def clip_page(cls,
page: fitz.Page,
bbox: fitz.Rect = None,
zoom: float = 3.0):
"""Clip page pixmap (without text) according to ``bbox``.
Args:
page (fitz.Page): pdf page to extract.
bbox (fitz.Rect, optional): Target area to clip. Defaults to None, i.e. entire page.
zoom (float, optional): Improve resolution by this rate. Defaults to 3.0.
Returns:
dict: Raw dict of the extracted pixmap.
"""
# hide text before clip the image only
# render Tr: set the text rendering mode
# - 3: neither fill nor stroke the text -> invisible
# read more:
# - https://github.com/pymupdf/PyMuPDF/issues/257
# - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/pdf_reference_archives/PDFReference.pdf
doc = page.parent
for xref in page.get_contents():
stream = doc.xrefStream(xref).replace(b'BT', b'BT 3 Tr') \
.replace(b'Tm', b'Tm 3 Tr') \
.replace(b'Td', b'Td 3 Tr')
doc.updateStream(xref, stream)
# improve resolution
# - https://pymupdf.readthedocs.io/en/latest/faq.html#how-to-increase-image-resolution
# - https://github.com/pymupdf/PyMuPDF/issues/181
bbox = page.rect if bbox is None else bbox & page.rect
image = page.getPixmap(clip=bbox,
matrix=fitz.Matrix(zoom,
zoom)) # type: fitz.Pixmap
return cls.to_raw_dict(image, bbox)
def _hide_page_text(cls, page: fitz.Page):
"""Hide page text before clipping page.
Args:
page (fitz.Page): pdf page to extract.
"""
# render Tr: set the text rendering mode
# - 3: neither fill nor stroke the text -> invisible
# read more:
# - https://github.com/pymupdf/PyMuPDF/issues/257
# - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/pdf_reference_archives/PDFReference.pdf
doc = page.parent
for xref in page.get_contents():
stream = doc.xrefStream(xref).replace(b'BT', b'BT 3 Tr') \
.replace(b'Tm', b'Tm 3 Tr') \
.replace(b'Td', b'Td 3 Tr')
doc.updateStream(xref, stream)
def _hide_page_text(page: fitz.Page):
'''Hide page text before clipping page.'''
# NOTE: text might exist in both content stream and form object stream
# - content stream, i.e. direct page content
# - form object, i.e. contents referenced by this page
xref_list = [
xref for (xref, name, invoker, bbox) in page.get_xobjects()
]
xref_list.extend(page.get_contents())
# render Tr: set the text rendering mode
# - 3: neither fill nor stroke the text -> invisible
# read more:
# - https://github.com/pymupdf/PyMuPDF/issues/257
# - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/pdf_reference_archives/PDFReference.pdf
doc = page.parent # type: fitz.Document
for xref in xref_list:
stream = doc.xref_stream(xref).replace(b'BT', b'BT 3 Tr') \
.replace(b'Tm', b'Tm 3 Tr') \
.replace(b'Td', b'Td 3 Tr')
doc.update_stream(xref, stream)