Update scenario mining evaluation.

[CainanD]

PR Summary

• Updated scenario mining evaluation to accept both temporal and spatio-temporal predictions in the same format.
• Changed tracking evaluation to evaluate thresholds in parallel.
• Temporal evaluation is 100x faster and spatio-temporal evaluation is 3x faster.
• Fixed balanced accuracy metric calculation

Testing

• Tested with tests/unit/evaluation/scenario_mining/test_eval.py.
• Tested the fork on EvalAI Scenario Mining and it produces the same metrics at 3x faster speeds.

In order to ensure this PR works as intended, it is:

• [yes] unit tested.
• other or not applicable (additional detail/rationale required)

Compliance with Standards

As the author, I certify that this PR conforms to the following standards:

• [yes] Code changes conform to PEP8 and docstrings conform to the Google Python style guide.
• [yes] A well-written summary explains what was done and why it was done.
• [yes] The PR is adequately tested and the testing details and links to external results are included.