def __init__(self):
LTTextContainer.__init__(self)
self.start_row = -1
self.end_row = -1
self.start_col = -1
self.end_col = -1
self.children = []
self.father = None
def callback(container: LTTextContainer):
"""Checks whether an container includes the provided text."""
if not hasattr(container, 'get_text'):
raise TypeError
text = container.get_text().strip().replace('\n', '')
return all(map(lambda s: s in text, strings))
def group(characters, delta=5, use_h_axis=True):
""" group characters based on proximity
Args:
use_h_axis: whether to group characters based on their distance
horizontally. When False, you group text into lines
"""
groups = []
for c in characters:
obj = None
for obj in groups:
if use_h_axis:
if obj.vdistance(c) < delta and obj.hdistance(c) < delta:
break
else:
if obj.vdistance(c) < delta:
break
obj = None
if obj is None:
obj = LTTextContainer()
obj.add(c)
groups.append(obj)
else:
obj.add(c)
return groups
def merge(lines, delta=5):
""" merge text elements where they should be in a single group
Some text elements might get placed into separate groups if they are are on
separate lines, and their lines start at different x-positions. This function
cleans those up.
"""
merged = []
for line in lines:
elem = list(line)
for a, b in itertools.combinations(range(len(elem)), 2):
if elem[a] is None or elem[b] is None:
continue
if elem[a].hdistance(elem[b]) < delta:
elem[a].extend(elem[b])
elem[b] = None
line = LTTextContainer()
line.extend([x for x in elem if x is not None])
merged.append(line)
return merged
def process_paragraph(self, paragraph: LTTextContainer, index: int,
page_containers: List[LTTextContainer]):
"""Process a paragraph. The default add a paragraph to the CorpusDocument"""
self._result.add_paragraph(paragraph.get_text().strip(), str(index))
def check_text_is_date(paragraph: LTTextContainer) -> bool:
"""Returns true if the text of a paragraph contains only a date."""
return re.match(r'\d\d? \w+ \d\d\d\d\s*$',
paragraph.get_text().strip()) is not None
def callback(element: LTTextContainer):
"""Checks whether an element's text does not match the provided text."""
if hasattr(element, 'get_text'):
return text != element.get_text().strip().replace('\n', '')
else:
raise TypeError
from pdfminer.high_level import extract_pages
from pdfminer.layout import LTTextContainer, LTChar, LTAnno
import os
def label_func():
return '1. Определение'
labels_txt = open('labels.txt', 'w')
for page_layout in extract_pages("text1.pdf"):
for element in page_layout:
if isinstance(element, LTTextContainer):
for text_line in element:
if isinstance(element, LTTextContainer):
if str(LTTextContainer.get_text(
self=text_line)).startswith(label_func()):
print(str(LTTextContainer.get_text(self=text_line)))