How to compare the result in terms on word and letter accuracy using python

[Ido]

I am running a tesseract custom trained model using synthetic data I have created.

and I am trying to test it using pytesseract  on my real data (after a pre-processing phase implemented in python ).

can some one recommend  a python existing model in order to calculate the error  in term of letter and word error?