Minor bug fixes

• Fix date correction via url

• Support python 3.10 whl

• Use url to validate publish date to prevent website traffic test

• Remove error when html blocks are limited (< 3) and raise Blockify error

• Bug fixes includes dozen of metadata parsing syntax error

• Fix video extraction
• Minor bug fix
• Fix tar.gz missing pyx file

Major changes include:

• New backend model : replace catboost with neural based model with 2x speedup and better author and headline extraction

• Rewrite processing code for better readability

• Reduce installation footprint by 10x

Added pipeline parameters to do post processing after meta data extraction and ML data extraction phase. Users should be able to add custom callbacks functions to modify results in these stages.

First release version

Initial release for first alpha review version, we will start internal testing before pushing the final release