changes upper and lower bounds for inference grading

changed grader to return 0 or 1

fixed uv module activation issue in inference.py during test evaluation

black format and isort code

fixed issues in SQL schema not available to agent and code parsing issues

added 2 new hard tasks 1 medium and grader for each, SQL DB mock data for hard task

fixed action parsing issues in inference.py and increased concurrent env in server/app.py

added configurations for black and formatted code

first commit