):
"""Filter out directories based on ignore criteria.""" """Filter out directories based on ignore criteria."""
if ignore_hidden: if ignore_hidden:
dirs[:] = [d for d in dirs if not d.startswith('.')] dirs[:] = [d for d in dirs if not d.startswith('.')]
dirs[:] = [d for d in dirs if not any(ig in d for ig in ignore_file_strings)] dirs[:] = [d for d in dirs if not any(ig in d for ig in ignore_strings)]
@staticmethod def should_print_file(file_path, file_filters, ignore_strings, ignore_hidden, whitelist_file, whitelist_content):
def path_or_content_contains(file_path, path_contains, content_contains): """Determine if a file should be printed based on new criteria."""
# Check if the file name contains specific strings (whitelist) if ignore_hidden and os.path.basename(file_path).startswith('.'):
if path_contains and any(whitelist_str in file_path for whitelist_str in path_contains): return False
return True if file_filters and not any(file_path.endswith(file_type) for file_type in file_filters):
return False
# Check file content for specific strings (if specified) if any(ignore_str in file_path for ignore_str in ignore_strings):
if content_contains: return False
if whitelist_file and not any(whitelist_str in file_path for whitelist_str in whitelist_file):
return False
if whitelist_content:
try: try:
with open(file_path, 'r') as f: with open(file_path, 'r') as f:
content = f.read() content = f.read()
# Return True if any of the content_contains strings are found in the content if not any(whitelist_str in content for whitelist_str in whitelist_content):
if any(whitelist_str in content for whitelist_str in content_contains): return False
return True
except UnicodeDecodeError: except UnicodeDecodeError:
# Return False if there's a Unicode decode error (file can't be read)
return False return False
return False return True
@staticmethod
def should_print_file(file_path, file_types, ignore_file_strings, ignore_hidden, path_contains, content_contains):
"""
Determine if a file should be printed based on various criteria.
Args:
file_path (str): The path of the file to be checked.
file_types (list): List of allowed file extensions.
ignore_file_strings (list): List of strings; if any are found in the file path, the file is ignored.
ignore_hidden (bool): If True, hidden files (starting with '.') are ignored.
path_contains (list): List of strings; the file is processed only if its path contains one of these strings.
content_contains (list): List of strings; the file is processed only if its content contains one of these strings.
Returns:
bool: True if the file should be printed, False otherwise.
"""
# Check for hidden files if ignore_hidden is enabled
if ignore_hidden and os.path.basename(file_path).startswith('.'):
return False
# Check if the file type is in the allowed list (if specified)
if file_types and not any(file_path.endswith(file_type) for file_type in file_types):
return False
# Check if the file should be ignored based on the presence of specific strings in its path
if any(ignore_str in file_path for ignore_str in ignore_file_strings):
return False
return DirectoryHandler.path_or_content_contains(file_path, path_contains, content_contains)
@staticmethod @staticmethod
def print_file_content(file_path, no_comments, compress): def print_file_content(file_path, no_comments, compress):
@ -132,9 +102,9 @@ class DirectoryHandler:
def handle_directory(directory, **kwargs): def handle_directory(directory, **kwargs):
"""Handle scanning and printing for directories.""" """Handle scanning and printing for directories."""
for root, dirs, files in os.walk(directory): for root, dirs, files in os.walk(directory):
DirectoryHandler.filter_directories(dirs, kwargs['ignore_file_strings'], kwargs['ignore_hidden']) DirectoryHandler.filter_directories(dirs, kwargs['ignore_strings'], kwargs['ignore_hidden'])
for file in files: for file in files:
if DirectoryHandler.should_print_file(os.path.join(root, file), kwargs['file_types'], kwargs['ignore_file_strings'], kwargs['ignore_hidden'], kwargs['path_contains'], kwargs['content_contains']): if DirectoryHandler.should_print_file(os.path.join(root, file), kwargs['file_filters'], kwargs['ignore_strings'], kwargs['ignore_hidden'], kwargs['whitelist_file'], kwargs['whitelist_content']):
DirectoryHandler.print_file_content(os.path.join(root, file), kwargs['no_comments'], kwargs['compress']) DirectoryHandler.print_file_content(os.path.join(root, file), kwargs['no_comments'], kwargs['compress'])
elif kwargs['verbose']: elif kwargs['verbose']:
print(f"Skipped file: {file}") print(f"Skipped file: {file}")
@ -148,23 +118,22 @@ class DirectoryHandler:
def main(): def main():
parser = argparse.ArgumentParser(description="Scan directories and print/compile file contents.") parser = argparse.ArgumentParser(description="Scan directories and print/compile file contents.")
parser.add_argument("paths", nargs='+', help="List of files or directories to scan.") parser.add_argument("paths", nargs='+', help="List of files or directories to scan.")
parser.add_argument("--file-types", nargs='+', default=[], help="Filter by file types (e.g., .txt .log).") parser.add_argument("-f", "--filetype", nargs='+', default=[], help="Filter by file types (e.g., .txt .log).")
parser.add_argument("--ignore-file-strings", nargs='+', default=[], help="Ignore files and folders containing these strings.") parser.add_argument("-i", "--ignore", nargs='+', default=[], help="Ignore files and folders containing these strings.")
parser.add_argument("--ignore-hidden", action='store_true', help="Ignore hidden directories and files.") parser.add_argument("--ignore-hidden", action='store_true', help="Ignore hidden directories and files.")
parser.add_argument("-v", "--verbose", action='store_true', help="Enable verbose mode.") parser.add_argument("-v", "--verbose", action='store_true', help="Enable verbose mode.")
parser.add_argument("--no-comments", action='store_true', help="Remove comments from the displayed content based on file type.") parser.add_argument("--no-comments", action='store_true', help="Remove comments from the displayed content based on file type.")
parser.add_argument("--compress", action='store_true', help="Compress code (for Python files).") parser.add_argument("--compress", action='store_true', help="Compress code (for Python files).")
parser.add_argument("--path-contains", nargs='+', default=[], help="Display files whose paths contain one of these strings.") parser.add_argument("--whitelist-file", nargs='+', default=[], help="Only display files whose paths contain one of these strings.")
parser.add_argument("--content-contains", nargs='+', default=[], help="Display files containing one of these strings in their content.") parser.add_argument("--whitelist-content", nargs='+', default=[], help="Only display files containing one of these strings in their content.")
args = parser.parse_args() args = parser.parse_args()
for path in args.paths: for path in args.paths:
if os.path.isdir(path): if os.path.isdir(path):
DirectoryHandler.handle_directory(path, file_types=args.file_types, ignore_file_strings=args.ignore_file_strings, ignore_hidden=args.ignore_hidden, verbose=args.verbose, no_comments=args.no_comments, compress=args.compress, path_contains=args.path_contains, content_contains=args.content_contains) DirectoryHandler.handle_directory(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, verbose=args.verbose, no_comments=args.no_comments, compress=args.compress, whitelist_file=args.whitelist_file, whitelist_content=args.whitelist_content)
elif os.path.isfile(path): elif os.path.isfile(path):
if DirectoryHandler.should_print_file(path, file_types=args.file_types, ignore_file_strings=args.ignore_file_strings, ignore_hidden=args.ignore_hidden, path_contains=args.path_contains, content_contains=args.content_contains): if DirectoryHandler.should_print_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, whitelist_file=args.whitelist_file, whitelist_content=args.whitelist_content):
DirectoryHandler.handle_file(path, file_types=args.file_types, ignore_file_strings=args.ignore_file_strings, ignore_hidden=args.ignore_hidden, no_comments=args.no_comments, compress=args.compress) DirectoryHandler.handle_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, no_comments=args.no_comments, compress=args.compress)
else: else:
print(f"Error: {path} is neither a valid file nor a directory.") print(f"Error: {path} is neither a valid file nor a directory.")
exit(1) exit(1)