feat: require Python 3.14 (STL release) + unconditional video + EPUB scoping#2138
Open
skilfoy wants to merge 8 commits into
Open
feat: require Python 3.14 (STL release) + unconditional video + EPUB scoping#2138skilfoy wants to merge 8 commits into
skilfoy wants to merge 8 commits into
Conversation
Add first-class PowerPoint ingestion that preserves slide structure, notes, charts, diagrams, images, audio/video, and attachments without python-pptx. Hardens converter caches (limits + artifact integrity), transcript cache keys (content/model/prompt), incremental liveness, and agent transcription workflows. Closes the gap tracked by Graphify-Labs#1573 and supersedes the thinner text-only approach in Graphify-Labs#1821.
- Bumps requires-python to >=3.14 - Makes [video] extra unconditional (faster-whisper always available) - Removes all python_version markers (tomli, leiden, svg, video) - Updates Ruff target-version and Pyright pythonVersion - Collapses CI matrix to ["3.14"] - Cleans legacy 3.10/3.11 comments Unconditional local video transcription for graph knowledge bases. Running clean on 3.14.3. STL release, baby.
…UB scoping doc - Classifiers for 3.14 - Updated table + leiden description for STL release - docs/epub-scoping.md (full semantic, safety-first, separate-PR plan)
- New graphify/epub.py with safe ZIP + defusedxml OPF/spine parsing - Chapter text extraction, crude media discovery (images/audio/video) - Manifest + sidecar output - Wired into detect.py (special case like PPTX, .epub in OFFICE_EXTENSIONS) - Unconditional video support benefits apply (3.14+) - Smoke-tested import and basic structure Follow-up will add more robust XHTML parsing, better media href resolution, adversarial tests, and full provenance matching PPTX quality. Part of STL 3.14 work + EPUB semantic extraction.
- Use stdlib HTMLParser for better text/title/media extraction from XHTML - Proper relative href resolution - OPF metadata (title, creator, language) in manifest - Per-chapter titles + richer Markdown - Improved manifest with chapter details and source_part - Expanded tests with multiple synthetic EPUB fixtures (order, titles, media routing, errors) - Added EPUB cleanup call in detect.py - All targeted tests (epub + presentation) green on 3.14.3 Bumped converter to v2.
- Bump defaults: claude-sonnet-5, gpt-5.6-mini, gemini-3.5-flash (and equivalents) - Raise default _MAX_IMAGE_BYTES to 32MB, add GRAPHIFY_MAX_IMAGE_MB - Make _MAX_IMAGES_PER_CHUNK auto-configured by backend + token_budget (env override only) - Update README (env table, vision section, recommendations) - Refresh skill.md and test compatibility - Clean stale 4.x/4o references This aligns skill mode (Hermes etc.) and headless with current models and better vision handling for images/PPTX/EPUB.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
[video]extra unconditional (faster-whisperno longer gated behindpython_version >= '3.11')python_versionconditionals and legacy 3.10/3.11/3.13 shimsdocs/epub-scoping.mdwith full plan for dedicated semantic EPUB ingestion (separate follow-up PR)Full test suite on 3.14.3: 3449 passed, 165 skipped, 6 unrelated pre-existing failures (ollama + install_references tests).
Targeted PPTX/transcribe/extract tests: all green on 3.14.3.
Motivation
Python 3.10 reaches EOL in October 2026. This gets us to a modern baseline with unconditional high-quality local video transcription — critical for treating video archives as first-class citizens in the knowledge graph.
Changes
requires-python = ">=3.14"allextra cleanedTesting
This is a breaking change for anyone still on <3.14.
Next: dedicated EPUB semantic ingestion PR (scoping doc included here).