Feature: PDF structure parser and metadata extraction

[coderabbitai]

Overview

Implement PDF parser to extract document structure, metadata, and text while skipping binary image data.

Parent Epic

Part of #91 - Document & Office Format Awareness

Description

Parse PDF structure (objects, streams, cross-reference tables) and extract meaningful strings from metadata, annotations, bookmarks, and text streams.

Implementation Details

• Use lopdf or pdf crate
• Parse PDF object structure
• Extract document info dictionary (Title, Author, Subject, Keywords)
• Parse catalog and page tree
• Extract text from content streams
• Identify and skip image streams
• Parse annotations and form fields
• Extract JavaScript from actions

String Sources

• Document metadata (Title, Author, Subject, Keywords, Creator, Producer)
• Bookmark titles
• Annotation text
• Form field names and values
• Font names
• JavaScript code
• Hyperlink URLs
• Named destinations

Acceptance Criteria

• Parse PDF structure (v1.4-1.7)
• Extract all metadata dictionary entries
• Parse page content streams for text
• Skip binary image streams (entropy-based)
• Extract annotations and bookmarks
• Handle encrypted PDFs (metadata only)
• Tests with diverse PDF samples

Test Cases

• Simple text PDFs
• PDFs with images
• PDFs with forms
• PDFs with JavaScript
• Encrypted PDFs
• Large PDFs (>100MB)

Related

Project: #76

[traycerai]

Warning

Rate limit exceeded. Please try after 1188 seconds.