annextube 0.14.0__tar.gz

annextube-0.14.0/.autorc

@@ -0,0 +1,83 @@
+{
+  "baseBranch": "master",
+  "noVersionPrefix": false,
+  "plugins": [
+    "git-tag",
+    [
+      "exec",
+      {
+        "afterRelease": "python -m build && twine upload dist/*"
+      }
+    ],
+    "released"
+  ],
+  "labels": [
+    {
+      "name": "release-major",
+      "releaseType": "major",
+      "changelogTitle": "💥 Breaking Change",
+      "description": "Increment the major version when merged",
+      "color": "C5000B"
+    },
+    {
+      "name": "release-minor",
+      "releaseType": "minor",
+      "changelogTitle": "🚀 Enhancement",
+      "description": "Increment the minor version when merged",
+      "color": "F1A60E"
+    },
+    {
+      "name": "release-patch",
+      "releaseType": "patch",
+      "changelogTitle": "🐛 Bug Fix",
+      "description": "Increment the patch version when merged",
+      "color": "870048"
+    },
+    {
+      "name": "release",
+      "releaseType": "release",
+      "description": "Create a release when this PR is merged",
+      "color": "16a34a"
+    },
+    {
+      "name": "skip-release",
+      "releaseType": "skip",
+      "description": "Preserve the current version when merged",
+      "color": "bf5416"
+    },
+    {
+      "name": "internal",
+      "releaseType": "none",
+      "changelogTitle": "🏠 Internal",
+      "description": "Changes only affect the internal API",
+      "color": "696969"
+    },
+    {
+      "name": "documentation",
+      "releaseType": "none",
+      "changelogTitle": "📝 Documentation",
+      "description": "Changes only affect the documentation",
+      "color": "cfd3d7"
+    },
+    {
+      "name": "tests",
+      "releaseType": "none",
+      "changelogTitle": "🧪 Tests",
+      "description": "Add or improve existing tests",
+      "color": "e4e669"
+    },
+    {
+      "name": "performance",
+      "releaseType": "patch",
+      "changelogTitle": "🏎 Performance",
+      "description": "Improve performance of an existing feature",
+      "color": "f4b2d8"
+    },
+    {
+      "name": "released",
+      "releaseType": "none",
+      "description": "This PR has been released",
+      "color": "84cc16"
+    }
+  ]
+}

annextube-0.14.0/.gitignore

@@ -0,0 +1,119 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# Virtual environments
+.venv/
+venv/
+ENV/
+env/
+.virtualenv
+uv.lock
+
+# Testing
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.log
+.pytest_cache/
+htmlcov/
+
+# Type checking
+.mypy_cache/
+.dmypy.json
+dmypy.json
+.pyre/
+.pytype/
+
+# IDEs
+.vscode/
+.idea/
+.vim/
+.viminfo
+*.swp
+*.swo
+*~
+.DS_Store
+Thumbs.db
+
+# Project specific
+.annextube/
+test-archive/
+test-archives/
+*.tmp
+
+# Node.js (frontend)
+node_modules/
+.npm/
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+.lesshst
+
+# Build artifacts
+*.min.js
+*.min.css
+web/
+demo/
+frontend/dist/
+frontend/build/
+pagefind/
+
+# Environment files
+.env
+.env.local
+.env.*.local
+*.key
+*.pem
+credentials.json
+
+# Documentation
+docs/_build/
+docs/.hugo_build.lock
+
+# Shell history and local config
+.bash_history
+.config/  # User config created during testing
+.anaconda/
+.conda/
+.claude/settings.local.json
+annextube/_version.py
+.deno
+.local
+.worktrees
+.config
+.ssh
+
+# Various files claude likes to create
+CHANGES_SUMMARY.md
+DEMO_SETUP.md
+EXAMPLE_UNAVAILABLE_FILTERING.md
+
+# no duct logs here -- only for tests etc
+.duct
+.claude/worktrees

annextube-0.14.0/.jscpd.json

@@ -0,0 +1,18 @@
+{
+  "threshold": 0,
+  "reporters": ["console"],
+  "ignore": [
+    "**/node_modules/**",
+    "**/.git/**",
+    "**/dist/**",
+    "**/.tox/**",
+    "**/.venv/**",
+    "**/__pycache__/**",
+    "**/web/**",
+    "**/.specify/**",
+    "**/specs/**",
+    "**/tools/**"
+  ],
+  "minTokens": 50,
+  "minLines": 5
+}

annextube-0.14.0/CHANGELOG.md

@@ -0,0 +1,359 @@
+# v0.14.0 (Fri Sep 04 2026)
+
+#### 🚀 Enhancement
+
+- Searchable video descriptions + Metadata/Full search modes (FR-042d–g) [#6](https://github.com/con/annextube/pull/6) ([@claude](https://github.com/claude) [@yarikoptic](https://github.com/yarikoptic) [@yarikoptic-gitmate](https://github.com/yarikoptic-gitmate))
+- Flag approximate caption search matches and reorder results [#7](https://github.com/con/annextube/pull/7) ([@claude](https://github.com/claude) [@yarikoptic-gitmate](https://github.com/yarikoptic-gitmate))
+
+#### ⚠️ Pushed to `master`
+
+- fix(release): use releaseType "none" for the "released" label ([@yarikoptic](https://github.com/yarikoptic))
+
+#### 🏠 Internal
+
+- feat(release): introduce intuit/auto for automated releases [#8](https://github.com/con/annextube/pull/8) ([@yarikoptic](https://github.com/yarikoptic))
+
+#### Authors: 3
+
+- Claude ([@claude](https://github.com/claude))
+- GitMate for @yarikoptic ([@yarikoptic-gitmate](https://github.com/yarikoptic-gitmate))
+- Yaroslav Halchenko ([@yarikoptic](https://github.com/yarikoptic))
+
+---
+
+# Changelog
+
+<!-- Entries from v0.14.0 onward are auto-generated by intuit/auto. -->
+<!-- Entries prior to v0.14.0 use Keep a Changelog format. -->
+
+## [0.13.0] - 2026-07-13
+
+### Added
+
+- **Multi-channel collection management (feature 003)**: DataLad superdataset workflow for archives that aggregate multiple channels — `annextube collection add`, `annextube aggregate`, breadcrumb navigation in the web UI, and the corresponding contract / integration / E2E test suites. Ships alongside a spec artifact under `.specify/specs/003-multi-channel-collections/`
+- **Retroactive caption curation during backup** (FR-062a): backups can now run the caption curation pipeline over existing archived videos, not just newly-fetched ones
+- **`annextube completion` command** (FR-057i, T141): outputs shell completion scripts for bash, zsh, and fish via Click's `ShellComplete.source()`; auto-detects the current shell from `$SHELL`
+- **MKV video playback: browser-specific error messages**: the web player now explains which browsers can natively decode MKV and links users to workarounds. Feature spec `.specify/specs/002-mkv-video-playback/`
+- **GitHub Actions test workflow**: tox-based Python matrix (3.10 – 3.14) plus lint / type / coverage jobs, running on every push and PR
+- **JSON backup output (T030) + batch API optimization (T039)**: completes remaining MVP gaps identified in `speckit.analyze`
+
+### Fixed
+
+- **yt-dlp playlist pagination truncation**: older yt-dlp releases silently truncated large YouTube playlists at ~100 entries even though `playlist_count` reported the true size, causing videos to appear-and-disappear across runs. Fixed upstream in yt-dlp/yt-dlp#16948; minimum requirement bumped to `yt-dlp>=2026.07.04`. Layered defence in the archive too: `Archiver._save_playlist_metadata` refuses to persist truncated lists, and when a YouTube Data API key is configured `YouTubeService.get_playlist_metadata` cross-checks against `playlistItems.list` and prefers the API's authoritative list
+- **YouTube API robustness**: HTTP error bodies decoded with `errors='replace'`; mid-pagination quota-exceeded events resume from the current `pageToken` instead of restarting from page 1 (applies to `get_playlist_video_ids` and `fetch_comments`)
+- **`annextube init` on an existing git-annex repo**: skip repo init when the target is already a git-annex working tree, and configure `annex.security.allowed-*` policies so the reused repo doesn't drop content
+- **Single-channel archive detection requires `.annextube/config.toml`**: previously a bare git repo could be misdetected as a single-channel archive
+- **Thumbnail download failures downgraded to warnings**: no longer aborts a backup
+
+### Changed
+
+- **Zero-tolerance code duplication**: dedicated jscpd tox env with threshold 0 %; the entire codebase now reports 0 clones after helper extraction (batch prefetch, video-processing loop, symlink writer, unavailable-video recorder, playlist-tab-entry parser, atomic-write prep, shared `--output-dir` Click decorator)
+- **Constitution amended to v1.5.0**: adds Principle XIII "DataLad-Native Operations"
+
+## [0.12.0] - 2026-03-11
+
+### Added
+
+- **`annextube init` modular flags**: `--curation`, `--search`, `--enable-all` for selective feature initialization
+- **`[full]` extra**: `pip install annextube[full]` installs all optional dependencies
+- **Auto-build search index after backup**: Pagefind caption index is rebuilt automatically when captions change
+- **Real-pagefind integration tests**: 7 tests exercise the actual pagefind binary with 30s timeout to catch IPC stalls
+- **WCAG 2.1 AA accessibility tests**: Automated contrast-ratio and a11y checks for the web frontend
+- **Zero-tolerance code duplication**: jscpd enforced at threshold 0 across the codebase
+
+### Fixed
+
+- **Pagefind IPC deadlock on DataLad subdatasets**: The pagefind binary hangs when its output directory contains git-annex symlinks (read-only targets) from a previous build. Fixed by writing to a clean temp directory, then syncing into the subdataset
+- **Pagefind IPC pipe interlocking**: Separate stderr from stdout, bypass the upstream library's fragile polling task, use direct IPC with 300s timeout for WriteFiles
+- **Color contrast violation**: Fixed insufficient contrast ratio in frontend CSS
+
+### Changed
+
+- **DataLad is now a core dependency**: Removed plain git submodule fallback; all subdataset operations use DataLad exclusively
+- **DataLad cfg_text2git for pagefind subdataset**: Use DataLad's built-in text2git configuration instead of manual .gitattributes
+- **yt-dlp[default] co-installs challenge solver**: No longer need separate `yt-dlp-youtube-oauth2` or manual solver setup
+- **Consolidate scripts/ into tools/**: All helper scripts live in `tools/`, tox.ini calls them directly
+- **Spec harmonization**: Aligned specification artifacts with implementation reality
+
+## [0.11.0] - 2026-03-04
+
+### Added
+
+- **Full-text caption search via Pagefind**: Search across all video captions with highlighted results and click-to-seek
+- **`build-search-index` CLI command**: Standalone pagefind index generation
+- **DataLad subdataset management for pagefind index**: Automatic subdataset creation with cfg_text2git
+- **Caption search UX polish**: Improved search result display and git submodule for pagefind frontend assets
+
+### Fixed
+
+- **Duplicate pagefind/ output at archive root**: Index files now go only to `web/pagefind/`
+- **Pagefind 100ms polling sleep**: Removed upstream library's polling delay causing ~140x slowdown
+- **Curated caption writes on annexed files**: Use AtomicFileWriter to handle symlink-to-readonly targets
+- **Pagefind URL resolution**: Fixed incorrect paths in subdirectory deployments
+
+## [0.10.0] - 2026-03-02
+
+### Added
+
+- **Aggregated error reporting**: Record all errors during backup and fail at end with summary instead of aborting on first error
+
+### Fixed
+
+- **metadata.json writes on annexed files**: Use UTF-8 encoding and handle annexed file targets correctly
+- **git diff quoting for non-ASCII paths**: Fixed diff parsing for filenames with special characters
+- **Unavailable video handling**: Skip downloads for known-unavailable videos in incremental channel backup
+- **Thumbnail re-download**: Skip unless in force mode
+- **TSV commit failure**: Target specific files and deduplicate commit logic
+- **Playlist-exclusive detection**: Include existing and unavailable video IDs
+
+### Changed
+
+- **Stop writing volatile timestamps to channel.json**: Removed last_sync/fetched_at fields that caused unnecessary commits
+
+## [0.9.0] - 2026-02-18
+
+### Added
+
+- **`embed-config` command**: Propagate shared TOML settings (e.g. `[curation]` glossary) from a super-dataset config into subdataset configs with comment-preserving deep merge. Supports `--existing keep` (default) and `--existing update` modes. Skips `[[sources]]` (per-subdataset)
+- **Caption curation design spec**: Planning document for the curation pipeline
+- **`[curation]` section in config template**: Generated configs now include documented curation settings (glossary, fuzzy matching, LLM, audio alignment)
+- **Glossary discovery with parent collation**: `glossary_path` + `glossary_collate_parents` in `[curation]` config for multi-channel setups — walk up parent dirs and merge all glossaries found
+- **Caption curation pipeline**: 8-stage ASR correction (`curate-captions` command) — glossary regex, LLM corrections, fuzzy matching, filler removal, ASR artifact fixes, sentence segmentation, cue chunking, timestamp restoration
+- **Clone command panel**: Web UI header button to clone the archive repository
+- **`tomlkit` dependency**: Comment-preserving TOML editing for `embed-config`
+- **`types-PyYAML` dev dependency**: Fix mypy type checking for yaml imports
+
+### Fixed
+
+- **`curate-captions` ignoring archive config for VIDEO_PATH**: Direct video directory mode was creating bare `CurationConfig()` defaults instead of loading `.annextube/config.toml`, silently ignoring `glossary_path` and other `[curation]` settings
+- **Transcript language selector reverting**: Fixed CaptionBrowser resetting user language picks on reactive updates
+- **Clone command UI**: Fixed reactive updates, tab switching, and inline layout
+- **Curated captions not default in transcript browser**: CaptionBrowser now auto-selects the curated variant (e.g. `en-curated`) over the base language; `en-curated` displays as "EN (curated)"
+
+### Changed
+
+- **Skip already-archived videos during playlist backup**: Incremental playlist backup no longer re-processes existing videos
+
+## [0.8.0] - 2026-02-16
+
+### Added
+
+- **yt-dlp rate-limit detection and concurrency limiting**: Detects YouTube 429/rate-limit responses and retries with exponential backoff. Cross-process concurrency limiting via file locks prevents parallel yt-dlp calls from triggering rate limits
+- **`tox -e sdist-check`**: Automated pre-release verification — builds sdist, installs in clean venv, verifies built frontend is included and `generate-web` works
+- **`tox -e full`**: Run all non-network tests with all optional deps (including playwright)
+- **Pre-release sdist verification checklist**: Documented manual tarball inspection steps
+
+### Fixed
+
+- **Catastrophically slow sdist build**: `skip-excluded-dirs = true` prevents hatchling from walking into multi-GB cache directories (201s → 0.04s)
+- **Built frontend missing from sdist**: `force-include` and `artifacts` config ensures `web/` is included; dropped frontend source from sdist
+- **Skip known-unavailable videos during playlist backup**: Avoid re-attempting videos that are known to be unavailable
+- **Rate-limit and concurrency implementation gaps**: Fixed three issues in the initial rate-limit implementation
+- **e2e test fixture**: Added pytest-playwright dependency
+
+## [0.7.0] - 2026-02-14
+
+### Added
+
+- **Extra metadata with DataCite vocabulary**: User-managed `extra_metadata.json` per video for supplemental metadata (related slides, papers, code repos). LinkML schema defines `RelatedResource` using DataCite v4.6 relationType and resourceTypeGeneral vocabularies. Merged additively into `metadata.json` during export (never overwrites archiver-managed fields)
+- **Related resources display**: Frontend shows linked resources in video details with type badges (`[Presentation slides]`, `[Software]`, etc.)
+- **Human-readable caption labels**: Caption language selector shows readable labels for yt-dlp variant suffixes (e.g. `en-cur1` → `EN (curated)`, `en-orig` → `EN (original)`)
+- **Caption download button**: Download VTT caption files directly from the transcript panel header
+- **Deno as core dependency**: Required for yt-dlp YouTube JS challenge solver
+- **Skip network tests by default**: pytest addopts `-m "not network"` prevents flaky CI from YouTube bot detection. Use `tox -e network` for full network test sweep
+
+### Fixed
+
+- **Phased backup_channel() flow**: Refactored into discovery → fetch → store → link phases, fixing O(P^2) TSV regeneration and symlink overhead
+- **Version injection after appVersion refactor**: Fixed placeholder matching in `deploy_frontend()` after Vite inlining changed from `v0.0.0-unknown` to bare `0.0.0-unknown`
+- **TypeError on None playlist titles**: Handle yt-dlp returning None for playlist titles
+- **Empty videos.tsv crash**: Handle missing `videos/` directory gracefully
+- **Frontend build warnings**: A11y labels, version variable reference, video caption attributes
+- **Encoding consistency**: Use `encoding="utf-8"` instead of `text=True` in all subprocess calls
+- **yt-dlp options consistency**: Use `_get_ydl_opts()` for all yt-dlp calls, apply env vars in UserConfig
+
+### Changed
+
+- **Header renamed to AnnexTube**: Links to GitHub repo https://github.com/con/annextube
+- **ejs:github enabled by default**: Remote JavaScript components for yt-dlp challenge solver
+- **Test infrastructure**: pytest tmp_path fixture, datalad in test deps, playwright skip when missing, dedicated network test environment
+
+## [0.6.0] - 2026-02-14
+
+### Added
+
+- **URL permalink for caption search state**: Shared links now also restore caption search context — query, case-sensitive/regex/filter toggles, and match position. URL params `q`, `cs`, `re`, `filter`, `match` are appended alongside existing player params (e.g. `#/channel/X/video/Y?tab=local&t=90&lang=en&q=neuroimaging&cs=1&re=1&filter=1&match=3`). Match position is restored on first load, then resets normally on new searches.
+
+## [0.5.0] - 2026-02-13
+
+### Added
+
+- **URL permalink for video player state**: Shared links now restore the full viewing context — active player tab (archive/youtube), wide mode, transcript visibility, playback position, and caption language. URL params are appended to the hash route (e.g. `#/channel/X/video/Y?tab=youtube&wide=1&t=90&transcript=0&lang=es`). Only non-default values are written, keeping URLs clean. Uses `history.replaceState` to avoid history pollution.
+
+## [0.4.0] - 2026-02-13
+
+### Added
+
+- **Interactive caption/transcript browser**: Side-by-side panel next to video player with full transcript display, click-to-seek, and active cue highlighting synced to playback position
+- **Transcript search**: Search within captions with case-sensitive (C), regex (.*), and filter (F) modes. Filter mode hides non-matching cues; dim mode preserves timeline context. Enter/Shift+Enter keyboard navigation between matches with N/M position indicator
+- **WebVTT parser with auto-caption dedup**: Zero-dependency parser handles YouTube auto-caption rolling 3-cue pattern (display→snapshot→carry-over), merging overlapping cues and stripping carry-over lines. Reduces a 1-hour video from ~4000 raw cues to ~1400 clean cues
+- **Wide/theater mode**: Toggle (button or `t` key) to use full browser width for video+transcript layout, persisted in localStorage
+- **Language selector**: Switch between available caption languages in the transcript panel
+- **Auto-scroll with manual override**: Transcript auto-scrolls to active cue during playback, pauses on manual scroll, re-enables after 5s idle or cue click. "Resume auto-scroll" button when disabled
+- **GitHub Pages sharing**: `annextube unannex` command and `prepare-ghpages` workflow for sharing archives via GitHub Pages
+- **Frontend download button**: Download local video files directly from the player
+- **YouTube external links**: "View on YouTube" button with source URL
+
+### Fixed
+
+- **Per-playlist videos.tsv**: Use `os.walk` to follow directory symlinks when indexing playlist videos
+- **Playlist symlink indexing**: Fixed per-playlist videos.tsv generation
+- **DataLoader archive root discovery**: Probe ascending paths at runtime instead of assuming fixed `../` relative path, enabling deployment at any URL depth
+- **Click-to-seek scroll glitch**: Set activeCueIndex immediately on cue click to prevent afterUpdate scrolling to the old active cue while the async seek completes
+- **Search regex statefulness**: Removed `g` flag from shared searchRegex — `RegExp.test()` with `g` advances `lastIndex`, breaking match detection in loops across different strings
+- **Wide→normal height revert**: CSS Grid `align-items: start` breaks the circular dependency where `--player-height` was measured from the grid-row height inflated by the caption browser
+- **Search nav button clipping**: Redesigned as two-row layout — search options on row 1, navigation (Filter + Prev/Next) on conditional row 2
+- **Filter→unfilter scroll position**: Scroll to current match when exiting filter mode instead of jumping to top
+- **Caption file path**: Fixed `getCaptionPath` in DataLoader to use `video.{lang}.vtt` naming (was `caption_{lang}.vtt`)
+- **Missing captions in transcript browser**: Reconcile `captions_available` in metadata.json with actual VTT files on disk during export. Fixes videos where captions were downloaded but metadata wasn't re-saved
+- **YouTube customUrl field**: Strip leading `@` from API response
+
+### Changed
+
+- **Simplified unannex**: Replaced 632-line implementation with thin git-annex wrapper
+- **Version injection**: Use `0.0.0-unknown` placeholder in frontend build, replaced at deploy time by `deploy_frontend()`
+- **Video+transcript layout**: CSS Grid with `align-items: start` and dynamic `--player-height` variable ensures transcript panel matches video height exactly
+
+## [0.3.0] - 2026-02-11
+
+### Added
+
+- **Batch YouTube API calls**: Pre-fetch statistics and metadata in bulk instead of per-video API calls
+- **`--comments-depth` CLI option**: Override comments depth at backup time (-1=unlimited, 0=disabled, N=limit)
+- **DataLad dataset support**: `--datalad` flag for `init` command to create DataLad-managed archives
+- **Version display in web UI**: Show annextube version in the header, injected during `generate-web`
+- **Channel permalink routing**: Direct URLs to channels and per-channel playlist support in web UI
+- **TSV regeneration for multi-channel collections**: `--update=tsv_metadata` works across collections
+- **`serve --regenerate=web`**: Allow creating the web directory if it doesn't exist yet
+
+### Fixed
+
+- **Incremental backup discarding new videos**: Social date window filter was incorrectly applied to new video discovery in incremental mode, silently dropping videos published outside the 1-week window
+- **Frontend URL state feedback loop**: Changing any filter caused a hashchange → restoreFromURL → re-render cycle, visibly resetting selections. Fixed with `history.replaceState()` instead of `window.location.hash`
+- **URL defaults pollution**: Default sort params (`sort=date&dir=desc`) no longer appended to clean URLs
+- **Video player paths at subpath deployments**: Absolute `/videos/...` paths replaced with relative `../videos/...` for correct resolution at any deployment subpath
+- **Video player reactivity**: Fixed state not updating when video prop changes
+- **Video availability checking**: Use HEAD requests instead of relying on TSV download_status; include 'tracked' status; centralized checking logic
+- **VideoPlayer thumbnail fallback**: Use YouTube CDN thumbnail when video not downloaded locally
+- **Video metadata/comments loading**: Use file_path from video object instead of reconstructing paths
+- **Browser back/forward navigation**: Fixed multi-channel video metadata loading on history navigation
+- **Page reloads and playlist selection**: Fixed routing to preserve context and support nested routes
+- **TSV field escaping**: Properly escape newlines and special characters in TSV output
+- **Unicode encoding errors**: Fixed encoding issues in channel avatar display
+- **Metadata sync**: Sync metadata.json download_status with actual file system state during TSV generation
+
+### Changed
+
+- **Asset filenames**: Vite build now produces clean filenames (`index.js`, `index.css`) without content hashes
+- **FilterPanel initialization**: Uses Svelte `tick()` for clean reactive cycle during mount
+
+## [0.2.2] - 2026-02-09
+
+### Added
+
+- **Multi-channel collections**: Hierarchical structure with channels.tsv, aggregate/export/generate-web/serve commands
+- **Auto-generate channel.json**: Automatically created during backup for seamless multi-channel integration
+- **YouTube API channel metadata**: Three-tier extraction (API → yt-dlp → archive) for complete channel information
+- **Channel avatar download**: Automatic download with MIME type detection and git-annex tracking
+- **Timestamp filtering for channel.json**: Prevents unnecessary commits when only timestamps change
+- **Quota exceeded auto-retry**: Automatic wait and retry when YouTube API quota is exceeded
+- **Periodic checkpoint commits**: Configurable intermediate commits during long backup operations with interruption recovery
+- **Configurable playlist symlink patterns**: Use {video_id}, {video_title}, {position} in playlist symlinks
+- **Archive discovery helpers**: Central discovery system for single-channel vs multi-channel detection
+
+### Fixed
+
+- **Playlist video reprocessing**: Skip reprocessing existing videos in 'playlists' update mode
+- **Caption download for existing videos**: Properly download captions when adding to existing archive
+- **Incremental playlist optimization**: Skip unavailable videos in incremental mode for faster processing
+- **Path pattern validation**: Fail early on unknown placeholders in video_path_pattern
+- **E2E test environment**: Use 'uv run annextube' to ensure correct environment activation
+- **Type checking**: Complete mypy compliance across all modules
+- **Linting**: Full ruff compliance across all modules
+
+### Changed
+
+- **Default include-playlists**: Changed from 'none' to 'all' for better discoverability
+- **Default include-podcasts**: Changed from 'none' to 'all' for consistency
+- **Default playlist path pattern**: Changed from {playlist_id} to {playlist_title} for readability
+- **Config template formatting**: Simplified double-brace escaping for better readability
+
+### Documentation
+
+- **Quota handling guide**: Complete documentation for YouTube API quota management
+- **Checkpoint system guide**: Documentation for periodic commits and recovery
+- **Multi-channel collections**: End-to-end guide for creating and managing collections
+- **yt-dlp challenge solver**: Documented dependency issues and solutions
+
+## [0.2.0] - 2026-02-05
+
+### Added
+
+- **Incremental social data updates**: Efficient API-optimized updates with statistics gate and early stopping (96% API request reduction)
+- **Containerization support**: Complete Podman/Docker/Singularity recipes with documentation
+- **Progress indicators**: Real-time feedback for long-running playlist/channel fetches
+- **Video file validation**: Magic byte detection to identify HTML error pages disguised as videos
+- **YouTube Data API integration**: Enhanced metadata and comment fetching with yt-dlp fallback
+- **Per-source component configuration**: Override global settings per channel/playlist
+- **User-wide configuration**: Cookies, proxy, network settings in ~/.config/annextube/config.toml
+- **Hierarchical video structure**: Year/month subdirectories for better organization
+- **Range server with annextube serve**: Local playback with seeking support
+- **Comprehensive E2E test suite**: Playwright tests for web UI and complete workflows
+- **Full web frontend**: Search, filter, sort, video player with local/YouTube tabs
+
+### Fixed
+
+- **Nested directory bug**: git-annex addurl with relative paths creating nested structures
+- **EJS solver integration**: Support for authenticated access with cookies + deno
+- **Playlist creation bug**: Ensure playlists are created with correct video associations
+- **Download status detection**: Proper differentiation between tracked and downloaded videos
+- **Video player issues**: Blank screen, iframe sizing, seeking with range requests
+- **git-annex addurl**: Remove --no-raw flag to avoid conflicts with --fast
+- **Cookie path handling**: Remove quotes from git-annex cookie configuration
+- **Code quality**: Complete mypy and ruff linting fixes
+
+### Changed
+
+- **Test infrastructure**: Dedicated AnnexTube test channel with 10 videos and 5 playlists
+- **E2E browser setup**: Avoid sudo prompts by checking existing installations
+- **Default update mode**: Changed to all-incremental (videos + social data updates)
+
+### Documentation
+
+- Complete frontend MVP documentation with phase-by-phase progress
+- Container deployment guide for Podman/Docker/Singularity
+- User configuration documentation with cookie and network settings examples
+
+## [0.1.0] - 2026-01-28
+
+Initial release with core YouTube archival functionality:
+- Channel and playlist backup
+- git-annex integration for efficient storage
+- Metadata extraction (JSON + TSV)
+- Caption and thumbnail downloads
+- Comment fetching
+- Basic web UI for browsing archives
+
+[0.12.0]: https://github.com/con/annextube/compare/v0.11.0...v0.12.0
+[0.11.0]: https://github.com/con/annextube/compare/v0.10.0...v0.11.0
+[0.10.0]: https://github.com/con/annextube/compare/v0.9.0...v0.10.0
+[0.9.0]: https://github.com/con/annextube/compare/v0.8.0...v0.9.0
+[0.8.0]: https://github.com/con/annextube/compare/v0.7.0...v0.8.0
+[0.7.0]: https://github.com/con/annextube/compare/v0.6.0...v0.7.0
+[0.6.0]: https://github.com/con/annextube/compare/v0.5.0...v0.6.0
+[0.5.0]: https://github.com/con/annextube/compare/v0.4.0...v0.5.0
+[0.4.0]: https://github.com/con/annextube/compare/v0.3.0...v0.4.0
+[0.3.0]: https://github.com/con/annextube/compare/v0.2.2...v0.3.0
+[0.2.2]: https://github.com/con/annextube/compare/v0.2.0...v0.2.2
+[0.2.0]: https://github.com/con/annextube/compare/v0.1.0...v0.2.0
+[0.1.0]: https://github.com/con/annextube/releases/tag/v0.1.0