Reverse engineer a string masking/obfuscating function

I would like to unmask or unobfuscate a piece of base64 encoded string containing some characters.

For instance, I know VovL5SaV_iSBx6oAFGypsTCO matches 00000000. Is there a way, knowing more data (the input and the output) to find out the algorithm that obfuscate the string? I couldn't find anything.

Some example:

• The result of f(A) will vary in length if A is longer f(A) will be longer, if A is shorter f(A) will be shorter.
• If we have A and B with a similar ending (e.g. aaaaa@gmail.com and bbbbb@gmail.com), then f(A) and f(B) will also have the same ending.