feat(aggregation): Implement hash-based summary cache

tests/test_summary_cache.py

@@ -0,0 +1,46 @@
+import os
+import shutil
+from pathlib import Path
+from src.summary_cache import SummaryCache, get_file_hash
+
+def test_get_file_hash():
+ content = "hello world"
+ # sha256 of "hello world"
+ expected = "b94d27b9934d3e08a52e52d7da7dabfac484efe37a5380ee9088f7ace2efcde9"
+ assert get_file_hash(content) == expected
+
+def test_summary_cache():
+ cache_dir = Path(".test_cache")
+ if cache_dir.exists():
+  shutil.rmtree(cache_dir)
+ cache_file = cache_dir / "cache.json"
+ 
+ cache = SummaryCache(str(cache_file))
+ 
+ file_path = "test.py"
+ content = "print('hello')"
+ content_hash = get_file_hash(content)
+ summary = "**Python** - 1 lines"
+ 
+ # Test empty cache
+ assert cache.get_summary(file_path, content_hash) is None
+ 
+ # Test set and get
+ cache.set_summary(file_path, content_hash, summary)
+ assert cache.get_summary(file_path, content_hash) == summary
+ 
+ # Test cache invalidation
+ assert cache.get_summary(file_path, "different_hash") is None
+ 
+ # Test persistence
+ cache2 = SummaryCache(str(cache_file))
+ assert cache2.get_summary(file_path, content_hash) == summary
+ 
+ # Cleanup
+ if cache_dir.exists():
+  shutil.rmtree(cache_dir)
+
+if __name__ == "__main__":
+ test_get_file_hash()
+ test_summary_cache()
+ print("Tests passed!")