Update code with new Data Pool

[trungthanhnguyen0502]

OVERVIEW:

1. Change the task generation mechanism. Validator gets random tasks from Data Pool Server. Remove Mathgenerator, HuggingFace dataset from ValidatorCode
2. Tasks will be continuously added and refreshed to avoid miners trying to solve OVERFIT on a certain dataset or math library. Miners should try to build a comprehensive solution that handles a variety of math problems, not only Mathgenerator Lib or any public dataset.
3. The implementation of DataPool server also aims to prevent fake validators
4. Update the number and frequency of tasks that validators send to miners, to save API/LLM costs for both miners and validators
5. Fixbug EvenLoop when use multi-thread + asynceIO (used by bittensor)
6. Refactor code
7. Add sample config file to run with PM2 Tools for miner/validator