Set input encoding to UTF-8 to fix parsing issues.

[SamFlt]

This PR tries to fix #1391, by forcing utf-8 decoding when reading headers before generating the bindings.

The issue can be seen in the generation log:

ERROR:root:Traceback (most recent call last):
  File "E:\user\visp-conda-ws\lib\site-packages\visp_python_bindgen\generator.py", line 54, in header_preprocess
    header.preprocess()
  File "E:\user\visp-conda-ws\lib\site-packages\visp_python_bindgen\header.py", line 110, in preprocess
    self.preprocessed_header_str = self.run_preprocessor() # Run preprocessor, get only code that can be compiled with current visp
  File "E:\uservisp-conda-ws\lib\site-packages\visp_python_bindgen\header.py", line 148, in run_preprocessor
    for line in input_header_file.readlines():
UnicodeDecodeError: 'gbk' codec can't decode byte 0x99 in position 4482: illegal multibyte sequence

ERROR:root:<class 'UnicodeDecodeError'>

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 54.00%. Comparing base (0a5ef8d) to head (c14edae).
Report is 80 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1470      +/-   ##
==========================================
+ Coverage   45.25%   54.00%   +8.74%     
==========================================
  Files        1121      442     -679     
  Lines      112131    50841   -61290     
  Branches    19077        0   -19077     
==========================================
- Hits        50750    27457   -23293     
+ Misses      61381    23384   -37997     

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