Add source files: models and parser

This commit is contained in:
2026-02-02 13:56:56 +00:00
parent 1f432035c8
commit 9e7fd72429

108
src/gdiffer/models.py Normal file
View File

@@ -0,0 +1,108 @@
"""Data models for git diff parsing and analysis."""
from dataclasses import dataclass, field
from typing import Optional
@dataclass
class DiffHunk:
"""Represents a single hunk (chunk) of changes in a diff."""
old_start: int
old_lines: int
new_start: int
new_lines: int
old_lines_content: list[str] = field(default_factory=list)
new_lines_content: list[str] = field(default_factory=list)
header: str = ""
def get_added_lines(self) -> list[tuple[int, str]]:
result = []
for i, line in enumerate(self.new_lines_content):
if line.startswith('+') and not line.startswith('+++'):
result.append((self.new_start + i, line[1:]))
return result
def get_removed_lines(self) -> list[tuple[int, str]]:
result = []
for i, line in enumerate(self.old_lines_content):
if line.startswith('-') and not line.startswith('---'):
result.append((self.old_start + i, line[1:]))
return result
def get_modified_lines(self) -> list[tuple[int, str, str]]:
result = []
added = self.get_added_lines()
removed = self.get_removed_lines()
for i, (old_line, old_content) in enumerate(removed):
if i < len(added):
new_line, new_content = added[i]
if old_content != new_content:
result.append((old_line, old_content, new_content))
return result
@dataclass
class DiffFile:
"""Represents a file in the diff with its changes."""
old_path: Optional[str]
new_path: Optional[str]
new_file_mode: Optional[str] = None
deleted_file_mode: Optional[str] = None
similarity_index: Optional[str] = None
rename_from: Optional[str] = None
rename_to: Optional[str] = None
hunks: list[DiffHunk] = field(default_factory=list)
change_type: str = "modify"
@property
def filename(self) -> str:
if self.new_path:
return self.new_path
return self.old_path or ""
@property
def is_new(self) -> bool:
return self.new_file_mode is not None or self.old_path in [None, "/dev/null"]
@property
def is_deleted(self) -> bool:
return self.deleted_file_mode is not None
@property
def is_rename(self) -> bool:
return self.rename_from is not None
@property
def extension(self) -> str:
filename = self.filename
if '.' in filename:
return filename.rsplit('.', 1)[-1].lower()
return ""
@dataclass
class CodeChange:
"""Represents a code change with context."""
file: DiffFile
hunk: Optional[DiffHunk]
old_code: str
new_code: str
language: str = "unknown"
summary: str = ""
issues: list[dict] = field(default_factory=list)
suggestions: list[str] = field(default_factory=list)
@dataclass
class DiffAnalysis:
"""Complete analysis result for a diff."""
files: list[DiffFile] = field(default_factory=list)
total_files: int = 0
files_added: int = 0
files_deleted: int = 0
files_modified: int = 0
files_renamed: int = 0
total_changes: int = 0
language_breakdown: dict[str, int] = field(default_factory=dict)
all_issues: list[dict] = field(default_factory=list)
all_suggestions: list[str] = field(default_factory=list)