def page_contains_tables(layout, device):
# TODO: hide doc, interpreter, device inside a higher level Pdf class. It's
# silly that we have to care about these (see function signature!!)
box_list = LeafList().populate(layout)
for item in box_list:
assert isinstance(item, Leaf), "NOT LEAF"
yhist = box_list.histogram(Leaf._top).rounder(1)
test = [k for k, v in yhist.items() if v > IS_TABLE_COLUMN_COUNT_THRESHOLD]
return len(test) > IS_TABLE_ROW_COUNT_THRESHOLD
def page_contains_tables(pdf_page):
if not isinstance(pdf_page, PDFPage):
raise TypeError("Page must be PDFPage, not {}".format(
pdf_page.__class__))
box_list = LeafList().populate(pdf_page)
for item in box_list:
assert isinstance(item, Leaf), "NOT LEAF"
yhist = box_list.histogram(Leaf._top).rounder(1)
test = [k for k, v in yhist.items() if v > IS_TABLE_COLUMN_COUNT_THRESHOLD]
return len(test) > IS_TABLE_ROW_COUNT_THRESHOLD
def page_contains_tables(pdf_page):
if not isinstance(pdf_page, PDFPage):
raise TypeError("Page must be PDFPage, not {}".format(
pdf_page.__class__))
box_list = LeafList().populate(pdf_page)
for item in box_list:
assert isinstance(item, Leaf), "NOT LEAF"
yhist = box_list.histogram(Leaf._top).rounder(1)
test = [k for k, v in yhist.items() if v > IS_TABLE_COLUMN_COUNT_THRESHOLD]
return len(test) > IS_TABLE_ROW_COUNT_THRESHOLD
def page_contains_tables(pdf_page, interpreter, device):
# TODO: hide doc, interpreter, device inside a higher level Pdf class. It's
# silly that we have to care about these (see function signature!!)
interpreter.process_page(pdf_page)
# receive the LTPage object for the page.
layout = device.get_result()
box_list = LeafList().populate(layout)
for item in box_list:
assert isinstance(item, Leaf), "NOT LEAF"
yhist = box_list.histogram(Leaf._top).rounder(1)
test = [k for k, v in yhist.items() if v > IS_TABLE_COLUMN_COUNT_THRESHOLD]
return len(test) > IS_TABLE_ROW_COUNT_THRESHOLD
