Merge pull request #23 from kirill-markin/refactoring/cleanup

Refactoring/cleanup

.github/workflows/tests.yml

@@ -14,7 +14,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"]
+        python-version: ["3.8", "3.11", "3.13"]
 
     steps:
     - uses: actions/checkout@v4
@@ -28,6 +28,9 @@ jobs:
       run: |
         python -m pip install --upgrade pip
         pip install .[dev]
+    - name: Run pylint
+      run: |
+        pylint repo_to_text
     - name: Run tests
       run: |
         pytest tests/ 

README.md

@@ -180,7 +180,7 @@ To install `repo-to-text` locally for development, follow these steps:
 2. Install the package with development dependencies:
 
     ```bash
-    pip install -e .[dev]
+    pip install -e ".[dev]"
     ```
 
 ### Requirements

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "repo-to-text"
-version = "0.5.3"
+version = "0.5.4"
 authors = [
     { name = "Kirill Markin", email = "markinkirill@gmail.com" },
 ]
@@ -41,4 +41,10 @@ dev = [
     "isort",
     "build",
     "twine",
+    "pylint",
+]
+
+[tool.pylint]
+disable = [
+    "C0303",
 ]

repo_to_text/__init__.py

@@ -1,2 +1,4 @@
+"""This is the main package for the repo_to_text package."""
+
 __author__ = 'Kirill Markin'
 __email__ = 'markinkirill@gmail.com'

repo_to_text/cli/__init__.py

@@ -1,3 +1,5 @@
+"""This module contains the CLI interface for the repo_to_text package."""
+
 from .cli import create_default_settings_file, parse_args, main
 
 __all__ = ['create_default_settings_file', 'parse_args', 'main']

repo_to_text/cli/cli.py

@@ -1,3 +1,7 @@
+"""
+CLI for repo-to-text
+"""
+
 import argparse
 import textwrap
 import os
@@ -12,7 +16,10 @@ def create_default_settings_file() -> None:
     """Create a default .repo-to-text-settings.yaml file."""
     settings_file = '.repo-to-text-settings.yaml'
     if os.path.exists(settings_file):
-        raise FileExistsError(f"The settings file '{settings_file}' already exists. Please remove it or rename it if you want to create a new default settings file.")
+        raise FileExistsError(
+            f"The settings file '{settings_file}' already exists. "
+            "Please remove it or rename it if you want to create a new default settings file."
+        )
 
     default_settings = textwrap.dedent("""\
         # Details: https://github.com/kirill-markin/repo-to-text
@@ -32,7 +39,7 @@ def create_default_settings_file() -> None:
           - "README.md"
           - "LICENSE"
     """)
-    with open('.repo-to-text-settings.yaml', 'w') as f:
+    with open('.repo-to-text-settings.yaml', 'w', encoding='utf-8') as f:
         f.write(default_settings)
     print("Default .repo-to-text-settings.yaml created.")
 
@@ -42,13 +49,25 @@ def parse_args() -> argparse.Namespace:
     Returns:
         argparse.Namespace: Parsed command line arguments
     """
-    parser = argparse.ArgumentParser(description='Convert repository structure and contents to text')
+    parser = argparse.ArgumentParser(
+        description='Convert repository structure and contents to text'
+    )
     parser.add_argument('input_dir', nargs='?', default='.', help='Directory to process')
     parser.add_argument('--debug', action='store_true', help='Enable debug logging')
     parser.add_argument('--output-dir', type=str, help='Directory to save the output file')
-    parser.add_argument('--create-settings', '--init', action='store_true', help='Create default .repo-to-text-settings.yaml file')
+    parser.add_argument(
+        '--create-settings',
+        '--init',
+        action='store_true',
+        help='Create default .repo-to-text-settings.yaml file'
+    )
     parser.add_argument('--stdout', action='store_true', help='Output to stdout instead of a file')
-    parser.add_argument('--ignore-patterns', nargs='*', help="List of files or directories to ignore in both tree and content sections. Supports wildcards (e.g., '*').")
+    parser.add_argument(
+        '--ignore-patterns',
+        nargs='*',
+        help="List of files or directories to ignore in both tree and content sections. "
+        "Supports wildcards (e.g., '*')."
+    )
     return parser.parse_args()
 
 def main() -> NoReturn:
@@ -75,6 +94,6 @@ def main() -> NoReturn:
 
         logging.debug('repo-to-text script finished')
         sys.exit(0)
-    except Exception as e:
+    except (FileNotFoundError, FileExistsError, PermissionError, OSError) as e:
-        logging.error(f'Error occurred: {str(e)}')
+        logging.error('Error occurred: %s', str(e))
         sys.exit(1)

repo_to_text/core/__init__.py

@@ -1,3 +1,5 @@
+"""This module contains the core functionality of the repo_to_text package."""
+
 from .core import get_tree_structure, load_ignore_specs, should_ignore_file, save_repo_to_text
 
 __all__ = ['get_tree_structure', 'load_ignore_specs', 'should_ignore_file', 'save_repo_to_text'] 

repo_to_text/core/core.py

@@ -1,69 +1,116 @@
+"""
+Core functionality for repo-to-text
+"""
+
 import os
 import subprocess
+from typing import Tuple, Optional, List, Dict, Any, Set
+from datetime import datetime, timezone
+from importlib.machinery import ModuleSpec
 import logging
 import yaml
-from datetime import datetime, timezone
-from typing import Tuple, Optional, List
 import pathspec
 from pathspec import PathSpec
 
-from ..utils.utils import check_tree_command, is_ignored_path, remove_empty_dirs
+from ..utils.utils import check_tree_command, is_ignored_path
 
-def get_tree_structure(path: str = '.', gitignore_spec: Optional[PathSpec] = None, tree_and_content_ignore_spec: Optional[PathSpec] = None) -> str:
+def get_tree_structure(
-    """Generate tree structure of the directory.
+        path: str = '.',
-    
+        gitignore_spec: Optional[PathSpec] = None,
-    Args:
+        tree_and_content_ignore_spec: Optional[PathSpec] = None
-        path: Directory path to generate tree for
+    ) -> str:
-        gitignore_spec: PathSpec object for gitignore patterns
+    """Generate tree structure of the directory."""
-        tree_and_content_ignore_spec: PathSpec object for tree and content ignore patterns
-        
-    Returns:
-        str: Generated tree structure
-    """
     if not check_tree_command():
         return ""
 
-    logging.debug(f'Generating tree structure for path: {path}')
+    logging.debug('Generating tree structure for path: %s', path)
-    result = subprocess.run(['tree', '-a', '-f', '--noreport', path], stdout=subprocess.PIPE)
+    tree_output = run_tree_command(path)
-    tree_output = result.stdout.decode('utf-8')
+    logging.debug('Tree output generated:\n%s', tree_output)
-    logging.debug(f'Tree output generated:\n{tree_output}')
 
     if not gitignore_spec and not tree_and_content_ignore_spec:
         logging.debug('No .gitignore or ignore-tree-and-content specification found')
         return tree_output
 
-    logging.debug('Filtering tree output based on .gitignore and ignore-tree-and-content specification')
+    logging.debug('Filtering tree output based on ignore specifications')
-    filtered_lines: List[str] = []
+    return filter_tree_output(tree_output, path, gitignore_spec, tree_and_content_ignore_spec)
 
-    for line in tree_output.splitlines():
+def run_tree_command(path: str) -> str:
-        idx = line.find('./')
+    """Run the tree command and return its output."""
-        if idx == -1:
+    result = subprocess.run(
-            idx = line.find(path)
+        ['tree', '-a', '-f', '--noreport', path],
-        if idx != -1:
+        stdout=subprocess.PIPE,
-            full_path = line[idx:].strip()
+        check=True
-        else:
+    )
-            continue
+    return result.stdout.decode('utf-8')
 
-        if full_path == '.':
+def filter_tree_output(
-            continue
+        tree_output: str,
+        path: str,
+        gitignore_spec: Optional[PathSpec],
+        tree_and_content_ignore_spec: Optional[PathSpec]
+    ) -> str:
+    """Filter the tree output based on ignore specifications."""
+    lines: List[str] = tree_output.splitlines()
+    non_empty_dirs: Set[str] = set()
 
-        relative_path = os.path.relpath(full_path, path)
+    filtered_lines = [
-        relative_path = relative_path.replace(os.sep, '/')
+        process_line(line, path, gitignore_spec, tree_and_content_ignore_spec, non_empty_dirs)
-        if os.path.isdir(full_path):
+        for line in lines
-            relative_path += '/'
+    ]
 
-        if not should_ignore_file(full_path, relative_path, gitignore_spec, None, tree_and_content_ignore_spec):
+    filtered_tree_output = '\n'.join(filter(None, filtered_lines))
-            display_line = line.replace('./', '', 1)
+    logging.debug('Filtered tree structure:\n%s', filtered_tree_output)
-            filtered_lines.append(display_line)
-        else:
-            logging.debug(f'Ignored: {relative_path}')
-
-    filtered_tree_output = '\n'.join(filtered_lines)
-    logging.debug(f'Filtered tree structure:\n{filtered_tree_output}')
-    logging.debug('Tree structure filtering complete')
     return filtered_tree_output
 
-def load_ignore_specs(path: str = '.', cli_ignore_patterns: Optional[List[str]] = None) -> Tuple[Optional[PathSpec], Optional[PathSpec], PathSpec]:
+def process_line(
+        line: str,
+        path: str,
+        gitignore_spec: Optional[PathSpec],
+        tree_and_content_ignore_spec: Optional[PathSpec],
+        non_empty_dirs: Set[str]
+    ) -> Optional[str]:
+    """Process a single line of the tree output."""
+    full_path = extract_full_path(line, path)
+    if not full_path or full_path == '.':
+        return None
+
+    relative_path = os.path.relpath(full_path, path).replace(os.sep, '/')
+
+    if should_ignore_file(
+        full_path,
+        relative_path,
+        gitignore_spec,
+        None,
+        tree_and_content_ignore_spec
+    ):
+        logging.debug('Ignored: %s', relative_path)
+        return None
+
+    if not os.path.isdir(full_path):
+        mark_non_empty_dirs(relative_path, non_empty_dirs)
+
+    if not os.path.isdir(full_path) or os.path.dirname(relative_path) in non_empty_dirs:
+        return line.replace('./', '', 1)
+    return None
+
+def extract_full_path(line: str, path: str) -> Optional[str]:
+    """Extract the full path from a line of tree output."""
+    idx = line.find('./')
+    if idx == -1:
+        idx = line.find(path)
+    return line[idx:].strip() if idx != -1 else None
+
+def mark_non_empty_dirs(relative_path: str, non_empty_dirs: Set[str]) -> None:
+    """Mark all parent directories of a file as non-empty."""
+    dir_path = os.path.dirname(relative_path)
+    while dir_path:
+        non_empty_dirs.add(dir_path)
+        dir_path = os.path.dirname(dir_path)
+
+def load_ignore_specs(
+        path: str = '.',
+        cli_ignore_patterns: Optional[List[str]] = None
+    ) -> Tuple[Optional[PathSpec], Optional[PathSpec], PathSpec]:
     """Load ignore specifications from various sources.
     
     Args:
@@ -71,7 +118,8 @@ def load_ignore_specs(path: str = '.', cli_ignore_patterns: Optional[List[str]]
         cli_ignore_patterns: List of patterns from command line
         
     Returns:
-        Tuple[Optional[PathSpec], Optional[PathSpec], PathSpec]: Tuple of gitignore_spec, content_ignore_spec, and tree_and_content_ignore_spec
+        Tuple[Optional[PathSpec], Optional[PathSpec], PathSpec]: Tuple of gitignore_spec, 
+        content_ignore_spec, and tree_and_content_ignore_spec
     """
     gitignore_spec = None
     content_ignore_spec = None
@@ -80,14 +128,16 @@ def load_ignore_specs(path: str = '.', cli_ignore_patterns: Optional[List[str]]
 
     repo_settings_path = os.path.join(path, '.repo-to-text-settings.yaml')
     if os.path.exists(repo_settings_path):
-        logging.debug(f'Loading .repo-to-text-settings.yaml from path: {repo_settings_path}')
+        logging.debug('Loading .repo-to-text-settings.yaml from path: %s', repo_settings_path)
-        with open(repo_settings_path, 'r') as f:
+        with open(repo_settings_path, 'r', encoding='utf-8') as f:
-            settings = yaml.safe_load(f)
+            settings: Dict[str, Any] = yaml.safe_load(f)
             use_gitignore = settings.get('gitignore-import-and-ignore', True)
             if 'ignore-content' in settings:
-                content_ignore_spec = pathspec.PathSpec.from_lines('gitwildmatch', settings['ignore-content'])
+                content_ignore_spec: Optional[PathSpec] = pathspec.PathSpec.from_lines(
+                    'gitwildmatch', settings['ignore-content']
+                )
             if 'ignore-tree-and-content' in settings:
-                tree_and_content_ignore_list.extend(settings['ignore-tree-and-content'])
+                tree_and_content_ignore_list.extend(settings.get('ignore-tree-and-content', []))
 
     if cli_ignore_patterns:
         tree_and_content_ignore_list.extend(cli_ignore_patterns)
@@ -95,15 +145,22 @@ def load_ignore_specs(path: str = '.', cli_ignore_patterns: Optional[List[str]]
     if use_gitignore:
         gitignore_path = os.path.join(path, '.gitignore')
         if os.path.exists(gitignore_path):
-            logging.debug(f'Loading .gitignore from path: {gitignore_path}')
+            logging.debug('Loading .gitignore from path: %s', gitignore_path)
-            with open(gitignore_path, 'r') as f:
+            with open(gitignore_path, 'r', encoding='utf-8') as f:
                 gitignore_spec = pathspec.PathSpec.from_lines('gitwildmatch', f)
 
-    tree_and_content_ignore_spec = pathspec.PathSpec.from_lines('gitwildmatch', tree_and_content_ignore_list)
+    tree_and_content_ignore_spec = pathspec.PathSpec.from_lines(
+        'gitwildmatch', tree_and_content_ignore_list
+    )
     return gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec
 
-def should_ignore_file(file_path: str, relative_path: str, gitignore_spec: Optional[PathSpec], 
+def should_ignore_file(
-                      content_ignore_spec: Optional[PathSpec], tree_and_content_ignore_spec: Optional[PathSpec]) -> bool:
+    file_path: str,
+    relative_path: str,
+    gitignore_spec: Optional[PathSpec],
+    content_ignore_spec: Optional[PathSpec],
+    tree_and_content_ignore_spec: Optional[PathSpec]
+) -> bool:
     """Check if a file should be ignored based on various ignore specifications.
     
     Args:
@@ -126,44 +183,73 @@ def should_ignore_file(file_path: str, relative_path: str, gitignore_spec: Optio
 
     result = (
         is_ignored_path(file_path) or
-        bool(gitignore_spec and gitignore_spec.match_file(relative_path)) or
+        bool(
-        bool(content_ignore_spec and content_ignore_spec.match_file(relative_path)) or
+            gitignore_spec and
-        bool(tree_and_content_ignore_spec and tree_and_content_ignore_spec.match_file(relative_path)) or
+            gitignore_spec.match_file(relative_path)
+        ) or
+        bool(
+            content_ignore_spec and
+            content_ignore_spec.match_file(relative_path)
+        ) or
+        bool(
+            tree_and_content_ignore_spec and
+            tree_and_content_ignore_spec.match_file(relative_path)
+        ) or
         os.path.basename(file_path).startswith('repo-to-text_')
     )
 
-    logging.debug(f'Checking if file should be ignored:')
+    logging.debug('Checking if file should be ignored:')
-    logging.debug(f'    file_path: {file_path}')
+    logging.debug('    file_path: %s', file_path)
-    logging.debug(f'    relative_path: {relative_path}')
+    logging.debug('    relative_path: %s', relative_path)
-    logging.debug(f'    Result: {result}')
+    logging.debug('    Result: %s', result)
     return result
 
-def save_repo_to_text(path: str = '.', output_dir: Optional[str] = None, to_stdout: bool = False, cli_ignore_patterns: Optional[List[str]] = None) -> str:
+def save_repo_to_text(
-    """Save repository structure and contents to a text file.
+        path: str = '.',
+        output_dir: Optional[str] = None,
+        to_stdout: bool = False,
+        cli_ignore_patterns: Optional[List[str]] = None
+    ) -> str:
+    """Save repository structure and contents to a text file."""
+    logging.debug('Starting to save repo structure to text for path: %s', path)
+    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(
+        path, cli_ignore_patterns
+    )
+    tree_structure: str = get_tree_structure(
+        path, gitignore_spec, tree_and_content_ignore_spec
+    )
+    logging.debug('Final tree structure to be written: %s', tree_structure)
 
-    Args:
+    output_content = generate_output_content(
-        path: Repository path
+        path,
-        output_dir: Directory to save output file
+        tree_structure,
-        to_stdout: Whether to output to stdout instead of file
+        gitignore_spec,
-        cli_ignore_patterns: List of patterns from command line
+        content_ignore_spec,
+        tree_and_content_ignore_spec
+    )
 
-    Returns:
+    if to_stdout:
-        str: Path to the output file or the output text if to_stdout is True
+        print(output_content)
-    """
+        return output_content
-    logging.debug(f'Starting to save repo structure to text for path: {path}')
-    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(path, cli_ignore_patterns)
-    tree_structure = get_tree_structure(path, gitignore_spec, tree_and_content_ignore_spec)
-    tree_structure = remove_empty_dirs(tree_structure, path)
-    logging.debug(f'Final tree structure to be written: {tree_structure}')
 
-    timestamp = datetime.now(timezone.utc).strftime('%Y-%m-%d-%H-%M-%S-UTC')
+    output_file = write_output_to_file(output_content, output_dir)
-    output_file = f'repo-to-text_{timestamp}.txt'
+    copy_to_clipboard(output_content)
 
-    if output_dir:
+    print(
-        if not os.path.exists(output_dir):
+        "[SUCCESS] Repository structure and contents successfully saved to "
-            os.makedirs(output_dir)
+        f"file: \"./{output_file}\""
-        output_file = os.path.join(output_dir, output_file)
+    )
 
+    return output_file
+
+def generate_output_content(
+        path: str,
+        tree_structure: str,
+        gitignore_spec: Optional[PathSpec],
+        content_ignore_spec: Optional[PathSpec],
+        tree_and_content_ignore_spec: Optional[PathSpec]
+    ) -> str:
+    """Generate the output content for the repository."""
     output_content: List[str] = []
     project_name = os.path.basename(os.path.abspath(path))
     output_content.append(f'Directory: {project_name}\n\n')
@@ -181,7 +267,13 @@ def save_repo_to_text(path: str = '.', output_dir: Optional[str] = None, to_stdo
             file_path = os.path.join(root, filename)
             relative_path = os.path.relpath(file_path, path)
 
-            if should_ignore_file(file_path, relative_path, gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec):
+            if should_ignore_file(
+                file_path,
+                relative_path,
+                gitignore_spec,
+                content_ignore_spec,
+                tree_and_content_ignore_spec
+            ):
                 continue
 
             relative_path = relative_path.replace('./', '', 1)
@@ -192,35 +284,45 @@ def save_repo_to_text(path: str = '.', output_dir: Optional[str] = None, to_stdo
                 with open(file_path, 'r', encoding='utf-8') as f:
                     output_content.append(f.read())
             except UnicodeDecodeError:
-                logging.debug(f'Could not decode file contents: {file_path}')
+                logging.debug('Could not decode file contents: %s', file_path)
                 output_content.append('[Could not decode file contents]\n')
             output_content.append('\n```\n')
 
     output_content.append('\n')
     logging.debug('Repository contents written to output content')
 
-    output_text = ''.join(output_content)
+    return ''.join(output_content)
 
-    if to_stdout:
+def write_output_to_file(output_content: str, output_dir: Optional[str]) -> str:
-        print(output_text)
+    """Write the output content to a file."""
-        return output_text
+    timestamp = datetime.now(timezone.utc).strftime('%Y-%m-%d-%H-%M-%S-UTC')
+    output_file = f'repo-to-text_{timestamp}.txt'
 
-    with open(output_file, 'w') as file:
+    if output_dir:
-        file.write(output_text)
+        if not os.path.exists(output_dir):
+            os.makedirs(output_dir)
+        output_file = os.path.join(output_dir, output_file)
 
+    with open(output_file, 'w', encoding='utf-8') as file:
+        file.write(output_content)
+
+    return output_file
+
+def copy_to_clipboard(output_content: str) -> None:
+    """Copy the output content to the clipboard if possible."""
     try:
-        import importlib.util
+        import importlib.util  # pylint: disable=import-outside-toplevel
-        if importlib.util.find_spec("pyperclip"):
+        spec: Optional[ModuleSpec] = importlib.util.find_spec("pyperclip")
-            import pyperclip # type: ignore
+        if spec:
-            pyperclip.copy(output_text) # type: ignore
+            import pyperclip  # pylint: disable=import-outside-toplevel # type: ignore
+            pyperclip.copy(output_content)  # type: ignore
             logging.debug('Repository structure and contents copied to clipboard')
         else:
             print("Tip: Install 'pyperclip' package to enable automatic clipboard copying:")
             print("     pip install pyperclip")
-    except Exception as e:
+    except ImportError as e:
-        logging.warning('Could not copy to clipboard. You might be running this script over SSH or without clipboard support.')
+        logging.warning(
-        logging.debug(f'Clipboard copy error: {e}')
+            'Could not copy to clipboard. You might be running this '
-    
+            'script over SSH or without clipboard support.'
-    print(f"[SUCCESS] Repository structure and contents successfully saved to file: \"./{output_file}\"")
+        )
-    
+        logging.debug('Clipboard copy error: %s', e)
-    return output_file 

repo_to_text/main.py

@@ -1,3 +1,5 @@
+"""This is the main entry point for the repo_to_text package."""
+
 from repo_to_text.cli.cli import main
 
 if __name__ == '__main__':

repo_to_text/utils/__init__.py

@@ -1,3 +1,5 @@
-from .utils import setup_logging, check_tree_command, is_ignored_path, remove_empty_dirs
+"""This module contains utility functions for the repo_to_text package."""
 
-__all__ = ['setup_logging', 'check_tree_command', 'is_ignored_path', 'remove_empty_dirs'] 
+from .utils import setup_logging, check_tree_command, is_ignored_path
+
+__all__ = ['setup_logging', 'check_tree_command', 'is_ignored_path']

repo_to_text/utils/utils.py

@@ -1,7 +1,8 @@
-import os
+"""This module contains utility functions for the repo_to_text package."""
+
 import shutil
 import logging
-from typing import List, Set
+from typing import List
 
 def setup_logging(debug: bool = False) -> None:
     """Set up logging configuration.
@@ -19,7 +20,10 @@ def check_tree_command() -> bool:
         bool: True if tree command is available, False otherwise
     """
     if shutil.which('tree') is None:
-        print("The 'tree' command is not found. Please install it using one of the following commands:")
+        print(
+            "The 'tree' command is not found. "
+            + "Please install it using one of the following commands:"
+        )
         print("For Debian-based systems (e.g., Ubuntu): sudo apt-get install tree")
         print("For Red Hat-based systems (e.g., Fedora, CentOS): sudo yum install tree")
         return False
@@ -40,43 +44,5 @@ def is_ignored_path(file_path: str) -> bool:
     is_ignored_file = any(file_path.startswith(prefix) for prefix in ignored_files_prefix)
     result = is_ignored_dir or is_ignored_file
     if result:
-        logging.debug(f'Path ignored: {file_path}')
+        logging.debug('Path ignored: %s', file_path)
     return result
-
-def remove_empty_dirs(tree_output: str, path: str = '.') -> str:
-    """Remove empty directories from tree output.
-    
-    Args:
-        tree_output: Output from tree command
-        path: Base path for the tree
-        
-    Returns:
-        str: Tree output with empty directories removed
-    """
-    logging.debug('Removing empty directories from tree output')
-    lines = tree_output.splitlines()
-    non_empty_dirs: Set[str] = set()
-    filtered_lines: List[str] = []
-
-    for line in lines:
-        parts = line.strip().split()
-        if parts:
-            full_path = parts[-1]
-            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)):
-                logging.debug(f'Directory is empty and will be removed: {full_path}')
-                continue
-            non_empty_dirs.add(os.path.dirname(full_path))
-            filtered_lines.append(line)
-    
-    final_lines: List[str] = []
-    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')
-    return '\n'.join(filtered_lines) 

tests/test_cli.py

@@ -1,15 +1,19 @@
+"""Test the CLI module."""
+
 import os
-import pytest
 import tempfile
 import shutil
 from typing import Generator
 from unittest.mock import patch, MagicMock
+import pytest
 from repo_to_text.cli.cli import (
     create_default_settings_file,
     parse_args,
     main
 )
 
+# pylint: disable=redefined-outer-name
+
 @pytest.fixture
 def temp_dir() -> Generator[str, None, None]:
     """Create a temporary directory for testing."""
@@ -52,7 +56,7 @@ def test_create_default_settings_file(temp_dir: str) -> None:
     settings_file = '.repo-to-text-settings.yaml'
     assert os.path.exists(settings_file)
 
-    with open(settings_file, 'r') as f:
+    with open(settings_file, 'r', encoding='utf-8') as f:
         content = f.read()
         assert 'gitignore-import-and-ignore: True' in content
         assert 'ignore-tree-and-content:' in content
@@ -94,7 +98,10 @@ def test_main_create_settings(mock_create_settings: MagicMock) -> None:
 
 @patch('repo_to_text.cli.cli.setup_logging')
 @patch('repo_to_text.cli.cli.create_default_settings_file')
-def test_main_with_debug_logging(mock_create_settings: MagicMock, mock_setup_logging: MagicMock) -> None:
+def test_main_with_debug_logging(
+    mock_create_settings: MagicMock,
+    mock_setup_logging: MagicMock
+) -> None:
     """Test main function with debug logging enabled."""
     with patch('sys.argv', ['repo-to-text', '--debug', '--create-settings']):
         with pytest.raises(SystemExit) as exc_info:

tests/test_core.py

@@ -1,17 +1,21 @@
+"""Test the core module."""
+
 import os
 import tempfile
 import shutil
-import pytest
 from typing import Generator
+import pytest
+
 from repo_to_text.core.core import (
     get_tree_structure,
     load_ignore_specs,
     should_ignore_file,
     is_ignored_path,
-    remove_empty_dirs,
     save_repo_to_text
 )
 
+# pylint: disable=redefined-outer-name
+
 @pytest.fixture
 def temp_dir() -> Generator[str, None, None]:
     """Create a temporary directory for testing."""
@@ -20,11 +24,12 @@ def temp_dir() -> Generator[str, None, None]:
     shutil.rmtree(temp_path)
 
 @pytest.fixture
-def sample_repo(temp_dir: str) -> str:
+def sample_repo(tmp_path: str) -> str:
     """Create a sample repository structure for testing."""
+    tmp_path_str = str(tmp_path)
     # Create directories
-    os.makedirs(os.path.join(temp_dir, "src"))
+    os.makedirs(os.path.join(tmp_path_str, "src"))
-    os.makedirs(os.path.join(temp_dir, "tests"))
+    os.makedirs(os.path.join(tmp_path_str, "tests"))
 
     # Create sample files
     files = {
@@ -47,12 +52,12 @@ ignore-content:
     }
 
     for file_path, content in files.items():
-        full_path = os.path.join(temp_dir, file_path)
+        full_path = os.path.join(tmp_path_str, file_path)
         os.makedirs(os.path.dirname(full_path), exist_ok=True)
-        with open(full_path, "w") as f:
+        with open(full_path, "w", encoding='utf-8') as f:
             f.write(content)
 
-    return temp_dir
+    return tmp_path_str
 
 def test_is_ignored_path() -> None:
     """Test the is_ignored_path function."""
@@ -63,7 +68,9 @@ def test_is_ignored_path() -> None:
 
 def test_load_ignore_specs(sample_repo: str) -> None:
     """Test loading ignore specifications from files."""
-    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(sample_repo)
+    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(
+        sample_repo
+    )
 
     assert gitignore_spec is not None
     assert content_ignore_spec is not None
@@ -82,7 +89,9 @@ def test_load_ignore_specs(sample_repo: str) -> None:
 
 def test_should_ignore_file(sample_repo: str) -> None:
     """Test file ignoring logic."""
-    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(sample_repo)
+    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(
+        sample_repo
+    )
 
     # Test various file paths
     assert should_ignore_file(
@@ -113,38 +122,6 @@ def test_get_tree_structure(sample_repo: str) -> None:
     assert "test_main.py" in tree_output
     assert ".git" not in tree_output
 
-def test_remove_empty_dirs(temp_dir: str) -> None:
-    """Test removal of empty directories from tree output."""
-    # Create test directory structure
-    os.makedirs(os.path.join(temp_dir, "src"))
-    os.makedirs(os.path.join(temp_dir, "empty_dir"))
-    os.makedirs(os.path.join(temp_dir, "tests"))
-    
-    # Create some files
-    with open(os.path.join(temp_dir, "src/main.py"), "w") as f:
-        f.write("print('test')")
-    with open(os.path.join(temp_dir, "tests/test_main.py"), "w") as f:
-        f.write("def test(): pass")
-    
-    # Create a mock tree output that matches the actual tree command format
-    tree_output = (
-        f"{temp_dir}\n"
-        f"├── {os.path.join(temp_dir, 'src')}\n"
-        f"│   └── {os.path.join(temp_dir, 'src/main.py')}\n"
-        f"├── {os.path.join(temp_dir, 'empty_dir')}\n"
-        f"└── {os.path.join(temp_dir, 'tests')}\n"
-        f"    └── {os.path.join(temp_dir, 'tests/test_main.py')}\n"
-    )
-    
-    filtered_output = remove_empty_dirs(tree_output, temp_dir)
-    
-    # Check that empty_dir is removed but other directories remain
-    assert "empty_dir" not in filtered_output
-    assert os.path.join(temp_dir, "src") in filtered_output
-    assert os.path.join(temp_dir, "tests") in filtered_output
-    assert os.path.join(temp_dir, "src/main.py") in filtered_output
-    assert os.path.join(temp_dir, "tests/test_main.py") in filtered_output
-
 def test_save_repo_to_text(sample_repo: str) -> None:
     """Test the main save_repo_to_text function."""
     # Create output directory
@@ -153,7 +130,7 @@ def test_save_repo_to_text(sample_repo: str) -> None:
 
     # Create .git directory to ensure it's properly ignored
     os.makedirs(os.path.join(sample_repo, ".git"))
-    with open(os.path.join(sample_repo, ".git/config"), "w") as f:
+    with open(os.path.join(sample_repo, ".git/config"), "w", encoding='utf-8') as f:
         f.write("[core]\n\trepositoryformatversion = 0\n")
 
     # Test file output
@@ -162,7 +139,7 @@ def test_save_repo_to_text(sample_repo: str) -> None:
     assert os.path.dirname(output_file) == output_dir
 
     # Check file contents
-    with open(output_file, 'r') as f:
+    with open(output_file, 'r', encoding='utf-8') as f:
         content = f.read()
 
         # Basic content checks
@@ -204,7 +181,9 @@ def test_load_ignore_specs_with_cli_patterns(sample_repo: str) -> None:
 
 def test_load_ignore_specs_without_gitignore(temp_dir: str) -> None:
     """Test loading ignore specs when .gitignore is missing."""
-    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(temp_dir)
+    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(
+        temp_dir
+    )
     assert gitignore_spec is None
     assert content_ignore_spec is None
     assert tree_and_content_ignore_spec is not None
@@ -214,7 +193,7 @@ def test_get_tree_structure_with_special_chars(temp_dir: str) -> None:
     # Create files with special characters
     special_dir = os.path.join(temp_dir, "special chars")
     os.makedirs(special_dir)
-    with open(os.path.join(special_dir, "file with spaces.txt"), "w") as f:
+    with open(os.path.join(special_dir, "file with spaces.txt"), "w", encoding='utf-8') as f:
         f.write("test")
 
     tree_output = get_tree_structure(temp_dir)
@@ -223,7 +202,9 @@ def test_get_tree_structure_with_special_chars(temp_dir: str) -> None:
 
 def test_should_ignore_file_edge_cases(sample_repo: str) -> None:
     """Test edge cases for should_ignore_file function."""
-    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(sample_repo)
+    gitignore_spec, content_ignore_spec, tree_and_content_ignore_spec = load_ignore_specs(
+        sample_repo
+    )
 
     # Test with dot-prefixed paths
     assert should_ignore_file(
@@ -264,7 +245,7 @@ def test_save_repo_to_text_with_binary_files(temp_dir: str) -> None:
 def test_save_repo_to_text_custom_output_dir(temp_dir: str) -> None:
     """Test save_repo_to_text with custom output directory."""
     # Create a simple file structure
-    with open(os.path.join(temp_dir, "test.txt"), "w") as f:
+    with open(os.path.join(temp_dir, "test.txt"), "w", encoding='utf-8') as f:
         f.write("test content")
 
     # Create custom output directory
@@ -281,5 +262,44 @@ def test_get_tree_structure_empty_directory(temp_dir: str) -> None:
     # Should only contain the directory itself
     assert tree_output.strip() == "" or tree_output.strip() == temp_dir
 
+def test_empty_dirs_filtering(tmp_path: str) -> None:
+    """Test filtering of empty directories in tree structure generation."""
+    # Create test directory structure with normalized paths
+    base_path = os.path.normpath(tmp_path)
+    src_path = os.path.join(base_path, "src")
+    empty_dir_path = os.path.join(base_path, "empty_dir")
+    tests_path = os.path.join(base_path, "tests")
+
+    os.makedirs(src_path)
+    os.makedirs(empty_dir_path)
+    os.makedirs(tests_path)
+
+    # Create some files
+    with open(os.path.join(src_path, "main.py"), "w", encoding='utf-8') as f:
+        f.write("print('test')")
+    with open(os.path.join(tests_path, "test_main.py"), "w", encoding='utf-8') as f:
+        f.write("def test(): pass")
+
+    # Get tree structure directly using the function
+    tree_output = get_tree_structure(base_path)
+
+    # Print debug information
+    print("\nTree output:")
+    print(tree_output)
+
+    # Basic structure checks for directories with files
+    assert "src" in tree_output
+    assert "tests" in tree_output
+    assert "main.py" in tree_output
+    assert "test_main.py" in tree_output
+
+    # Check that empty directory is not included by checking each line
+    for line in tree_output.splitlines():
+        # Skip the root directory line
+        if base_path in line:
+            continue
+        # Check that no line contains 'empty_dir'
+        assert "empty_dir" not in line, f"Found empty_dir in line: {line}"
+
 if __name__ == "__main__":
     pytest.main([__file__])

tests/test_utils.py

@@ -1,6 +1,10 @@
+"""Test the utils module."""
+
 import logging
-import pytest
 from typing import Generator
+import io
+import pytest
+
 from repo_to_text.utils.utils import setup_logging
 
 @pytest.fixture(autouse=True)
@@ -72,7 +76,8 @@ def test_setup_logging_multiple_calls() -> None:
 
     # Call setup_logging again
     setup_logging(debug=True)
-    assert len(root_logger.handlers) == initial_handler_count, "Should not create duplicate handlers"
+    assert len(root_logger.handlers) == \
+        initial_handler_count, "Should not create duplicate handlers"
 
 def test_setup_logging_level_change() -> None:
     """Test changing log levels between setup_logging calls."""
@@ -96,11 +101,12 @@ def test_setup_logging_message_format() -> None:
     logger = logging.getLogger()
 
     # Create a temporary handler to capture output
-    import io
     log_capture = io.StringIO()
     handler = logging.StreamHandler(log_capture)
     # Use formatter that includes pathname
-    handler.setFormatter(logging.Formatter('%(levelname)s %(name)s:%(pathname)s:%(lineno)d %(message)s'))
+    handler.setFormatter(
+        logging.Formatter('%(levelname)s %(name)s:%(pathname)s:%(lineno)d %(message)s')
+    )
     logger.addHandler(handler)
 
     # Ensure debug level is set
@@ -123,7 +129,6 @@ def test_setup_logging_error_messages() -> None:
     logger = logging.getLogger()
 
     # Create a temporary handler to capture output
-    import io
     log_capture = io.StringIO()
     handler = logging.StreamHandler(log_capture)
     handler.setFormatter(logger.handlers[0].formatter)