qurator-spk/page2tsv
1
0
Fork 0
mirror of https://github.com/qurator-spk/page2tsv.git synced 2025-10-30 16:14:13 +01:00
Code Issues Releases Wiki Activity

remove spam

Browse source
This commit is contained in:
Kai Labusch 2024-11-12 14:46:57 +01:00
parent ef8244a466
commit 8d8bf517b9
1 changed files with 0 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

7
qurator/tsvtools/cli.py
View file

@ -346,13 +346,6 @@ def tsv2tsv(tsv_in_file, tsv_out_file, ner_rest_endpoint, noproxy,
prev_pos = 0 prev_pos = 0
for idx, _ in tsv_out.iterrows(): for idx, _ in tsv_out.iterrows():
# if idx < len(tsv_out) and len(tsv_out.loc[idx, 'TOKEN']) == 0 and tsv_out.loc[idx+1, 'No.'] == 0:
# print("word_pos=0!!!!")
# word_pos = 0
#
# if 0 < tsv_out.loc[idx, 'No.'] < word_pos:
# word_pos = 0
if prev_pos != 0 and not tsv_out.loc[idx, 'NE-TAG'].startswith('I-') and \ if prev_pos != 0 and not tsv_out.loc[idx, 'NE-TAG'].startswith('I-') and \
tsv_out.loc[idx, 'No.'] == 0 or len(tsv_out.loc[idx, 'TOKEN']) == 0: tsv_out.loc[idx, 'No.'] == 0 or len(tsv_out.loc[idx, 'TOKEN']) == 0:
word_pos = 0 word_pos = 0