Skip to content

pdf_utils

langroid/parsing/pdf_utils.py

pdf_split_pages(input_pdf)

Splits a PDF into individual pages in a temporary directory.

Parameters:

Name Type Description Default
input_pdf Union[BytesIO, BinaryIO]

Input PDF file in bytes or binary mode

 required
max_workers

Maximum number of concurrent workers for parallel processing

 required

Returns:

Type Description
Tuple[List[Path], TemporaryDirectory[Any]]

Tuple containing: - List of paths to individual PDF pages - Temporary directory object (caller must call cleanup())
Example

paths, tmp_dir = split_pdf_temp("input.pdf")

Use paths...

tmp_dir.cleanup() # Clean up temp files when done

Source code in langroid/parsing/pdf_utils.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
def pdf_split_pages(
    input_pdf: Union[BytesIO, BinaryIO],
) -> Tuple[List[Path], TemporaryDirectory[Any]]:
    """Splits a PDF into individual pages in a temporary directory.

    Args:
        input_pdf: Input PDF file in bytes or binary mode
        max_workers: Maximum number of concurrent workers for parallel processing

    Returns:
        Tuple containing:
            - List of paths to individual PDF pages
            - Temporary directory object (caller must call cleanup())

    Example:
        paths, tmp_dir = split_pdf_temp("input.pdf")
        # Use paths...
        tmp_dir.cleanup()  # Clean up temp files when done
    """
    tmp_dir = tempfile.TemporaryDirectory()
    doc = fitz.open(stream=input_pdf, filetype="pdf")
    paths = []

    for page_num in range(len(doc)):
        new_doc = fitz.open()
        new_doc.insert_pdf(doc, from_page=page_num, to_page=page_num)
        output = Path(tmp_dir.name) / f"page_{page_num + 1}.pdf"
        new_doc.save(str(output))
        new_doc.close()
        paths.append(output)

    doc.close()
    return paths, tmp_dir