Someone who has good knowledge about Adaboost boosting

I have a dataset for multi classification. You need to implement Adaboost from scratch which means using no direct skicit methods. Than you should use one-vs-all encoding to get from binary classification to multi classificaion (7 classes in our case)