Fix comments
This commit is contained in:
@@ -247,10 +247,10 @@ class Stream(BaseParser):
|
||||
" should be equal")
|
||||
|
||||
def _nurminen_table_detection(self, textlines):
|
||||
# an general heuristic implementation of the table detection
|
||||
# a general heuristic implementation of the table detection
|
||||
# algorithm described by Anssi Nurminen's master's thesis:
|
||||
# https://dspace.cc.tut.fi/dpub/bitstream/handle/123456789/21520/Nurminen.pdf?sequence=3
|
||||
# assumes that tables vertically separated by some distance
|
||||
# assumes that tables are situated relatively apart vertically
|
||||
|
||||
# TODO: add support for arabic text #141
|
||||
# sort textlines in reading order
|
||||
@@ -263,9 +263,9 @@ class Stream(BaseParser):
|
||||
textedges.generate(text_grouped)
|
||||
# select relevant edges
|
||||
relevant_textedges = textedges.get_relevant()
|
||||
# guess table areas using relevant edges
|
||||
# guess table areas using textlines and relevant edges
|
||||
table_bbox = textedges.get_table_areas(textlines, relevant_textedges)
|
||||
# treat whole page as table if not table areas found
|
||||
# treat whole page as table area if no table areas found
|
||||
if not len(table_bbox):
|
||||
table_bbox = {(0, 0, self.pdf_width, self.pdf_height): None}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user