Normalize GitHub IDs for miner matching

[hoangtrung1108]

Summary

• Normalize miner GitHub IDs to strings when building duplicate-detection buckets.
• Normalize issue-bounty registered miner lookup keys so int and string IDs match consistently.
• Add regressions for mixed int/string duplicate detection and bounty solver lookup.

Fixes #1413

Tests

• PYTHONPATH=C:\Users\49159\money-engine\test_stubs;C:\Users\49159\money-engine\bounties\gittensor .\.venv\Scripts\python -m pytest tests\validator\test_github_id_normalization.py tests\validator\test_issue_competitions_forward.py::test_int_github_id_miner_receives_solution_vote -q
• .\.venv\Scripts\python -m ruff check gittensor\validator\oss_contributions\inspections.py gittensor\validator\issue_competitions\forward.py tests\validator\test_github_id_normalization.py tests\validator\test_issue_competitions_forward.py

Note: full dependency install on this Windows host failed because bittensor native dependencies require the MSVC linker (link.exe), so the focused tests were run with a minimal import stub for external bittensor modules.