[🧹 CHORE]: Rewrite multipart text values parsing algorithm

[rustatian]

No duplicates 🥲.

• [X] I have searched for a similar issue.

What should be improved or cleaned up?

Currently, the multipart parsing algorithm is not efficient enough since it involves several steps to search the data inside the multipart request. Moreover, it prevents for using protobuf since it has an any type inside, which is not presented in proto. I think we can use a stack to parse this data structure in one pass - O(n).