def identify_num_columns(self):
column_count = dict()
self.column_dim = list()
column_dim_list = list()
for page in self.pages:
column_dim = list()
for segment in page.segments:
if segment.font_family == self.default_font_family and math.fabs(segment.font_size - self.default_size) < 0.1:
column_dim_copy = list(column_dim)
updated_dim = False
for dim in column_dim_copy:
if math.fabs(segment.bbox[0] - dim[0]) < 10.0:
column_dim.remove(dim)
minx = segment.bbox[0] if segment.bbox[0] < dim[0] else dim[0]
maxx = segment.bbox[2] if segment.bbox[2] > dim[1] else dim[1]
column_dim.append( (minx, maxx) )
updated_dim = True
break
if updated_dim == False:
column_dim.append( (segment.bbox[0], segment.bbox[2]) )
minx_flow = self.flow_bbox[0] if self.flow_bbox[0] < segment.bbox[0] else segment.bbox[0]
miny_flow = self.flow_bbox[1] if self.flow_bbox[1] < segment.bbox[1] else segment.bbox[1]
maxx_flow = self.flow_bbox[2] if self.flow_bbox[2] > segment.bbox[2] else segment.bbox[2]
maxy_flow = self.flow_bbox[3] if self.flow_bbox[3] > segment.bbox[3] else segment.bbox[3]
self.flow_bbox = (minx_flow, miny_flow, maxx_flow, maxy_flow)
column_dim_list.append( column_dim )
if len(column_dim) in column_count:
column_count[ len(column_dim) ] += 1
else:
column_count[ len(column_dim) ] = 1
self.num_columns = find_most_frequent_item(column_count)
for column_dim in column_dim_list:
for dim in column_dim:
updated = False
for column_index in range(self.num_columns):
if dim[0] >= self.flow_bbox[0] + ((self.flow_bbox[2] - self.flow_bbox[0]) / self.num_columns) * column_index and \
dim[1] <= self.flow_bbox[0] + ((self.flow_bbox[2] - self.flow_bbox[0]) / self.num_columns) * (column_index+1):
for i in range(len(self.column_dim)):
test_dim = self.column_dim[i]
if math.fabs(test_dim[0] - dim[0]) < 10.0 and math.fabs(test_dim[1] - dim[1]) < 10.0:
minx = test_dim[0] if test_dim[0] < dim[0] else dim[0]
maxx = test_dim[1] if test_dim[1] > dim[1] else dim[1]
self.column_dim[i] = (minx, maxx)
updated = True
break
if updated == False:
self.column_dim.append( (dim[0], dim[1]) )
def find_default_fonts(self):
page_font_count = dict()
size_count = dict()
for page in self.pages:
for segment in page.segments:
if segment.contains_text():
for k, v in segment.font_count.items():
if k in page_font_count:
page_font_count[k] += v
else:
page_font_count[k] = v
if segment.font_size in size_count:
size_count[segment.font_size] += segment.font_count[segment.font]
else:
size_count[segment.font_size] = segment.font_count[segment.font]
font = find_most_frequent_item(page_font_count).split(",")
self.default_font_family = font[0]
self.default_font_type = "Regular" if len(font) == 1 else font[1]
self.default_size = find_most_frequent_item(size_count)
def find_default_fonts(self):
page_font_count = dict()
size_count = dict()
for page in self.pages:
for segment in page.segments:
if segment.contains_text():
for k, v in segment.font_count.items():
if k in page_font_count:
page_font_count[k] += v
else:
page_font_count[k] = v
if segment.font_size in size_count:
size_count[segment.font_size] += segment.font_count[
segment.font]
else:
size_count[segment.font_size] = segment.font_count[
segment.font]
font = find_most_frequent_item(page_font_count).split(",")
self.default_font_family = font[0]
self.default_font_type = "Regular" if len(font) == 1 else font[1]
self.default_size = find_most_frequent_item(size_count)
def identify_num_columns(self):
column_count = dict()
self.column_dim = list()
column_dim_list = list()
for page in self.pages:
column_dim = list()
for segment in page.segments:
if segment.font_family == self.default_font_family and math.fabs(
segment.font_size - self.default_size) < 0.1:
column_dim_copy = list(column_dim)
updated_dim = False
for dim in column_dim_copy:
if math.fabs(segment.bbox[0] - dim[0]) < 10.0:
column_dim.remove(dim)
minx = segment.bbox[
0] if segment.bbox[0] < dim[0] else dim[0]
maxx = segment.bbox[
2] if segment.bbox[2] > dim[1] else dim[1]
column_dim.append((minx, maxx))
updated_dim = True
break
if updated_dim == False:
column_dim.append((segment.bbox[0], segment.bbox[2]))
minx_flow = self.flow_bbox[0] if self.flow_bbox[
0] < segment.bbox[0] else segment.bbox[0]
miny_flow = self.flow_bbox[1] if self.flow_bbox[
1] < segment.bbox[1] else segment.bbox[1]
maxx_flow = self.flow_bbox[2] if self.flow_bbox[
2] > segment.bbox[2] else segment.bbox[2]
maxy_flow = self.flow_bbox[3] if self.flow_bbox[
3] > segment.bbox[3] else segment.bbox[3]
self.flow_bbox = (minx_flow, miny_flow, maxx_flow,
maxy_flow)
column_dim_list.append(column_dim)
if len(column_dim) in column_count:
column_count[len(column_dim)] += 1
else:
column_count[len(column_dim)] = 1
self.num_columns = find_most_frequent_item(column_count)
for column_dim in column_dim_list:
for dim in column_dim:
updated = False
for column_index in range(self.num_columns):
if dim[0] >= self.flow_bbox[0] + ((self.flow_bbox[2] - self.flow_bbox[0]) / self.num_columns) * column_index and \
dim[1] <= self.flow_bbox[0] + ((self.flow_bbox[2] - self.flow_bbox[0]) / self.num_columns) * (column_index+1):
for i in range(len(self.column_dim)):
test_dim = self.column_dim[i]
if math.fabs(test_dim[0] -
dim[0]) < 10.0 and math.fabs(
test_dim[1] - dim[1]) < 10.0:
minx = test_dim[
0] if test_dim[0] < dim[0] else dim[0]
maxx = test_dim[
1] if test_dim[1] > dim[1] else dim[1]
self.column_dim[i] = (minx, maxx)
updated = True
break
if updated == False:
self.column_dim.append((dim[0], dim[1]))
