Corporate Diff Engine

A pluggable Python library for creating semantic and line-by-line diffs for various file formats. Designed for use in backend systems working with file storages like S3/MinIO.

Key Features

• Strategy Pattern: Choose the best diff algorithm based on file type (.md, .json, etc.).
• Extensible: Easily add new strategies for custom formats (e.g., .py, .yaml, .xml).
• Storage Agnostic: Works with any async storage client that implements a simple protocol.
• Modern: Built with Python 3.10+, Pydantic, and async support in mind.

Installation

From source:

pip install .

Or to install in editable mode for development:

pip install -e .[dev]

Usage

The library is designed to be integrated into a service that has access to a file storage.

import asyncio
from diff_engine import StorageDiffProvider, IAsyncStorageClient

# 1. Create a storage client adapter that matches the protocol
# This is an example for a local file system. In a real app,
# this would be your MinIO or S3 client wrapper.
class MockStorageClient(IAsyncStorageClient):
    async def get_object_as_text(self, path: str) -> str:
        with open(path, 'r', encoding='utf-8') as f:
            return f.read()

async def main():
    # 2. Instantiate the provider with your storage client
    storage_client = MockStorageClient()
    diff_provider = StorageDiffProvider(storage_client)

    # 3. Get a diff between two files
    # The engine will automatically select the TextDiffStrategy for .md
    md_patch = await diff_provider.get_diff("tests/fixtures/sample_v1.md", "tests/fixtures/sample_v2.md")
    print("--- Markdown Diff ---")
    print(f"Lines Added: {md_patch.added_lines}, Removed: {md_patch.removed_lines}")
    print(md_patch.diff_text)

    # The engine will automatically select the JsonDiffStrategy for .json
    json_patch = await diff_provider.get_diff("tests/fixtures/sample_v1.json", "tests/fixtures/sample_v2.json")
    print("\n--- JSON Diff ---")
    print(json_patch.diff_text)


if __name__ == "__main__":
    asyncio.run(main())

For Developers

To run tests:

pytest

To format code: black . isort . (Файлы для tests/ и exceptions.py опущены для краткости, но их структура очевидна из контекста)