Released: Aug 17, 2025

8.2.0 での更新項目

機能

• AI-Powered PDF Processing - This release introduces a powerful new package, DsPdfAI, designed to showcase how Document Solutions for PDF can integrate with AI services to enhance PDF document workflows. This feature allows developers to leverage OpenAI or Azure OpenAI to generate summaries, create outline trees, and extract structured tables data directly from PDFs.
  • Supported Scenarios - Currently, DsPdfAI supports the following three AI-driven PDF processing capabilities:
    • Generate a Document Summary - Quickly produce short or long summaries of any PDF document.
    • Programmatically Generate a PDF Document Summary using DsPdfAI in C# or VB.
    • Build a Document Outline Tree - Automatically generate a structured table of contents, complete with nested outline levels.
    • Automatically generate a structured table of contents using AI capabilities included in a .NET PDF API.
    • Extract Tables from a Document - Use natural language prompts to identify and extract tabular data from a PDF.
    • Extract Tables from a PDF using PDF API AI Capabilities.
    • These features are built on top of the OpenAI Chat API, with options to connect either via the official OpenAI .NET client library or the Azure OpenAI service.
  • API Overview - The new package introduces two main classes, both inheriting from a shared base class for core functionality:
    • OpenAIDocumentAssistant - Connects to the OpenAI REST API using the official .NET client library.
    • AzureOpenAIDocumentAssistant - Connects to Azure OpenAI services using the Azure.AI.OpenAI package.
    • Both derive from OpenAIDocumentAssistantBase, which implements the logic for extracting text via Page.GetText() and sending content to the AI service for processing. The pageRange parameter allows developers to limit requests to specific sections of a PDF.
  • Platforms Supported:
    • OpenAI REST API (via the official .NET client).
    • Azure OpenAI Service (via the Azure.AI.OpenAI package).

Document Solutions PDF Viewer

• DsPdfViewer.wasm Now Supports Export to SVG - In this release, the WebAssembly implementation of DsPdfViewer now supports saving and exporting a loaded-in PDF file to SVG format, which will be included within a .zip file after exporting to SVG format.
• Redesigned Toolbar and Navigation for Better Usability - With this release, DsPdfViewer introduces a redesigned toolbar and navigation system to improve usability, consistency, and customization across viewing and editing modes. These changes ensure that the most frequently used actions are easier to access, while still allowing backward compatibility for users who prefer the legacy layout.
• Toolbar Layout Improvements - The toolbar has been restructured to prioritize commonly used actions and streamline the interface. If the old layout/behavior is desired, it can be restored by following the commands shown in the CHANGELOG.md file included with the release.
  • New Default Order
    • View mode: Open → Save → Print → Search (moved Open to the front for faster access).
    • Edit mode: Added Open before Save.
  • Removed from Default Toolbar
    • Full Screen.
    • Themes.
    • Extract Table (moved to the side panel).
  • Form Filler Button
    • Now appears only in the mobile toolbar by default.
    • Optimized for smaller screens like tablets and phones, where form-filling is common.
• Side Panel Enhancements - The side panel has been updated for a more intuitive workflow, with a new default order and icon updates.
  • New Default Order:
    • Document List → Shared Documents → Search → Thumbnails → Outline → Structure Tree → Articles → Layers → Attachments → Extract Table.
    • Updated Extract Table Icon - Now displays a table with a right-pointing arrow for clarity.
• Navigation Controls & Keyboard Support - Navigation is now more compact, customizable, and accessible.
  • Visual Updates
    • Compact page navigation input.
    • First/Last page buttons hidden by default.
    • Improved page counter display logic.
  • Keyboard Shortcuts (when page input is focused):
    • ← / → : Navigate pages (context-aware).
    • ↑ / ↓ : Force page navigation.
    • PageUp / PageDown : Jump to previous/next page.
    • Home / End : Jump to first/last page.
    • Enter : Submit manual entry.
    • Escape : Cancel input.
    • Ctrl + Click on page arrows : Jump to first/last page.
  • New Semantic CSS Classes:
    • .gc-btn.go-to-first.
    • .gc-btn.go-to-last.
    • .current-page-input.
    • .gcv-page-input__text.
• Overlapping Annotations Handling (Editor Mode) - When multiple annotations overlap, users now get clearer and more precise control:
  • Primary Context Menu shows actions for the topmost annotation.
  • Additional Annotations Section lists other overlapping annotations below a separator.
  • Clicking any listed annotation switches focus for editing.
• Context Menu & UI Standardization - To ensure consistency and professionalism, the interface now follows standardized formatting for all text elements:
  • Buttons / Menu Items / Headers: Title Case.
  • Tooltips: Sentence case.
  • Labels: Title Case.
  • Descriptions / Status Text: Sentence case.
  • Placeholders: Sentence case.
  • Technical Terms: Original case (e.g., PDF, URL, iOS).
  • Dynamic Values:
    • In Actions: Title Case → "Add {{count}} Files".
    • In Descriptions: Original case → "{{count}} files processed".
• API Enhancements - Developers can now customize navigation more precisely with new configuration options.

不具合の修正

• Fixed an issue where GcPdfDocument.Page.GetText() and GcPdfDocument.Page.GetTextMap().GetText() could return different text when GcPdfDocument.RecognitionAlgorithm was Advanced (the default).