fixing https://github.com/qurator-spk/eynollah/issues/169 : This fix actually resolves the error. Additionally, the contour with intersections no longer appears in the results.

[kba]

For discussion: @vahidrezanezhad has a workaround for the multipolygon issue where eynollah produces polygons with self-intersections.

[codecov]

Codecov Report

❌ Patch coverage is 33.33333% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 58.36%. Comparing base (c65b345) to head (67be590).

Files with missing lines	Patch %	Lines
ocrd_tesserocr/common.py	33.33%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #221      +/-   ##
==========================================
- Coverage   58.45%   58.36%   -0.09%     
==========================================
  Files          12       12              
  Lines        1372     1374       +2     
  Branches      323      325       +2     
==========================================
  Hits          802      802              
- Misses        434      435       +1     
- Partials      136      137       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[bertsky]

I can see how this would help – as a workaround.

But really, this is not the way to fix this. The code above in particular tries to avoid yielding multiple geometries – by finding the best path through all parts and connecting them with little "bridges". So if that failed, for some reason, we must fix that directly.

Could you please provide me with the input polygons that caused this?

[bertsky]

One simple way to share example polygon data is by using Shapely's wkt serialization, which can easily be printed, plotted, re-imported etc. So here, e.g. if you do...

Suggested change

	jointp = max(jointp.geoms, key=lambda p: p.area) # Choose the largest polygon
	for geom in jointp.geoms:
	print(geom.wkt)
	jointp = max(jointp.geoms, key=lambda p: p.area) # Choose the largest polygon

...and then send me the stdout, that would really help.