Merge pull request #27 from tawanda-kembo/feat/improve-ignore-rules

fix: Improve .gitignore handling and add support for directory patterns

.gitignore

@@ -139,4 +139,6 @@ __pycache__/
 # Distribution / packaging
 *.egg-info/
 build/
-dist/
+dist/
+.aider*
+.aider.*

code_collator/__init__.py

@@ -0,0 +1 @@
+from .collate import collate_codebase, main  # noqa: F401

code_collator/collate.py

@@ -2,14 +2,14 @@
 import argparse
 from pathlib import Path
 import logging
+from fnmatch import fnmatch
 
 
 def setup_logging():
     """Set up logging configuration."""
     logging.basicConfig(
         level=logging.INFO,
-        format='%(asctime)s - %(levelname)s - %(message)s',
-        force=True
+        format='%(asctime)s - %(levelname)s - %(message)s'
     )
 
 
@@ -42,10 +42,27 @@ def read_gitignore(path):
 
 def should_ignore(file_path, ignore_patterns):
     """Check if a file should be ignored based on .gitignore patterns and if it's in the .git directory."""
-    from fnmatch import fnmatch
     if '.git' in Path(file_path).parts:
         return True
-    return any(fnmatch(file_path, pattern) for pattern in ignore_patterns)
+
+    relative_path = os.path.relpath(file_path)
+    path_parts = relative_path.split(os.sep)
+
+    for pattern in ignore_patterns:
+        if pattern.endswith('/'):
+            # Directory pattern
+            if any(fnmatch(part, pattern[:-1]) for part in path_parts):
+                return True
+        elif '/' in pattern:
+            # Path pattern
+            if fnmatch(relative_path, pattern):
+                return True
+        else:
+            # File pattern
+            if fnmatch(os.path.basename(file_path), pattern):
+                return True
+
+    return False
 
 
 def collate_codebase(path, output_file):

setup.py

@@ -4,10 +4,12 @@
 
 here = pathlib.Path(__file__).parent.resolve()
 
+
 def get_version():
     version = os.environ.get('PACKAGE_VERSION', '0.0.0')
     return version
 
+
 setup(
     name="code-collator",
     version=get_version(),
@@ -29,7 +31,7 @@ def get_version():
         "Programming Language :: Python :: 3.11",
     ],
     keywords="cli, development, documentation",
-    packages=find_packages(),
+    packages=find_packages(include=['code_collator', 'code_collator.*']),
     python_requires=">=3.6, <4",
     install_requires=[
         # Add any dependencies here
@@ -43,4 +45,4 @@ def get_version():
         "Bug Reports": "https://github.com/tawanda-kembo/code-collator/issues",
         "Source": "https://github.com/tawanda-kembo/code-collator",
     },
-)
+)

tests/test_collate.py

@@ -1,28 +1,33 @@
+import sys
+import os
 import pytest
-from unittest.mock import mock_open, patch
-from code_collator import collate
 import logging
+from unittest.mock import mock_open, patch
+from code_collator.collate import is_binary_file, read_gitignore, should_ignore, main
+
+# Add the parent directory to sys.path
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
 
 
 def test_is_binary_file():
     with patch('builtins.open', mock_open(read_data=b'\x00binary\xff')):
-        assert collate.is_binary_file('test.bin') is True
+        assert is_binary_file('test.bin') is True
 
     with patch('builtins.open', mock_open(read_data=b'hello world')):
-        assert collate.is_binary_file('test.txt') is False
+        assert is_binary_file('test.txt') is False
 
 
 def test_read_gitignore():
     with patch('builtins.open', mock_open(read_data='*.pyc\n__pycache__\n')):
-        patterns = collate.read_gitignore('.')
+        patterns = read_gitignore('.')
         assert patterns == ['*.pyc', '__pycache__']
 
 
 def test_should_ignore():
     patterns = ['*.pyc', '__pycache__']
-    assert collate.should_ignore('test.pyc', patterns)
-    assert collate.should_ignore('test.py', patterns) is False
-    assert collate.should_ignore('.git/config', patterns)
+    assert should_ignore('test.pyc', patterns)
+    assert should_ignore('test.py', patterns) is False
+    assert should_ignore('.git/config', patterns)
 
 
 @pytest.fixture
@@ -37,7 +42,7 @@ def mock_file_system(tmp_path):
 # def test_collate_codebase(mock_file_system, caplog):
 #     caplog.set_level(logging.INFO)
 #     output_file = mock_file_system / "output.md"
-#     collate.collate_codebase(str(mock_file_system), str(output_file))
+#     collate_codebase(str(mock_file_system), str(output_file))
 
 #     with open(output_file, 'r') as f:
 #         content = f.read()
@@ -57,7 +62,7 @@ def mock_file_system(tmp_path):
 def test_main(mock_file_system, caplog):
     caplog.set_level(logging.INFO)
     with patch('sys.argv', ['collate', '-p', str(mock_file_system), '-o', 'output.md']):
-        collate.main()
+        main()
 
     assert "Starting code collation" in caplog.text
     assert "Code collation completed" in caplog.text