OCR-D / ocrd_calamari

Recognize text using Calamari OCR and the OCR-D framework
Apache License 2.0
13 stars 6 forks source link

descend to all available TextRegions recursively #75

Closed bertsky closed 2 years ago

bertsky commented 2 years ago

Since OcrdPage.get_AllRegions has been introduced in core, we support region-region recursion (which is unavoidable for cases like drop-cap text-regions, tables or other region types with text) in most processors (including all OCR wrappers). Not using this construct effectively makes a processor go blind on these structures. (This is especially bad if you want to do multi-OCR alignment or OCR evaluation.)

codecov-commenter commented 2 years ago

Codecov Report

Merging #75 (36e5136) into master (ecb0d67) will decrease coverage by 1.18%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #75      +/-   ##
==========================================
- Coverage   88.16%   86.98%   -1.19%     
==========================================
  Files           3        3              
  Lines         169      169              
  Branches       38       38              
==========================================
- Hits          149      147       -2     
- Misses         12       13       +1     
- Partials        8        9       +1     
Impacted Files Coverage Δ
ocrd_calamari/recognize.py 86.50% <100.00%> (-1.23%) :arrow_down:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update ecb0d67...36e5136. Read the comment docs.

mikegerber commented 2 years ago

I haven't actually tested it (I go on vacation tomorrow), but tests run successful so this seems to OK to merge.

@bertsky I'll prepare a new release after my vacation