Skip to content

Changelog

All notable changes to pyfs-watcher are documented here.

v0.3.0

Breaking: Reorganised the public Python API to keep the top-level namespace clean.

  • pyfs_watcher.errors — All exception classes (WalkError, HashError, CopyError, etc.) now live in a dedicated errors submodule.
  • pyfs_watcher.types — All dataclass/result types (DiffResult, SearchMatch, SnapshotResult, DuEntry, RenamePreview, RenameResult) now live in a dedicated types submodule.
  • Top-level exports are now limited to the public functions (walk, walk_collect, hash_file, hash_files, copy_files, move_files, find_duplicates, search, search_iter, diff_dirs, sync, snapshot, verify, disk_usage, bulk_rename) and the FileWatcher class.

Migration:

# Before (0.2.x)
from pyfs_watcher import WalkError, DiffResult

# After (0.3.0)
from pyfs_watcher.errors import WalkError
from pyfs_watcher.types import DiffResult

v0.2.0

Six new feature modules, transforming pyfs-watcher from fast filesystem utils into a comprehensive filesystem toolkit:

  • Search — Parallel content search (search(), search_iter()) using Rust's regex crate. Supports glob filtering, case-insensitive matching, context lines, max file size limits, and binary file detection.
  • Diff — Directory comparison (diff_dirs()) with content-aware diffing and optional move detection. Reports added, removed, modified, unchanged, and moved files.
  • Sync — Incremental directory sync (sync()) that only copies changed files. Supports dry_run preview, delete_extra cleanup, progress callbacks, and error-tolerant operation.
  • Snapshot — File integrity monitoring (snapshot(), verify()) with JSON-based snapshots. Captures file hashes, sizes, mtimes, and permissions. Detects additions, removals, and modifications.
  • Disk Usage — Parallel size calculation (disk_usage()) with per-child breakdown sorted by size. Supports glob filtering and hidden file control.
  • Rename — Regex-based batch rename (bulk_rename()) with dry_run=True by default for safety. Supports recursive operation, directory renaming, and undo().

Additional changes:

  • Added shared WalkFilter infrastructure in utils.rs for consistent filtering across all features
  • Added 6 new typed exception classes (SearchError, DirDiffError, SyncError, SnapshotError, DiskUsageError, RenameError)
  • Full type stubs for all new classes and functions
  • Fixed jwalk skip_hidden default — hidden files now correctly appear when skip_hidden=False
  • New Cargo dependencies: regex, serde, serde_json, chrono

v0.1.1

  • Bumped project version
  • Set up CI with GitHub Actions (lint, typecheck, Rust checks, tests)
  • Added pre-commit hooks for code quality
  • Renamed package to pyfs-watcher / pyfs_watcher
  • Created virtual environment workflow for CI test execution

v0.1.0

Initial release with five feature modules:

  • Walk — Parallel recursive directory traversal (walk(), walk_collect()) powered by jwalk
  • Hash — BLAKE3 and SHA-256 file hashing (hash_file(), hash_files()) with memory-mapped I/O and parallel batch processing
  • Copy/Move — Bulk file copy and move (copy_files(), move_files()) with progress callbacks and cross-device move fallback
  • Watch — Cross-platform filesystem watcher (FileWatcher, async_watch()) with debouncing and ignore patterns
  • Dedup — Three-stage duplicate file detection (find_duplicates()) with size grouping, partial hash, and full hash

Additional features:

  • Full type stubs (py.typed) for IDE autocompletion and static analysis
  • Typed exception hierarchy (FsWatcherError and subclasses)
  • Logging bridge from Rust to Python via pyo3-log
  • GitHub Actions workflow for PyPI publishing