text-extraction

Text Extraction Example

This example demonstrates how to extract text with positional information from PDF documents using GoPDF.

Features

• Extract text from PDF pages
• Get X, Y coordinates for each text element
• Font information (name and size)
• Text bounding boxes
• Statistics (font usage, page coverage)

Usage

go run main.go <pdf-file>

Example Output

Opening PDF: sample.pdf
PDF has 1 pages

Extracting text from page 1...

Found 25 text elements:

Position details for first 20 elements:
-------------------------------------------------------
[1] Text: "Hello, World!"
    Position: (100.00, 700.00)
    Size: 72.00 x 12.00
    Font: Helvetica, Size: 12.0pt

[2] Text: "This is a sample PDF."
    Position: (100.00, 686.00)
    Size: 108.00 x 12.00
    Font: Helvetica, Size: 12.0pt

...

-------------------------------------------------------
All extracted text:
-------------------------------------------------------
Hello, World!
This is a sample PDF.
...

-------------------------------------------------------

Statistics:
  Total text elements: 25
  Unique fonts: 2
    Helvetica: 20 elements
    Helvetica-Bold: 5 elements
  Text bounding box:
    Bottom-left: (72.00, 100.00)
    Top-right: (540.00, 720.00)
    Dimensions: 468.00 x 620.00 points

Phase 2.5 Implementation

This example demonstrates Phase 2.5 of the GoPDF roadmap: Text Extraction with Positional Information.

What's Implemented:

1. Text Elements - Each piece of text with position data
2. Matrix Transformations - PDF text matrix handling
3. Text State Tracking - Font, size, spacing, leading
4. Content Stream Parsing - Parse PDF operators
5. Text Operators - BT, ET, Tj, TJ, Tm, Td, Tf, etc.
6. FlateDecode Support - Decompress compressed streams

Critical for Table Extraction:

The positional information extracted here (X, Y coordinates) is essential for Phase 2.6 (Table Detection). Table extraction algorithms need to know where text is located to determine:

• Column boundaries (vertical alignment)
• Row boundaries (horizontal alignment)
• Cell grouping (proximity analysis)
• Table regions (spatial clustering)

Implementation Details

Text Extraction Process:

1. Open PDF file with parser.OpenPDF()
2. Create extractor.NewTextExtractor(reader)
3. Call ExtractFromPage(pageNum) to get text elements
4. Process elements with position data

Text Operators Supported:

• BT/ET - Text object delimiters
• Tf - Set font and size
• Tm - Set text matrix
• Td/TD - Move text position
• T* - Move to next line
• Tj - Show text string
• TJ - Show text with positioning
• Tc/Tw/Tz/TL/Ts - Text state parameters

Stream Decoding:

• FlateDecode - zlib decompression (most common)
• Other filters can be added in future phases

Next Steps

After text extraction, the next phase is:

Phase 2.6 - Table Detection:

• Use text positions to detect table regions
• Identify ruling lines (lattice mode)
• Analyze whitespace (stream mode)
• Group text into cells
• Build table structure

Testing

Test with your own PDF files:

# Simple text document
go run main.go document.pdf

# Invoice with tables
go run main.go invoice.pdf

# Financial report
go run main.go report.pdf

Requirements

• Go 1.25+
• PDF file (unencrypted)

Limitations (Phase 2.5)

• Font width estimation is approximate (full font metrics in Phase 3)
• Encryption not supported (Phase 4)
• Some advanced encoders not supported (Phase 4)

Reference

• PDF 1.7 Specification, Section 9.4 (Text Objects)
• PDF 1.7 Specification, Section 9.3 (Text State)
• GoPDF ROADMAP.md, Phase 2.5

License

MIT License - See LICENSE in project root