Jesse_Chen
45 lines
1.7 KiB
Python
45 lines
1.7 KiB
Python
from scripts.minute_candidate_discriminability import analyze_candidate_rows, feature_fingerprint
|
|
|
|
|
|
def _row(time: str, points: float, rules: list[str] | None = None) -> dict:
|
|
return {
|
|
"time": time,
|
|
"score": points,
|
|
"evidence": [{
|
|
"event_id": "event-1",
|
|
"domain": "career",
|
|
"candidate_time": time,
|
|
"rule_ids": rules or ["same_rule"],
|
|
"points": points,
|
|
}],
|
|
"missing_layers": [],
|
|
}
|
|
|
|
|
|
def test_fingerprint_excludes_candidate_time_but_includes_evidence_features() -> None:
|
|
assert feature_fingerprint(_row("10:00", 1.0)) == feature_fingerprint(_row("10:01", 1.0))
|
|
assert feature_fingerprint(_row("10:00", 1.0)) != feature_fingerprint(_row("10:00", 2.0))
|
|
|
|
|
|
def test_diagnostic_reports_equivalent_adjacent_minutes() -> None:
|
|
report = analyze_candidate_rows([
|
|
_row("10:00", 1.0),
|
|
_row("10:01", 1.0),
|
|
_row("10:02", 2.0, ["different_rule"]),
|
|
])
|
|
|
|
assert report["unique_feature_fingerprint_count"] == 2
|
|
assert report["indistinguishable_adjacent_pair_count"] == 1
|
|
assert report["adjacent_transitions"][0]["feature_changed"] is False
|
|
assert report["adjacent_transitions"][0]["changed_event_ids"] == []
|
|
assert report["top_candidate_feature_unique"] is True
|
|
assert report["status"] == "minute_feature_unique"
|
|
|
|
|
|
def test_diagnostic_blocks_a_fully_equivalent_range() -> None:
|
|
report = analyze_candidate_rows([_row("10:00", 1.0), _row("10:01", 1.0)])
|
|
|
|
assert report["distinguishable_candidate_ratio"] == 0.5
|
|
assert report["top_candidate_feature_unique"] is False
|
|
assert report["status"] == "blocked_feature_equivalent_range"
|