qurator-spk/dinglehopper
1
0
Fork 0
mirror of https://github.com/qurator-spk/dinglehopper.git synced 2025-07-27 19:29:55 +02:00
Code Issues Releases Wiki Activity

🧹 dinglehopper: Remove broken implementation of the unordered word error rate

Browse source
This commit is contained in:
Gerber, Mike 2019-11-18 15:03:17 +01:00
parent f22228840e
commit 11a6341641
4 changed files with 2 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

3
qurator/dinglehopper/cli.py
View file

@ -56,7 +56,6 @@ def process(gt, ocr, report_prefix):
cer = character_error_rate(gt_text, ocr_text) cer = character_error_rate(gt_text, ocr_text)
wer = word_error_rate(gt_text, ocr_text) wer = word_error_rate(gt_text, ocr_text)
uwer = unordered_word_error_rate(gt_text, ocr_text)
char_diff_report = gen_diff_report(gt_text, ocr_text, css_prefix='c', joiner='', none='·', align=align) char_diff_report = gen_diff_report(gt_text, ocr_text, css_prefix='c', joiner='', none='·', align=align)
@ -72,7 +71,7 @@ def process(gt, ocr, report_prefix):
template = env.get_template(template_fn) template = env.get_template(template_fn)
template.stream( template.stream(
gt=gt, ocr=ocr, gt=gt, ocr=ocr,
cer=cer, wer=wer, uwer=uwer, cer=cer, wer=wer,
char_diff_report=char_diff_report, char_diff_report=char_diff_report,
word_diff_report=word_diff_report word_diff_report=word_diff_report
).dump(out_fn) ).dump(out_fn)

1
qurator/dinglehopper/templates/report.html.j2
View file

@ -35,7 +35,6 @@
<h2>Metrics</h2> <h2>Metrics</h2>
<p>CER: {{ cer|round(4) }}</p> <p>CER: {{ cer|round(4) }}</p>
<p>WER: {{ wer|round(4) }}</p> <p>WER: {{ wer|round(4) }}</p>
<!-- FIXME <p>WER (unordered): {{ uwer|round(4) }}</p> -->
<h2>Character differences</h2> <h2>Character differences</h2>
{{ char_diff_report }} {{ char_diff_report }}

10
qurator/dinglehopper/tests/test_word_error_rate.py
View file

@ -2,7 +2,7 @@ from __future__ import division, print_function
import math import math
from .. import word_error_rate, words, unordered_word_error_rate from .. import word_error_rate, words
def test_words(): def test_words():
@ -35,11 +35,3 @@ def test_word_error_rate():
assert word_error_rate('', '') == 0 assert word_error_rate('', '') == 0
assert word_error_rate('Schlyñ lorem ipsum dolor sit amet,', 'Schlym̃ lorem ipsum dolor sit amet.') == 1/6 assert word_error_rate('Schlyñ lorem ipsum dolor sit amet,', 'Schlym̃ lorem ipsum dolor sit amet.') == 1/6
def test_unordered_word_error_rate():
assert unordered_word_error_rate('abc def ghi', 'ghi abc def') == 0
assert unordered_word_error_rate('abc def ghi', 'ghi abcX def') == 1/3
assert unordered_word_error_rate('abc def ghi jkl', 'abc ghi def jkl') == 0
assert unordered_word_error_rate('abc def ghi jkl', 'abc ghi defX jkl') == 1/4
# XXX There seem to be some cases where this does not work

6
qurator/dinglehopper/word_error_rate.py
View file

@ -61,9 +61,3 @@ def word_error_rate(reference, compared):
return float('inf') return float('inf')
return d / n return d / n
def unordered_word_error_rate(reference, compared):
reference_seq = sorted(words_normalized(reference))
compared_seq = sorted(words_normalized(compared))
return word_error_rate(reference_seq, compared_seq)