Fix offset mapping for uie-x

[linjieccc]

PR types

Bug fixes

PR changes

APIs

Description

• 修复Tokenizer对单一字符进行处理后 offset_mapping无法对齐的问题 [Bug]: uie文档抽取，索引越界 #4656

[paddle-bot]

Thanks for your contribution!

[codecov]

Codecov Report

Merging #4687 (0b03a0f) into develop (67d82fd) will increase coverage by 0.00%.
The diff coverage is 0.00%.

@@           Coverage Diff            @@
##           develop    #4687   +/-   ##
========================================
  Coverage    42.08%   42.09%           
========================================
  Files          434      434           
  Lines        61870    61915   +45     
========================================
+ Hits         26039    26063   +24     
- Misses       35831    35852   +21     

Impacted Files	Coverage Δ
paddlenlp/taskflow/information_extraction.py	8.08% <0.00%> (-0.03%)	⬇️
paddlenlp/transformers/image_processing_utils.py	24.41% <0.00%> (-1.59%)	⬇️
paddlenlp/transformers/mbart/tokenizer.py	89.69% <0.00%> (-0.78%)	⬇️
paddlenlp/transformers/t5/tokenizer.py	91.26% <0.00%> (-0.66%)	⬇️
paddlenlp/transformers/clip/tokenizer.py	95.30% <0.00%> (-0.57%)	⬇️
paddlenlp/transformers/model_utils.py	77.40% <0.00%> (-0.35%)	⬇️
paddlenlp/transformers/tokenizer_utils_base.py	78.14% <0.00%> (-0.24%)	⬇️
paddlenlp/transformers/tokenizer_utils.py	63.08% <0.00%> (+0.14%)	⬆️
paddlenlp/utils/downloader.py	62.99% <0.00%> (+0.44%)	⬆️
paddlenlp/transformers/ernie_m/tokenizer.py	82.14% <0.00%> (+0.71%)	⬆️
... and 4 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[sijunhe]

information_extraction taskflow这里逻辑很重，希望能加上单测

[linjieccc]

information_extraction taskflow这里逻辑很重，希望能加上单测

好的，后续会针对uie相关功能增加下对应单测