def filter_redactions_by_pixmap(
redactions: List[RedactionType],
page: Page,
) -> List[RedactionType]:
"""Convert each bad redaction to an image and check it for text
:param redactions: A list of redactions that might be bad
:param page: The PyMuPDF.Page object where the bad redactions might be
:return: The redactions, if they are valid
"""
bad_redactions = []
for redaction in redactions:
pixmap = page.get_pixmap(
# Use gray for simplicity and speed, though this risks missing a
# bad redaction.
colorspace=fitz.csRGB,
clip=fitz.Rect(redaction["bbox"]),
)
if not pixmap.is_unicolor:
# There's some degree of variation in the colors of the pixels.
# ∴ it's not a uniform box and it's not a bad redaction.
# filename = f'{redaction["text"].replace("/", "_")}.png'
# pixmap.save(filename)
continue
bad_redactions.append(redaction)
return bad_redactions
def image_from_page(page: fitz.Page, scale: float = 1) -> Image:
"""Converts a page to an image.
:param page: the page to be represented as an image
:param scale: the proportion by which to scale the image
"""
pix = page.get_pixmap(matrix=fitz.Matrix(scale, scale)) # type: ignore
return Image.frombytes( # type: ignore
("RGBA" if pix.alpha else "RGB"),
(pix.width, pix.height), pix.samples
)