Spaces:
Running
Running
Commit History
-
Add AI slide-to-transcript alignment for untimestamped imagesβ
Claude +1: Claude Opus 4.8When the images-input mode gets a transcript but the slide filenames carry no HH-MM-SS timestamps, place each slide on the timeline by content instead of failing. Locally OCR each slide (Tesseract) and ask Claude when the narration first discusses it, yielding real per-slide start times. Falls back to an even split across the transcript span when there's no API key or alignment fails. - image_input.py: add all_have_timestamps, order_by_name, build_frames_from_starts, build_frames_even_split (clamped, ordered). - llm_extract.py: add align_slides_to_transcript + SlideAlignError (text-based, so it works through Anthropic-compatible proxies). - /jobs/from-images: three-way timeline branch (order-only / timestamp-anchored / AI-aligned) run off the event loop. - Update images.html hint, README to note timestamps are now optional. - Tests for the new helpers and alignment cleanup. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01β¦e646192 -
feat: ANTHROPIC_MODEL env override for the LLM model IDβ
ClaudeDiagnosing a Parley run where every slide came back described as "blank"/"no content" (one hexagon diagram was even called "an instructor standing in front of a plain background") β the slides are content-rich, so the model isn't receiving the images. That's a proxy-side issue (image not forwarded, or claude-haiku-4-5 routed to a non-vision model). Add an ANTHROPIC_MODEL env override so a proxy that names its vision model differently can be pointed at the right one, instead of the model ID being hard-coded. Documented with a "must be vision-capable" note. https://claude.ai/code/session_01EM8jiYxdVFy1w3LeJ9GBJg5125240 -
feat: route MIT Parley keys to the Parley proxy (alongside normal Anthropic keys)β
ClaudeSupport both a normal Anthropic key (sk-ant-β¦) and an MIT Parley key (sk-parley-β¦) in the same API-key field. A Parley key must hit the Parley proxy rather than api.anthropic.com, so _build_client now resolves the base URL by key prefix: sk-parley β PARLEY_BASE_URL (default https://parley.api.mit.edu, env-overridable), everything else β the SDK default. An explicit ANTHROPIC_BASE_URL still routes non-Parley keys to a custom compatible proxy, but never overrides Parley routing. Detection is by prefix so no server config is needed β the user just pastes whichever key they have and it goes to the right endpoint. Tests cover Parley vs Anthropic routing and the env override. https://claude.ai/code/session_01EM8jiYxdVFy1w3LeJ9GBJg897f74e
-
rename: the app is now "The Describatron"β
ClaudeRebrand the user-facing name from "Video Guide Maker" to "The Describatron" β and de-video the copy now that it also builds guides from slide images: * Page titles + headers across index / images / edit pages. * FastAPI app title; published-guide and editor footers. * README (HF Space card title + emoji π¬βπ€ + heading), USER_GUIDE and DATA_HANDLING headings, and the intro prose. * Index eyebrow + sample copy: "lecture video (or slide images) and a transcript" instead of video-only. Infrastructure identifiers left untouched: the GitHub repo (VideoGuideMaker), the HF Space (Video2Guide), the internal package/ logger name (videoguidemaker), and DEPLOY.md's paths β those are real names, not branding. https://claude.ai/code/session_01EM8jiYxdVFy1w3LeJ9GBJg2f32611
-
docs: update README + USER_GUIDE for this session's featuresβ
ClaudeBring the docs in line with the current app: * Images-input mode (/images): build from timestamped slide exports + a transcript, no video. Documented in both the README ("Two ways in") and USER_GUIDE ("The two paths" + intro). * Self-contained edits export + /edit: Save edits (JSON) now embeds the slides, and /edit reopens the full editor from that one file. Added to the outputs / sharing sections. * Scene description: the recommended accessibility path (open paragraph under the slide that doubles as the long-form image description, alt emptied when present). Replaces the removed "key terms" entries in the AI-features table and editor descriptions. * Per-element include toggles + bulk "apply to all sections" bar. * Markdown download alongside HTML/ZIP. * Removed stale content: key terms, WCAG 2.1 β 2.2, deterministic frame extraction note. Env-var table gains MAX_IMAGE_MB / MAX_EDIT_MB. Project layout gains image_input.py, images.html, edit.html, jobs.py. * Points atβ¦637c2d1
-
Update README to reflect current feature setβ
Claude- Scene preview / picker step - LLM enrichment (alt text, key terms, math, section titles) - Review mode (in-browser editor) - Env var config table (MAX_VIDEO_MB, MAX_TRANSCRIPT_MB, ANTHROPIC_API_KEY) - Updated project layout (orchestrator, llm_extract, review template, _doc_styles) - HF Spaces note about not setting X-Frame-Options/CSP at app level - Python version constraint (3.10β3.13) https://claude.ai/code/session_01EGBMur51izertyiu7msWPQ71844d0
-
Add Dockerfile and HF Spaces metadata for free deploymentβ
ClaudeProvides a Docker image compatible with the Hugging Face Spaces free CPU tier: Python 3.11 slim base, ffmpeg + tesseract installed, runs uvicorn on port 7860 as a non-root user (UID 1000) per HF requirements. README frontmatter declares the Space metadata so a git push deploys the app end-to-end.2128d4b -
Build initial VideoGuideMaker MVPβ
ClaudeFastAPI web app that turns an instructional video plus VTT/SRT transcript into a WCAG 2.2 AA-ready HTML study guide. Pipeline: PySceneDetect + OpenCV for scene detection and frame sampling, Haar-cascade face detection to drop talking-head frames, Tesseract OCR for on-screen text, timestamp-based narration alignment, Jinja2 template with proper landmarks, headings, lang, skip link, time elements, and data-alt-draft="true" flags on auto-drafted alt-text. Output is either a self-contained HTML file (base64-inlined images) or a zip bundle (HTML + images/ folder). Includes unit tests for the transcript parser and HTML generator. https://claude.ai/code/session_01Q9yxSTNAMqkjcYLDJAbV3Zb6f361a