load_pdf

Load a PDF file and return an iterator that yields page images in BGR format.

Pages are rendered lazily one at a time to avoid loading all pages into memory at once, preventing OOM errors for large PDFs with hundreds of pages.

Parameters:

Name	Type	Description	Default
pdf_path	str	The path to the PDF file to be loaded.	required
dpi	int	The resolution (dots per inch) for rendering the PDF pages as images. Higher values result in higher resolution images. Defaults to 200.	200

Returns:

Name	Type	Description
PdfPageIterator	PdfPageIterator	An iterator yielding NumPy arrays (BGR format) for each page. Has a total_pages attribute and supports len().

Raises:

Type	Description
FileNotFoundError	If the specified PDF file does not exist.
ValueError	If the file format is not supported. If the file is not a valid PDF.
RuntimeError	If there is an error while processing the PDF file.

Example

from yomitoku.data.functions import load_pdf

pages = load_pdf("example.pdf", dpi=200)
print(len(pages))  # Output: Number of pages in the PDF file
for page_img in pages:
    print(page_img.shape)  # Output: (height, width, channels)

Source code in src/yomitoku/data/functions.py

def load_pdf(pdf_path: str, dpi=200) -> PdfPageIterator:
    """
    Load a PDF file and return an iterator that yields page images in BGR format.

    Pages are rendered lazily one at a time to avoid loading all pages into
    memory at once, preventing OOM errors for large PDFs with hundreds of pages.

    Args:
        pdf_path (str): The path to the PDF file to be loaded.
        dpi (int, optional): The resolution (dots per inch) for rendering the PDF pages
            as images. Higher values result in higher resolution images. Defaults to 200.

    Returns:
        PdfPageIterator: An iterator yielding NumPy arrays (BGR format) for each page.
            Has a `total_pages` attribute and supports `len()`.

    Raises:
        FileNotFoundError: If the specified PDF file does not exist.
        ValueError:
            - If the file format is not supported.
            - If the file is not a valid PDF.
        RuntimeError: If there is an error while processing the PDF file.

    Example:
        ```python
        from yomitoku.data.functions import load_pdf

        pages = load_pdf("example.pdf", dpi=200)
        print(len(pages))  # Output: Number of pages in the PDF file
        for page_img in pages:
            print(page_img.shape)  # Output: (height, width, channels)
        ```
    """
    pdf_path = Path(pdf_path)
    if not pdf_path.exists():
        raise make_error(ErrorCode.IMAGE_FILE_NOT_FOUND)

    ext = pdf_path.suffix[1:].lower()
    if ext not in SUPPORT_INPUT_FORMAT:
        raise make_error(ErrorCode.UNSUPPORTED_IMAGE_FORMAT)

    if ext != "pdf":
        raise make_error(ErrorCode.NOT_PDF_FILE)

    return PdfPageIterator(pdf_path, dpi=dpi)