[Feature Request]: Consider switching from pandas to polars for performance improvements

[bgrnwd]

What feature would you like to see?

https://docs.pola.rs/user-guide/getting-started/

[bgrnwd]

Doing some basic profiling with cProfile. Testing the code from the README

client = AmericanSoccerAnalysis()
asa_players = client.get_players(names="Andre")

Pandas

   240613 function calls (238081 primitive calls) in 21.651 seconds

   Ordered by: cumulative time
   List reduced from 1475 to 10 due to restriction <10>

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
      2/1    0.000    0.000   21.651   21.651 {built-in method builtins.exec}
        1    0.000    0.000   21.651   21.651 <string>:1(<module>)
        1    0.000    0.000   21.650   21.650 client.py:477(get_players)
        1    0.004    0.004   21.640   21.640 client.py:70(_get_entity)
        6    0.001    0.000   21.588    3.598 client.py:260(_execute_query)
       13    0.009    0.001   21.549    1.658 client.py:290(_single_request)
       13    0.000    0.000   21.117    1.624 sessions.py:593(get)
       13    0.000    0.000   21.116    1.624 sessions.py:502(request)
       13    0.000    0.000   20.950    1.612 sessions.py:673(send)
       13    0.000    0.000   20.679    1.591 adapter.py:50(send)

Polars

[bgrnwd]

Running in to an issue when switching over from pandas to polars.

SchemaError: failed to determine supertype of struct[9] and struct[4]

It seems like polars is running in to an issue reading the JSON structure returned by the API and converting that in to a DataFrame.