Skip to content
UniPDF
Guides
Extraction
Text Extraction

Text Extraction

Text extraction goes through the extractor package, one page at a time:

page, err := pdfReader.GetPage(pageNum)
if err != nil {
    return err
}

ex, err := extractor.New(page)
if err != nil {
    return err
}

text, err := ex.ExtractText()
if err != nil {
    return err
}

extractor.New uses the default mode. When the output is not what you expect, the mode is usually the thing to change.

Choosing a mode

Pass a mode through NewWithOptions instead:

ex, err := extractor.NewWithOptions(page, &extractor.Options{
    ExtractionMode: extractor.ExtractionModePlain,
})
ModeWhat it doesUse it when
ExtractionModeLayout (default)Detects paragraphs and tables, works out reading order. Supports Tables() and Marks().You need marks or tables, for instance search and replace.
ExtractionModePlainReads straight from the content streams, no paragraph grouping or table detection. Fastest, and produces the most complete text with no duplication.Bulk text extraction, and data-heavy PDFs.
ExtractionModeLayoutNoBreaksAs Layout, but without line breaks inside the same horizontal line.The page is one column and layout mode is breaking lines you want joined.
ExtractionModeGridPlaces words on a fixed-width character grid from their positions, keeping multi-column text and table columns aligned. Comparable to pdftotext -layout.You want the visual column structure preserved as text.

In ExtractionModeGrid, Marks() and Tables() return empty results in this mode, so use ExtractionModeLayout if you need either. And the grid layout is available from any mode through PageText.GridText(), so you do not have to switch modes just to get it once.

Layout mode is the one to reach for if you are unsure, and plain mode is the one to try first when extraction looks wrong.

Other options

extractor.Options carries a few more switches worth knowing:

  • IncludeAnnotations pulls annotation text into the output. Off by default.
  • ApplyCropBox limits extraction to the page’s crop box.
  • DisableDehyphenation keeps words broken across lines separate. By default a word ending in a hyphen is rejoined with the next line.
  • RelaxedMode tries to repair invalid parameter lengths and values rather than failing, which helps with malformed files.
  • DisableDocumentTags ignores document tags during list extraction.

Run the example

outputPdfText opens the file, loops over the pages, and prints the text of each one to the terminal using the default mode.

git clone https://github.com/unidoc/unipdf-examples.git
cd unipdf-examples/extract
go run pdf_extract_text.go input.pdf

If this is your first time using UniPDF, follow the getting started guide to create an API key and set up your development environment.

View the full source
Last updated on
Table Extraction