Fetching metadata from the HF Docker repository... 7 days ago
-
feat(layout): add layout grid tokens and breakpoint aliases
tfrere HF Staff
-
feat: improve PDF export with screenshot-based embed rendering and cover page layout
tfrere HF Staff - Replace dynamic D3/Plotly embeds with rasterized screenshots in PDF export,
preventing SVG scaling issues when viewport changes for print
- Redesign PDF Table of Contents with numbered sections and clean typography
- Add dedicated cover page (hero + meta) with page break before TOC
- Refactor export-pdf.mjs: extract makeMediaResponsive(), remove duplicated
fixSvg code (~135 lines removed)
- Add screenshot-elements.mjs utility for element-level screenshots
- Update components (Hero, HtmlEmbed, Stack, TableOfContents, ThemeToggle)
- Clean up unused data files and legacy remark plugin
- Add AUDIT-2026 and ROADMAP-2026 docs
Made-with: Cursor
-
feat: backport dataviz page, embed export, and visual extraction from smol-training-playbook
tfrere +1: Cursor HF Staff Major changes:
- Overhaul dataviz page with vertical layout, multi-visual support (charts, images, stacks, tables), metadata display, and per-item export
- Add robust visual extraction (extract-embeds.mjs): Stack grouping, stripCodeBlocks to avoid false positives, image import resolution
- Add HtmlEmbed download functionality (PNG export, batch ZIP export)
- Remove Trackio component (moved to smol-training-playbook only)
- Remove unused banner variants and old embeds
- Add AGENTS.md and CLAUDE.md for AI agent context
- Update styles, footer, SEO, and PDF export scripts
Co-authored-by: Cursor <cursoragent@cursor.com>
-
Clean repository - remove missing LFS files
tfrere HF Staff
