Documents & Forms

## 1.2.7 — 2026-09-09
- Rebuilt the public **Document List → Cards** layout to match the approved Documents & Forms resource-library mockup: hero/header, search and category/file-type/access filters, sorting, grid/list controls, responsive four-column cards, pagination, and per-page controls.
- Added permission-aware PDF first-page and image thumbnails. Thumbnails are generated lazily and cached as protected derivatives; Office and other formats receive recognizable styled file previews.
- Added separate **View / Open in New Tab** and **Download** actions. Public document open actions use a new tab; executable PHP/HTML uploads remain blocked.
- Added download counters for explicit downloads and surfaced them in the public cards.
- Added safe PPT/PPTX upload recognition so PowerPoint resources can participate in the library UI.
- Reprocessing now replaces ordinary un-published normalized/thumbnail derivatives instead of endlessly accumulating duplicates.
- Preserved full-text search, the fuzzy fallback, metadata suggestions, direct SEO-friendly URLs, permissions, and protected storage.

## 1.2.6 — 2026-09-09
- Removed the HTML document-view wrapper from the canonical public experience; documents now open directly in the browser native PDF/media viewer.
- Changed canonical friendly URLs to `document-download.php/<public-id>/<title-slug>.<ext>` so the final URL segment is a meaningful title-derived filename instead of the opaque public id.
- Added title-derived `Content-Disposition` filenames (ASCII + UTF-8) for inline viewing and downloads, improving browser tab/viewer filename fallback while leaving original protected files unchanged.
- Kept backward compatibility for 1.2.3–1.2.5 document URLs: old `document.php?doc=...` links redirect to the new direct canonical URL and older friendly download paths remain resolvable.
- Retained the conservative fuzzy-search fallback unchanged.

# Changelog

## 1.2.5 — Workspace Fuzzy Search & Shared-Host URL Compatibility
- Added typo-tolerant fallback matching to the Documents workspace search itself; 1.2.4 fuzzy matching had only been wired into the Core Site Search provider.
- Multi-word searches ignore short connector words for fuzzy comparison and allow conservative one/two-edit matching, so `compensation for srvices rendered` can match `compensation for services rendered`.
- Changed the canonical document-view URL to shared-host-safe `document.php?doc=<title-slug>--<public-id>` form, avoiding 404s on Apache configurations with PATH_INFO disabled while retaining a readable title slug.
- Kept legacy `?id=` and 1.2.3/1.2.4 PATH_INFO parsing for backward compatibility.
- Protected raw-download URLs and all 1.2.4 metadata/title/description/document-card improvements remain intact.
- No Core update required.

## 1.2.4 — Metadata Suggestions, Document View & Search Polish
- Fixed fuzzy multi-word phrase candidate selection so a one-character typo in a term can still find a document when other correctly-spelled terms identify the candidate.
- After extraction, blank/default Title and blank Description fields are conservatively populated from meaningful early document text. User-entered metadata is never overwritten and remains fully editable in the normal document detail form.
- Added a permission-aware SEO document view page with readable slug URLs, document-specific browser-tab title, meta description/canonical URL, protected inline PDF/image/audio/video preview where supported, and explicit Open/Download actions.
- Preserved the protected raw download endpoint and legacy links.
- Upgraded the existing Document List `cards` layout into a responsive public resource-card grid with visibility badge, title, description, and View / Download action.
- No Core update required.

## 1.2.3 — Search Relevance & Friendly Document URLs
- Added occurrence-aware ranking while preserving field importance: title matches remain strongest, followed by description, filename, and extracted document contents.
- Added restrained typo-tolerant fallback matching with Levenshtein distance only when exact/substring results are sparse. Exact results always outrank fuzzy matches.
- Added SEO/readability-friendly protected document URLs in the form `document-download.php/<title-slug>/<public-id>`.
- Preserved backward compatibility with existing `document-download.php?id=<public-id>` links.
- Search results, document widgets, PDF designer links, and protected download metadata now use the canonical friendly URL helper.
- Includes all 1.2.2 PDF extraction and modal-close fixes.

## 1.2.2 — PDF Extraction & Document Modal Fixes
- Added a pure-PHP PDF text extraction fallback for common Flate/ASCIIHex/ASCII85 compressed text streams when `pdftotext` is unavailable.
- Added common PDF literal/hex string decoding and embedded ToUnicode CMap handling for searchable text.
- Kept OCR out of scope: image-only/scanned PDFs still require OCR and can remain `Text: empty`.
- Updated Documents search placeholder to include document contents.
- Fixed document detail/upload modal closing with a visible sticky close button, backdrop click, and Escape key.
- Includes the 1.2.1 Core Site Search provider and Document List category-filter fixes.

## 1.2.1 — Site Search & Webpage Integration
- Registered Documents with the existing Core Site Search provider API; no Core update is required.
- Site Search can now match published documents by title, description, original filename, and already-extracted `search_text`.
- Search results enforce the module's existing visibility/role/member access checks before returning a document.
- Search matches inside extracted file contents return a nearby text snippet and a protected permission-aware document URL.
- Activated the existing Document List widget's category-slug filter so Builder resource sections can actually be scoped by category.
- Removed file-size display from normal Documents library cards; file type remains visible.
- Kept derivative file sizes only inside the technical Optimize/Convert derivative panel where size is useful for optimization decisions.

## 1.2.0 — Smart Document Analysis
- Added server-assisted classification from locally extracted document text; Core does not require an update.
- Added ranked document-type suggestions with confidence and matched-signal explanations.
- Added manual type override and new-type creation when the server guess is wrong or absent.
- Added **Remember this layout**, enabled by default after confirmation, with structural layout feedback to DivisionDesk Server.
- Added **No — none of the above** / ignore flow so normal uploads remain simple and suggestions never block document work.
- Added optional cross-module actions, including creation of a review-required Finance member-renewal starter template when Finance is installed.
- Added local fallback document-type choices and graceful operation when the knowledge service is unavailable.
- Replacing a source file invalidates stale knowledge analysis automatically.
- Added restore-from-Trash through fetch/AJAX.
- Added migration 003_knowledge.
- All normal admin mutations remain fetch/AJAX with CSRF and permission enforcement.

## 1.1.1
- Removed hard dependency on PHP Fileinfo; uploads now use guarded Fileinfo plus safe signature/extension fallbacks.
- Removed unguarded mbstring calls from upload, document, form, and list paths.
- Improved upload errors so missing optional PHP extensions return JSON errors instead of fatal exceptions.

## 1.1.0
- Added searchable full-text extraction for PDF, TXT, CSV, DOCX and supported Office documents.
- Library search now matches extracted document contents (for example, a search for `dues` can find a PDF containing that word).
- Added visible extraction status, word count, processing errors, and AJAX Re-extract / Reprocess.
- Added server capability detection for pdftotext, Ghostscript, LibreOffice, ImageMagick, and the PDF-form exporter.
- Added non-destructive PDF normalization to a PDF 1.4 working derivative for FPDI workflows.
- Added real AcroForm export from DivisionDesk mapped PDF fields when Python + PyMuPDF is available.
- Added Image Optimize & Convert with server-supported WebP/JPEG/PNG/AVIF output, quality presets, resizing, derivative history, download, and publish selection.
- Published derivatives are served by the protected document download endpoint; originals remain retained.
- Added migration 002_processing for search/process metadata and derivative records.
- All new admin operations use fetch/AJAX and JSON responses with permission and CSRF enforcement.

## 1.0.0
- Initial secure document library, versions, folders/categories, PDF detection, field designer, webforms and protected downloads.