Closed
Conversation
…phism This refactor introduces a clean inheritance architecture using proper Python polymorphism: Architecture: - Add Source base class in dedicated source.py file (common metadata/extra fields) - Refactor FileSystemNode to inherit from Source with full backward compatibility - Create specialized classes: FileSystemFile, FileSystemDirectory, FileSystemSymlink, GitRepository - render_tree() method belongs to FileSystemNode level (tree-specific, not all sources need it) Pure Python Polymorphism: - Each subclass implements its own get_sort_priority() and get_content() methods - NO type property or enum needed - use isinstance() directly - FileSystemFile.get_sort_priority() returns 0 (files first) - FileSystemDirectory.get_content() raises ValueError (directories can't have content) - FileSystemSymlink.get_content() returns target path (what symlink points to) - Clean, extensible design following Python best practices Removed Legacy Type System: - Completely removed FileSystemNodeType enum - No more type property - use isinstance() everywhere - Constructors now use specific classes: FileSystemFile(), FileSystemDirectory(), etc. - Pure polymorphism without any type checking properties Code Changes: - src/gitingest/schemas/source.py: New base Source class - src/gitingest/schemas/filesystem.py: Refactored with polymorphic methods, Path import in TYPE_CHECKING - src/gitingest/ingestion.py: Use specific constructors, populate symlink targets - src/gitingest/output_formatter.py: Use isinstance() instead of enum comparisons - Remove all FileSystemNodeType imports and usage - All pre-commit hooks pass (ruff-check, ruff-format, etc.) Benefits: - True Python polymorphism where each class knows its own behavior - No explicit type checking needed - Python dispatches automatically - More extensible - adding new source types just requires implementing methods - Cleaner code without enum/string type comparisons - Full backward compatibility maintained
Replace all isinstance() type checking with proper polymorphic methods: Pure Polymorphism Methods: - get_summary_info(): Each class returns its own summary format - is_single_file(): Boolean check without isinstance() - gather_contents(): Recursive content gathering via method dispatch - get_display_name(): Tree display formatting (/, -> target, etc.) - has_children(): Check for child nodes without type checking Benefits: - No isinstance() 'clochard' style code anywhere - True duck typing - just call methods and let Python dispatch - Cleaner, more maintainable code - Each class encapsulates its own behavior - Easy to extend with new node types Code Changes: - FileSystemNode: Base implementations for all methods - FileSystemFile: is_single_file()=True, summary with line count - FileSystemDirectory: get_display_name() adds '/', has_children() checks list - FileSystemSymlink: get_display_name() shows '-> target' - output_formatter.py: Use polymorphic methods instead of isinstance() This is proper OOP - objects know their own behavior!
NicolasIRAGNE
changed the title
cursor
bot
force-pushed
the
refactor/source-inheritance
branch
from
217afe2 to
87cbb5a
Compare
Contributor
There was a problem hiding this comment.
Pull Request Overview
This PR introduces a Source base class and refactors the filesystem schema to use proper inheritance and polymorphism. The refactoring eliminates the enum-based type system in favor of specialized classes for different filesystem node types.
Key changes:
- Introduces a new
Sourcebase class with common fields (metadataandextra) - Replaces
FileSystemNodeTypeenum with specialized classes (FileSystemFile,FileSystemDirectory,FileSystemSymlink,GitRepository) - Moves type-specific behavior into polymorphic methods on the respective classes
Reviewed Changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated 6 comments.
Show a summary per file
| File | Description |
|---|---|
src/gitingest/schemas/source.py |
New base class with common metadata fields |
src/gitingest/schemas/filesystem.py |
Major refactor replacing enum-based typing with inheritance hierarchy |
src/gitingest/schemas/__init__.py |
Updated exports to include new classes |
src/gitingest/output_formatter.py |
Updated to use polymorphic methods instead of type checking |
src/gitingest/ingestion.py |
Updated to instantiate specific node types instead of generic nodes |
.vscode/launch.json |
Added reload flag to debug configuration |
src/gitingest/schemas/filesystem.py
Outdated
Comment on lines
71
to
82
| """Return sort priority. Override in subclasses.""" | ||
| return 1 # Default: not a file | ||
|
|
There was a problem hiding this comment.
The comment '# Default: not a file' is unclear. Consider documenting the actual priority values and what they represent (e.g., '# Default priority for non-file nodes (directories, symlinks)').
Suggested change
| """Return sort priority. Override in subclasses.""" | |
| return 1 # Default: not a file | |
| """Return sort priority. Override in subclasses. | |
| Returns | |
| ------- | |
| int | |
| Sort priority value. By convention: | |
| 0 - file nodes (highest priority, e.g., README) | |
| 1 - default for non-file nodes (directories, symlinks, etc.) | |
| Subclasses should override as appropriate. | |
| """ | |
| return 1 # Default priority for non-file nodes (directories, symlinks, etc.) |
Comment on lines
+98
to
104
| from gitingest.utils.file_utils import _decodes, _get_preferred_encodings, _read_chunk | ||
| from gitingest.utils.notebook import process_notebook | ||
|
|
There was a problem hiding this comment.
Imports inside methods can impact performance and readability. Consider moving these imports to the top of the file or module level.
Suggested change
| from gitingest.utils.file_utils import _decodes, _get_preferred_encodings, _read_chunk | |
| from gitingest.utils.notebook import process_notebook | |
| from gitingest.utils.notebook import process_notebook |
Comment on lines
+99
to
104
| from gitingest.utils.notebook import process_notebook | ||
|
|
There was a problem hiding this comment.
Imports inside methods can impact performance and readability. Consider moving these imports to the top of the file or module level.
Suggested change
| from gitingest.utils.notebook import process_notebook |
src/gitingest/schemas/filesystem.py
Outdated
Comment on lines
170
to
167
| return f"File: {self.name}\nLines: {len(self.content.splitlines()):,}\n" | ||
|
|
||
| def is_single_file(self) -> bool: | ||
| """Files are single files.""" | ||
| return True | ||
|
|
||
| def render_tree(self, prefix: str = "", *, is_last: bool = True) -> list[str]: | ||
| """Render the tree representation of this file.""" | ||
| current_prefix = "└── " if is_last else "├── " | ||
| return [f"{prefix}{current_prefix}{self.name}"] | ||
|
|
||
|
|
There was a problem hiding this comment.
Calling self.content in summary info will trigger file reading, which could be expensive. Consider caching content or only reading when actually needed for display.
Suggested change
| return f"File: {self.name}\nLines: {len(self.content.splitlines()):,}\n" | |
| def is_single_file(self) -> bool: | |
| """Files are single files.""" | |
| return True | |
| def render_tree(self, prefix: str = "", *, is_last: bool = True) -> list[str]: | |
| """Render the tree representation of this file.""" | |
| current_prefix = "└── " if is_last else "├── " | |
| return [f"{prefix}{current_prefix}{self.name}"] | |
| if self._cached_content is not None: | |
| line_count = len(self._cached_content.splitlines()) | |
| return f"File: {self.name}\nLines: {line_count:,}\n" | |
| else: | |
| return f"File: {self.name}\nLines: (not loaded)\n" | |
| def is_single_file(self) -> bool: | |
| """Files are single files.""" | |
| return True | |
| @property | |
| def content(self) -> str: | |
| """Return file content, caching after first read.""" | |
| if self._cached_content is None: | |
| self._cached_content = self.get_content() | |
| return self._cached_content |
src/gitingest/schemas/filesystem.py
Outdated
| current_prefix = "└── " if is_last else "├── " | ||
| display_name = self.name + "/" | ||
| lines.append(f"{prefix}{current_prefix}{display_name}") | ||
| if hasattr(self, "children") and self.children: |
There was a problem hiding this comment.
Using hasattr check for children is unnecessary since children is defined as a field with default factory in the base class. This check can be simplified to just if self.children:.
Suggested change
| if hasattr(self, "children") and self.children: | |
| if self.children: |
src/gitingest/schemas/filesystem.py
Outdated
| # Mark as git repo in the tree | ||
| display_name = f"{self.name}/ (git repository)" | ||
| lines.append(f"{prefix}{current_prefix}{display_name}") | ||
| if hasattr(self, "children") and self.children: |
There was a problem hiding this comment.
Using hasattr check for children is unnecessary since children is defined as a field with default factory in the base class. This check can be simplified to just if self.children:.
Suggested change
| if hasattr(self, "children") and self.children: | |
| if self.children: |
cursor
bot
force-pushed
the
refactor/source-inheritance
branch
2 times, most recently
from
46c7f5d to
26ccecd
Compare
…thod Even simpler approach: - Replace node.has_children() with direct if node.children: - Remove unnecessary has_children() methods from all classes - Pythonic and direct - empty lists are falsy, non-empty are truthy - Less code, same functionality This is the most straightforward way to check for children in Python.
cursor
bot
force-pushed
the
refactor/source-inheritance
branch
from
26ccecd to
280d3f0
Compare
|
Hi there! We haven’t seen activity on this pull request for 45 days, so I’m marking it as stale. |
|
Hi there! We haven’t heard anything for 10 days, so I’m closing this pull request. Feel free to reopen if you’d like to continue working on it. Thanks! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.