def update_preview(self):
if self.pages is not None:
pdf_file = self.pages[self.current_page - 1]
pdf_file.seek(0)
document = Document(stream=pdf_file, filetype="PDF")
image = QPixmap()
page = document.loadPage(0)
container_size = self.preview_box.parent().size()
normalized_zoom_factor = min(
page.rect.height / container_size.height(),
page.rect.width / container_size.width(),
)
scale_mat = Matrix(normalized_zoom_factor, normalized_zoom_factor)
image.loadFromData(page.getPixmap(matrix=scale_mat).getPNGData())
image_size = container_size * self.zoom_factor
self.preview_box.setPixmap(
image.scaled(image_size, Qt.KeepAspectRatio))
self.preview_box.resize(image_size)
else:
self.preview_box.setTextFormat(Qt.RichText)
self.preview_box.setStyleSheet("QLabel { color : darkgray; }")
self.preview_box.setText(
f'<center><img src="{appctxt.get_resource("logo.png")}") /><br />Wähle auf der rechten Seite<br />eine Datei und ein Papierformat<br />aus um die Vorschau anzuzeigen</center>'
)
self.preview_box.resize(QSize(200, 200))
def manga_to_PDF(dir_path, one_file=True, width=None, height=None):
if dir_path[-1] == '\\' or dir_path[-1] == '/':
dir_path = dir_path[0:-1]
if not os.path.isdir(dir_path):
raise ValueError('传入的路径并非文件夹')
from fitz import Document, Pixmap, Rect
from glob import _iglob as glob
if one_file:
title = os.path.basename(dir_path)
with Document() as doc:
for file_path in glob(os.path.join(dir_path, "*", "*.jpg"), False,
False):
pixmap = Pixmap(file_path)
if width and height:
pixmap = Pixmap(pixmap, width, height, None)
elif width:
pixmap = Pixmap(pixmap, width,
int(pixmap.height / pixmap.width * width),
None)
elif height:
pixmap = Pixmap(pixmap,
int(pixmap.width / pixmap.height * height),
height, None)
rect = Rect(0, 0, pixmap.width, pixmap.height)
page = doc.newPage(width=pixmap.width, height=pixmap.height)
page.insertImage(rect, pixmap=pixmap)
doc.save(os.path.join(dir_path, title + ".pdf"), deflate=True)
else:
for chap in glob(os.path.join(dir_path, "*"), False, True):
title = os.path.basename(chap)
with Document() as doc:
for file_path in glob(os.path.join(chap, "*.jpg"), False,
False):
pixmap = Pixmap(file_path)
if width and height:
pixmap = Pixmap(pixmap, width, height, None)
elif width:
pixmap = Pixmap(
pixmap, width,
int(pixmap.height / pixmap.width * width), None)
elif height:
pixmap = Pixmap(
pixmap, int(pixmap.width / pixmap.height * height),
height, None)
rect = Rect(0, 0, pixmap.width, pixmap.height)
page = doc.newPage(width=pixmap.width,
height=pixmap.height)
page.insertImage(rect, pixmap=pixmap)
doc.save(os.path.join(dir_path, title + ".pdf"), deflate=True)
def inspect(file: Union[str, bytes, Path]) -> PdfRedactionsDict:
"""
Inspect a file for bad redactions and return a Dict with their info
:file: The PDF to process, as bytes if you have the file in memory (useful
if it's coming from the network), as a unicode string if you know the
path to the file on your local disk, or as a pathlib.Path object.
:return: A dict with the bad redaction information. If no bad redactions
are found, returns an empty dict.
"""
if type(file) == bytes:
pdf = Document(stream=file, filetype="pdf")
elif type(file) == str and file.startswith("https://"):
r = requests.get(file, timeout=10)
r.raise_for_status()
pdf = Document(stream=r.content, filetype="pdf")
else:
# str filepath or Pathlib Path
pdf = Document(file)
bad_redactions = {}
for page_number, page in enumerate(pdf, start=1):
redactions = get_bad_redactions(page)
if redactions:
bad_redactions[page_number] = redactions
pdf.close()
bad_redactions = check_if_all_dates(bad_redactions)
return bad_redactions
def extract_meta(doc: Document,
pattern: str,
page: Optional[int] = None,
ign_case: bool = False) -> List[dict]:
"""Extract meta for a `pattern` on `page` in a pdf document
Arguments
doc: document from pymupdf
pattern: a regular expression pattern
page: page number (1-based index), if None is given, search for the
entire document, but this is highly discouraged.
ign_case: ignore case?
"""
result = []
if page is None:
pages = doc.pages()
elif 1 <= page <= doc.pageCount:
pages = [doc[page - 1]]
else: # page out of range
return result
regex = re.compile(pattern,
re.IGNORECASE) if ign_case else re.compile(pattern)
# we could parallelize this, but I don't see a reason
# to *not* specify a page number
for p in pages:
result.extend(search_in_page(regex, p))
return result
def attact_annotation(annot: Annotation, doc: fitz.Document):
page: fitz.Page = doc.loadPage(annot.pageNum)
atype = annot.annotType[0]
ann_new = None
if atype == fitz.ANNOT_FREETEXT:
ann_new = page.addFreetextAnnot(annot.rect, annot.info['content'])
elif atype == fitz.ANNOT_TEXT:
ann_new = page.addTextAnnot(annot.point, annot.info['content'])
elif atype == fitz.ANNOT_HIGHLIGHT:
ann_new = page.addHighlightAnnot(annot.quads)
elif atype == fitz.ANNOT_STRIKEOUT:
ann_new = page.addStrikeOUTAnnot(annot.quads)
elif atype == fitz.ANNOT_SQUIGGLY:
ann_new = page.addSquigglyAnnot(annot.quads)
elif atype == fitz.ANNOT_UNDERLINE:
ann_new = page.addUnderlineAnnot(annot.quads)
else:
print(
f'Annotation type {annot.annotType} is not supported. Ignore: {annot}'
)
if ann_new:
ann_new: fitz.Annot
ann_new.setInfo(annot.info)
ann_new.setColors(annot.colors)
ann_new.setBorder(annot.border)
if annot.lineEnds:
ann_new.setLineEnds(*annot.lineEnds)
ann_new.setOpacity(annot.opacity)
def extract_info_from_pdf(doc: fitz.Document, user: str) -> [Train]:
trains_list = list()
for i in range(doc.pageCount):
page = doc.loadPage(i)
text = page.getText("text")
train = extract_info_from_text(text, user)
if train is not None:
trains_list.append(train)
return trains_list
def read_book(file):
"""
This function allows you to read book
:file: must be string formatted .pdf or .epub or .fb2 file.
Other formats will raise exception.
"""
if file.endswith('.pdf') or file.endswith('.epub') or file.endswith(
'.fb2'):
return Document(file)
def save_transform_pdf(
case_id: str,
pdf_file: FileStorage,
):
__init_time = time.time()
__total_rendering_time = 0
case_dir_path = __case_dir_path(case_id)
os.makedirs(case_dir_path, exist_ok=True)
pdf_path = "%s%08d.pdf" % (parent_dir, rd.randint(0, 1e7))
pdf_file.save(pdf_path)
pdf = Document(pdf_path)
refs_size = pdf._getXrefLength()
img_cnt = 0
q = Queue()
# 遍历每一个对象
for i in range(1, refs_size):
# 定义对象字符串
text = pdf._getXrefString(i)
is_object = re.search(checkXO, text)
# 使用正则表达式查看是否是图片
is_img = re.search(checkIM, text)
# 如果不是对象也不是图片，则continue
if not is_object or not is_img:
continue
img_cnt += 1
path = '%s%08d.png' % (case_dir_path, img_cnt)
# 根据索引生成图像
q.put((pdf_path, i, path))
procs = [
Process(target=render, name="P%d" % _pi, args=(q, ))
for _pi in range(64)
]
for proc in procs:
print('start', proc.name)
proc.start()
for proc in procs:
proc.join()
print("total time: %.4f" % (time.time() - __init_time))
def _is_device_cs(xref, doc: fitz.Document):
'''Check whether object xref is a device based color space.
'''
# cs definition
obj_contents = doc.xrefObject(xref)
# for now, just check /ICCBased CS:
# it's treated as a device based cs if /Device[Gray|RGB|CMYK] exists in /Alternate.
#
# [ /ICCBased 15 0 R ]
#
# <<
# /Alternate /DeviceRGB
# /Filter /FlateDecode
# /Length 2597
# /N 3
# >>
if '/ICCBased' in obj_contents:
name, x, *_ = obj_contents[1:-1].strip().split()
ICC_contents = doc.xrefObject(int(x))
return '/Alternate /Device' in ICC_contents
# ignore all other color spaces, may include if facing associated cases
return False
def return_page(doc: Document, page_num=0):
"""
This function allows you to read any page of your file in html format.
:doc: Document with all pages.
:page_num: Number of document page, default value = 0.
Wrong number raise exception.
"""
if page_num < 0 or page_num >= doc.page_count:
raise Exception("Wrong page")
page = doc.loadPage(page_num)
return page
def show_page(doc: Document, page_num=0):
"""
This function allows you to read any page of your file in html format.
:doc: Document with all pages.
:page_num: Number of document page, default value = 0.
Wrong number raise exception.
"""
if page_num < 0 or page_num >= doc.page_count:
raise Exception("Wrong page")
pix = doc.loadPage(page_num).get_pixmap()
fmt = QImage.Format_RGBA8888 if pix.alpha else QImage.Format_RGB888
qtimg = QImage(pix.samples, pix.width, pix.height, pix.stride, fmt)
return qtimg
def extract_toc(doc: Document, recipe: Recipe) -> List[ToCEntry]:
"""Extract toc entries from a document
Arguments
doc: a pdf document
recipe: recipe from user
Returns
a list of toc entries in the document
"""
result = []
for page in doc.pages():
for blk in page.getTextPage().extractDICT().get('blocks', []):
result.extend(
recipe.extract_block(blk, page.number + 1)
)
return result
def render(q: Queue):
__pdf_path = None
__pdf_doc = None
while not q.empty():
pdf_path, xref, save_to = q.get()
if __pdf_path == pdf_path:
pdf_doc = __pdf_doc
else:
print('Worker Read')
pdf_doc = Document(pdf_path)
__pdf_path = pdf_path
__pdf_doc = pdf_doc
pix = fitz.Pixmap(pdf_doc, xref)
if pix.n < 5:
pix.writePNG(save_to)
# 否则先转换CMYK
else:
fitz.Pixmap(fitz.csRGB, pix).writePNG(save_to)
def _check_device_cs(doc:fitz.Document, page:fitz.Page):
'''Get all color space name used in current page and check if they're device based color space.'''
# default device based cs
cs = {
'/DeviceGray': True,
'/DeviceRGB' : True,
'/DeviceCMYK': True
}
# content of page object, e.g.
# <<
# ...
# /Resources <<
# ...
# /ColorSpace <<
# /Cs6 14 0 R
# >>
# >>
# /Rotate 0
# /Type /Page
# >>
obj_contents = doc.xrefObject(page.xref)
cs_found = False
for line_ in obj_contents.splitlines():
line = line_.strip()
# check start/end of color space block
if not cs_found and line.startswith('/ColorSpace'):
cs_found = True
continue
if not cs_found:
continue
elif line=='>>':
break
# now within cs block, e.g. /Cs6 14 0 R
cs_name, xref, *_ = line.split()
cs[cs_name] = _is_device_cs(int(xref), doc)
return cs
def test_reading_books_test1(self):
test_file = "c++_trofumenko.pdf"
actual_result = read_book(test_file).page_count
expected_result = Document(test_file).page_count
self.assertEqual(actual_result, expected_result)
def write_toc(doc: Document, toc: List[ToCEntry]):
"""Write table of contents to a document"""
fitz_toc = list(map(lambda e: e.to_fitz_entry(), toc))
doc.set_toc(fitz_toc)
def read_toc(doc: Document) -> List[ToCEntry]:
"""Read table of contents from a document"""
return [ToCEntry(*entry) for entry in doc.get_toc()]
page_delta = args.delta[0] + 1
else:
page_delta = 1
if args.start:
start = args.start[0]
else:
start = 0
if args.end:
end = args.end[0]
results_file = args.output[0]
font_standard = args.font_size[0]
pdf_file = args.pdf_file[0]
names_file = args.names[0]
with open(names_file, "r") as file:
names_list = tuple(map(lambda x: str(x).strip("\n"), file.readlines()))
doc = Document(pdf_file)
if end is None:
end = len(doc)
results = open(results_file, "a")
for name in names_list:
print(f"Parsing: {name}")
word_in_pages = []
results.write(f"{name} ")
for page in range(start, end):
lines = unescape(doc[page].getText("html")).split("\n")
for line in lines:
item = line.replace("\t", " ")
bs = BeautifulSoup(item, "html.parser")
if bs.p is not None:
if name.title() in str(bs.p.text) and str(
bs.p["style"]).find("top:84pt") == -1:
