qurator-spk/ocrd_repair_inconsistencies
1
0
Fork 0
mirror of https://github.com/qurator-spk/ocrd_repair_inconsistencies.git synced 2025-10-31 00:44:13 +01:00
Code Issues Releases Wiki Activity

pylint karma

Browse source
This commit is contained in:
Robert Sachunsky 2019-11-29 18:20:28 +01:00
parent e16438a377
commit d5094ee42b
3 changed files with 33 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

24
.pylintrc Normal file
View file

@ -0,0 +1,24 @@
[MASTER]
[MESSAGES CONTROL]
disable =
ungrouped-imports,
bad-continuation,
missing-docstring,
no-self-use,
superfluous-parens,
invalid-name,
line-too-long,
too-many-arguments,
too-many-branches,
too-many-statements,
too-many-locals,
too-few-public-methods,
wrong-import-order,
duplicate-code
# allow indented whitespace (as required by interpreter):
no-space-check=empty-line
# allow non-snake-case identifiers:
good-names=n,i

2
ocrd_repair_inconsistencies/config.py
View file

@ -1,4 +1,4 @@
import json import json
from pkg_resources import resource_string from pkg_resources import resource_string
OCRD_TOOL = json.loads(resource_string(__name__, 'ocrd-tool.json').decode('utf8')) OCRD_TOOL = json.loads(resource_string(__name__, 'ocrd-tool.json').decode('utf8'))

18
ocrd_repair_inconsistencies/ocrd_repair_inconsistencies.py
View file

@ -64,7 +64,7 @@ class RepairInconsistencies(Processor):
page_id, region.id, textLineOrder) page_id, region.id, textLineOrder)
continue continue
_fix_segment(region, page_id, reverse=textLineOrder=='bottom-to-top') _fix_segment(region, page_id, reverse=(textLineOrder == 'bottom-to-top'))
lines = region.get_TextLine() lines = region.get_TextLine()
for line in lines: for line in lines:
@ -80,7 +80,7 @@ class RepairInconsistencies(Processor):
page_id, line.id, readingDirection) page_id, line.id, readingDirection)
continue continue
_fix_segment(line, page_id, reverse=readingDirection=='right-to-left') _fix_segment(line, page_id, reverse=(readingDirection == 'right-to-left'))
words = line.get_Word() words = line.get_Word()
for word in words: for word in words:
@ -96,7 +96,7 @@ class RepairInconsistencies(Processor):
page_id, word.id, readingDirection) page_id, word.id, readingDirection)
continue continue
_fix_segment(word, page_id, reverse=readingDirection=='right-to-left') _fix_segment(word, page_id, reverse=(readingDirection == 'right-to-left'))
file_id = input_file.ID.replace(self.input_file_grp, self.output_file_grp) file_id = input_file.ID.replace(self.input_file_grp, self.output_file_grp)
if file_id == input_file.ID: if file_id == input_file.ID:
@ -124,10 +124,8 @@ def get_text(thing, joiner=''):
texts = [_get_text_for_one(part) for part in thing] texts = [_get_text_for_one(part) for part in thing]
if all(texts): if all(texts):
return joiner.join(texts) return joiner.join(texts)
else: return None
return None return _get_text_for_one(thing)
else:
return _get_text_for_one(thing)
def _fix_segment(segment, page_id, reverse=False): def _fix_segment(segment, page_id, reverse=False):
"""Fix order of child elements of (region/line/word) segment.""" """Fix order of child elements of (region/line/word) segment."""
@ -160,8 +158,7 @@ def _fix_segment(segment, page_id, reverse=False):
polygon = Polygon(polygon_from_points(child.get_Coords().points)) polygon = Polygon(polygon_from_points(child.get_Coords().points))
if horizontal: if horizontal:
return polygon.centroid.x return polygon.centroid.x
else: return polygon.centroid.y
return polygon.centroid.y
sorted_children = sorted(children, reverse=reverse, key=polygon_position) sorted_children = sorted(children, reverse=reverse, key=polygon_position)
sorted_concat_text = get_text(sorted_children, joiner) sorted_concat_text = get_text(sorted_children, joiner)
@ -170,7 +167,8 @@ def _fix_segment(segment, page_id, reverse=False):
LOG.info('Fixing element order of page "%s" segment "%s"', page_id, segment.id) LOG.info('Fixing element order of page "%s" segment "%s"', page_id, segment.id)
adoption(sorted_children) adoption(sorted_children)
else: else:
LOG.debug('Resorting children of page "%s" segment "%s" from %s to %s does not suffice to turn "%s" into "%s"', LOG.debug('Resorting children of page "%s" segment "%s" from %s to %s' +
'does not suffice to turn "%s" into "%s"',
page_id, segment.id, page_id, segment.id,
str([seg.id for seg in children]), str([seg.id for seg in children]),
str([seg.id for seg in sorted_children]), str([seg.id for seg in sorted_children]),