Region-Based Extraction using Python

Extract text from a specific region of a page

Use TextAbsorber together with a Rectangle to limit extraction to a specific area of a page. This approach is useful for zone-based extraction from headers, footers, table cells, form fields, invoices, or other fixed-layout regions where the text position is known in advance.

1. Open the source PDF as a Document.
2. Create a TextAbsorber instance.
3. Configure text_search_options to limit extraction to a rectangle.
4. Accept the absorber on the target page.
5. Write the extracted text to an output file.

import aspose.pdf as ap


def extract_text_from_region(infile, page_number, rect_coords, outfile):
    """
    Extract text from a specified rectangular region on a given page.
    Args:
        infile (str): Path to input PDF file.
        page_number (int): 1-based index of the page.
        rect_coords (tuple): (llx, lly, urx, ury) coordinates of the rectangle.
        outfile (str): Output text file path.
    """
    document = ap.Document(infile)
    try:
        absorber = ap.text.TextAbsorber()
        # Set options to restrict search to the rectangle
        absorber.text_search_options.limit_to_page_bounds = True
        llx, lly, urx, ury = rect_coords
        absorber.text_search_options.rectangle = ap.Rectangle(llx, lly, urx, ury, True)
        # Accept on the specific page
        document.pages[page_number].accept(absorber)
        extracted_text = absorber.text
        with open(outfile, "w", encoding="utf-8") as tw:
            tw.write(extracted_text)
    finally:
        document.close()

Extract Paragraphs with bounding polygon rendering

You can also use ParagraphAbsorber to inspect paragraph geometry. In addition to extracting text, this approach records each section rectangle and paragraph polygon, which is useful for layout mapping, document analysis, accessibility tooling, or region-aware post-processing.

1. Open the source PDF as a Document.
2. Create a ParagraphAbsorber instance.
3. Visit the target page.
4. Read the page markup from absorber.page_markups.
5. Iterate through sections and paragraphs to capture geometry and text.
6. Write the rectangle, polygon, and text data to the output file.

import aspose.pdf as ap


def extract_paragraphs_with_geometry(infile, outfile):
    """
    Extract paragraphs and record geometry info (rectangle / polygon) for each paragraph in a PDF.
    Args:
        infile (str): Path to input PDF file.
        outfile (str): Path to output text file.
    """
    document = ap.Document(infile)
    try:
        absorber = ap.text.ParagraphAbsorber()
        absorber.visit(document.pages[1])  # Visit page 2 (pages are 1-indexed)

        page_markup = absorber.page_markups[0]
        with open(outfile, "w", encoding="utf-8") as tw:
            for sec_idx, section in enumerate(page_markup.sections, start=1):
                tw.write(f"Section {sec_idx}: rectangle = {section.rectangle}\n")
                for para_idx, paragraph in enumerate(section.paragraphs, start=1):
                    tw.write(f"  Paragraph {para_idx}: polygon = {paragraph.points}\n")
                    # Concatenate paragraph text
                    parts = []
                    for line in paragraph.lines:
                        for fragment in line:
                            parts.append(fragment.text)
                        parts.append("\r\n")
                    tw.write("    Text: " + "".join(parts) + "\n\n")
    finally:
        document.close()