fix(evaluation): resolve PR gate failures (lint, CI extras, remove flyradar example)

[miguelgfierro]

What failed

The PR gate for #279 failed on Lint, Type Check, Test, and Pre-commit.

Fixes

Remove flyradar example (per user instruction):

• Deleted examples/flyradar_eval_example.py

CI: add --extra evaluation to sync steps:

• typecheck job: uv sync now includes --extra evaluation
• test job: uv sync now includes --extra evaluation
• Without this, numpy is absent and all evaluation tests fail with ModuleNotFoundError

Ruff lint fixes across fireflyframework_agentic/evaluation/ and examples/:

• I001 — sorted all unsorted import blocks (__init__.py, cli.py, matcher.py, judge.py, judge_client.py, scorecard.py, corpus.py, gates.py, run_config_snapshot.py)
• F401 — removed unused compute_retrieval_metrics import in flycanon_eval_example.py
• SIM108 — replaced if/else block with ternary in flycanon_eval_example.py
• B905 — added strict=False to all bare zip() calls in judge.py
• N806 — lowercased PER_SURFACE_CAP and BLEND_PHRASES function-local constants
• UP035 — replaced from typing import Sequence with from collections.abc import Sequence in stats.py
• SIM102 — collapsed nested if into single and in matcher.py