def test_issue_13(self):
"""
Test slightly simplified from gist here: https://github.com/jsvine/pdfplumber/issues/13
"""
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/issue-13-151201DSP-Fond-581-90D.pdf")
)
# Only find checkboxes this size
RECT_WIDTH = 9.3
RECT_HEIGHT = 9.3
RECT_TOLERANCE = 2
def filter_rects(rects):
## Just get the rects that are the right size to be checkboxes
rects_found = []
for rect in rects:
if ( rect['height'] > ( RECT_HEIGHT - RECT_TOLERANCE )
and ( rect['height'] < RECT_HEIGHT + RECT_TOLERANCE)
and ( rect['width'] < RECT_WIDTH + RECT_TOLERANCE)
and ( rect['width'] < RECT_WIDTH + RECT_TOLERANCE) ):
rects_found.append(rect)
return rects_found
def determine_if_checked(checkbox, curve_list):
# This figures out if the bounding box of (either) line used to make
# one half of the 'x' is the right size and overlaps with a rectangle.
# This isn't foolproof, but works for this case.
# It's not totally clear (to me) how common this style of checkboxes
# are used, and whether this is useful approach to them.
# Also note there should be *two* matching LTCurves for each checkbox.
# But here we only test there's at least one.
for curve in curve_list:
if ( checkbox['height'] > ( RECT_HEIGHT - RECT_TOLERANCE )
and ( checkbox['height'] < RECT_HEIGHT + RECT_TOLERANCE)
and ( checkbox['width'] < RECT_WIDTH + RECT_TOLERANCE)
and ( checkbox['width'] < RECT_WIDTH + RECT_TOLERANCE) ):
xmatch = False
ymatch = False
if ( max(checkbox['x0'], curve['x0']) <= min(checkbox['x1'], curve['x1']) ):
xmatch = True
if ( max(checkbox['y0'], curve['y0']) <= min(checkbox['y1'], curve['y1']) ):
ymatch = True
if xmatch and ymatch:
return True
return False
p0 = pdf.pages[0]
curves = p0.objects["curve"]
rects = filter_rects(p0.objects["rect"])
n_checked = sum([ determine_if_checked(rect, curves)
for rect in rects ])
assert(n_checked == 5)
def test_rotation(self):
rotated = pdfplumber.from_path(
os.path.join(HERE, "pdfs/nics-background-checks-2015-11-rotated.pdf")
)
assert(self.pdf.pages[0].width == 1008)
assert(self.pdf.pages[0].height == 612)
assert(rotated.pages[0].width == 612)
assert(rotated.pages[0].height == 1008)
assert(rotated.pages[0].cropbox == self.pdf.pages[0].cropbox)
assert(rotated.pages[0].bbox != self.pdf.pages[0].bbox)
def extract(pdf_path: str, filter=None, flavor='lattice', lang: str = 'eng', **imgOcrSettings):
'''
: 抽取pdf中的表格数据
'''
pdf = pdfplumber.from_path(pdf_path)
total_page = len(pdf.pages)
tables: [PageTable] = []
# 使用camelot抽取表格
print('use camelot extract tables')
camelot_tables = camelot.read_pdf(pdf_path, pages='all',
flavor=flavor, suppress_stdout=False)
for t in camelot_tables:
text = pdf.pages[t.page - 1].extract_text()
merge_table(tables, t.page, t.data, text)
# 如果抽取完成则返回
if len(tables) == total_page:
return tables
# 否则使用ocr抽取其他页面的表格
extract_pages = [t.page for t in tables]
total_page_set = set(range(1, total_page + 1))
extracted_pages_set = set(extract_pages)
other_pages = list(total_page_set.difference(extracted_pages_set))
for page_number in other_pages:
other_tables = extract_imgbase(
pdf, page_number, flavor, lang, filter, **imgOcrSettings)
if other_tables is not None:
merge_tables(tables, page_number, other_tables)
pdf.close()
return tables
def test_issue_33(self):
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/issue-33-lorem-ipsum.pdf")
)
assert len(pdf.metadata.keys())
def test_issue_21(self):
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/150109DSP-Milw-505-90D.pdf")
)
assert len(pdf.objects)
def test_issue_14(self):
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/cupertino_usd_4-6-16.pdf")
)
assert len(pdf.objects)
def setUp(self):
path = os.path.join(HERE, "pdfs/nics-background-checks-2015-11.pdf")
self.pdf = pdfplumber.from_path(path)
def setUp(self):
print('i am running')
path = os.path.join('/home/sxs/yuhsuan/datasets/personal/003.pdf')
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def setUp(self):
path = os.path.join(HERE, "pdfs/nics-background-checks-2015-11.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def setUp(self):
path = os.path.join(HERE, "pdfs/nics-background-checks-2015-11.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def setUp(self):
path = os.path.join(HERE,
"pdfs/WARN-Report-for-7-1-2015-to-03-25-2016.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def setUp(self):
path = os.path.join(HERE, "pdfs/la-precinct-bulletin-2014-p1.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def test_load(self):
path = os.path.join(HERE, "pdfs/cupertino_usd_4-6-16.pdf")
pdf = pdfplumber.from_path(path)
def test_issue_53(self):
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/issue-53-example.pdf")
)
assert len(pdf.objects)
def setUp(self):
path = os.path.join(HERE, "pdfs/WARN-Report-for-7-1-2015-to-03-25-2016.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width
def test_issue_67(self):
pdf = pdfplumber.from_path(
os.path.join(HERE, "pdfs/issue-67-example.pdf")
)
assert len(pdf.metadata.keys())
def setUp(self):
path = os.path.join(HERE, "pdfs/la-precinct-bulletin-2014-p1.pdf")
self.pdf = pdfplumber.from_path(path)
self.PDF_WIDTH = self.pdf.pages[0].width