fix: fix false invalidation when decoding token User Ids (#1278)
* add a util method for padding base64 strings if they are not of an expected length
* return the original string if it already contains padding, do not throw
* add tests for padding method, and for token that needs padding
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.