🔬 File & Deduplication Analyzers Reference

Technical interface specifications for the 23 advanced file, media, and deduplication analyzers in cortex_unified.analyzers.

---

⚡ Core Analyzer Classes

1. CzkawkaDuplicateFinder

• Module: cortex_unified.analyzers.czkawka_duplicate_finder
• Algorithm: 3-stage validation (File Size $\rightarrow$ Header Hash $\rightarrow$ Full BLAKE3 digest).
• Methods:
• find_duplicates(directories: list[str]) -> list[DuplicateGroup]
• set_exclude_dirs(paths: list[str])
• cancel(): Immediately halts multi-threaded file walking.

2. PerceptualDuplicateFinder

• Module: cortex_unified.analyzers.perceptual_duplicate_finder
• Algorithm: Difference Hash (dHash) and Discrete Cosine Transform Perceptual Hash (pHash).
• Methods:
• compute_hash(image_path: str, kind: str = "phash") -> int
• hamming_distance(hash1: int, hash2: int) -> int
• find_similar(gallery_dir: str, threshold: int = 5) -> list[ImageGroup]

3. ContentDefinedChunker

• Module: cortex_unified.analyzers.content_defined_chunker
• Algorithm: FastCDC (Fast Content-Defined Chunking) with Gear hashing and Jaccard similarity index.
• Methods:
• chunk_file(file_path: str, min_size: int = 2048, avg_size: int = 8192, max_size: int = 32768) -> list[Chunk]
• jaccard_similarity(chunks_a: list[Chunk], chunks_b: list[Chunk]) -> float

4. FuzzyFinder

• Module: cortex_unified.analyzers.fuzzy_finder
• Algorithm: Context-Triggered Piecewise Hashing (CTPH / Ssdeep).
• Methods:
• fuzzy_hash(file_path: str) -> str
• compare_fuzzy(hash1: str, hash2: str) -> int: Returns similarity score from 0 to 100.