jeirmeister/repo-to-text
1
0
Fork 0
mirror of https://github.com/kirill-markin/repo-to-text.git synced 2025-12-06 03:22:23 -08:00
Code Issues Projects Releases Packages Wiki Activity Actions

empty dirs remove

Browse source
This commit is contained in:
Kirill Markin 2024-06-08 11:08:08 +02:00
parent d720a094dd
commit 2dd3413f0d
1 changed files with 23 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

35
repo_to_text/main.py
View file

@ -4,12 +4,13 @@ import pathspec
import logging import logging
# Configure logging # Configure logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s') logging.basicConfig(level=logging.DEBUG, format='%(asctime)s - %(levelname)s - %(message)s')
def get_tree_structure(path='.', gitignore_spec=None) -> str: def get_tree_structure(path='.', gitignore_spec=None) -> str:
logging.debug(f'Generating tree structure for path: {path}') logging.debug(f'Generating tree structure for path: {path}')
result = subprocess.run(['tree', '-a', '-f', '--noreport', path], stdout=subprocess.PIPE) result = subprocess.run(['tree', '-a', '-f', '--noreport', path], stdout=subprocess.PIPE)
tree_output = result.stdout.decode('utf-8') tree_output = result.stdout.decode('utf-8')
logging.debug(f'Tree output generated: {tree_output}')
if not gitignore_spec: if not gitignore_spec:
logging.debug('No .gitignore specification found') logging.debug('No .gitignore specification found')
@ -44,30 +45,40 @@ def is_ignored_path(file_path: str) -> bool:
logging.debug(f'Path ignored: {file_path}') logging.debug(f'Path ignored: {file_path}')
return result return result
def remove_empty_dirs(tree_output: str) -> str: def remove_empty_dirs(tree_output: str, path='.') -> str:
logging.debug('Removing empty directories from tree output') logging.debug('Removing empty directories from tree output')
lines = tree_output.splitlines() lines = tree_output.splitlines()
non_empty_dirs = set() non_empty_dirs = set()
filtered_lines = [] filtered_lines = []
for line in reversed(lines): for line in lines:
logging.debug(f'Processing line: {line}') parts = line.strip().split()
if line.strip().endswith('/'): if parts:
logging.debug('Line is a directory') full_path = parts[-1]
if any(line.strip() in dir_line for dir_line in non_empty_dirs): if os.path.isdir(full_path) and not any(os.path.isfile(os.path.join(full_path, f)) for f in os.listdir(full_path)):
filtered_lines.append(line) logging.debug(f'Directory is empty and will be removed: {full_path}')
else: continue
non_empty_dirs.add(line) non_empty_dirs.add(os.path.dirname(full_path))
filtered_lines.append(line) filtered_lines.append(line)
final_lines = []
for line in filtered_lines:
parts = line.strip().split()
if parts:
full_path = parts[-1]
if os.path.isdir(full_path) and full_path not in non_empty_dirs:
logging.debug(f'Directory is empty and will be removed: {full_path}')
continue
final_lines.append(line)
logging.debug('Empty directory removal complete') logging.debug('Empty directory removal complete')
return '\n'.join(reversed(filtered_lines)) return '\n'.join(final_lines)
def save_repo_to_text(path='.') -> None: def save_repo_to_text(path='.') -> None:
logging.debug(f'Starting to save repo structure to text for path: {path}') logging.debug(f'Starting to save repo structure to text for path: {path}')
gitignore_spec = load_gitignore(path) gitignore_spec = load_gitignore(path)
tree_structure = get_tree_structure(path, gitignore_spec) tree_structure = get_tree_structure(path, gitignore_spec)
tree_structure = remove_empty_dirs(tree_structure) tree_structure = remove_empty_dirs(tree_structure, path)
output_file = 'repo_structure.txt' output_file = 'repo_structure.txt'
with open(output_file, 'w') as file: with open(output_file, 'w') as file: