Jesse_Chen
151 lines
4.3 KiB
Python
151 lines
4.3 KiB
Python
from __future__ import annotations
|
|
|
|
from scripts.active_rectification_events import (
|
|
CandidateScoreRow,
|
|
adjudicate_candidate_rows,
|
|
precision_weight,
|
|
score_life_events,
|
|
)
|
|
|
|
|
|
def _row(time: str, score: float) -> CandidateScoreRow:
|
|
return {
|
|
"time": time,
|
|
"score": score,
|
|
"evidence": [],
|
|
"missing_layers": [],
|
|
}
|
|
|
|
|
|
def test_high_confidence_requires_four_events_three_domains_and_narrow_leader() -> None:
|
|
result = adjudicate_candidate_rows(
|
|
[
|
|
_row("14:22", 16),
|
|
_row("14:23", 16),
|
|
_row("14:24", 16),
|
|
_row("14:25", 16),
|
|
_row("14:26", 16),
|
|
_row("14:27", 10),
|
|
],
|
|
event_count=4,
|
|
domain_count=3,
|
|
request_fingerprint="high-fixture",
|
|
)
|
|
|
|
assert result["confidence"] == "high"
|
|
assert result["can_apply"] is True
|
|
assert result["winning_segment"] == {
|
|
"start_time": "14:22",
|
|
"end_time": "14:26",
|
|
"representative_time": "14:24",
|
|
"width_minutes": 5,
|
|
}
|
|
assert result["margin_percent"] == 37.5
|
|
|
|
|
|
def test_tied_disjoint_candidates_abstain() -> None:
|
|
result = adjudicate_candidate_rows(
|
|
[_row("14:20", 10), _row("14:21", 8), _row("14:22", 10)],
|
|
event_count=4,
|
|
domain_count=3,
|
|
request_fingerprint="tie-fixture",
|
|
)
|
|
|
|
assert result["confidence"] == "low"
|
|
assert result["can_apply"] is False
|
|
assert "tied_leader" in result["reasons"]
|
|
assert result["winning_segment"] is None
|
|
|
|
|
|
def test_medium_confidence_never_allows_application() -> None:
|
|
result = adjudicate_candidate_rows(
|
|
[_row("14:20", 10), _row("14:21", 8)],
|
|
event_count=3,
|
|
domain_count=2,
|
|
request_fingerprint="medium-fixture",
|
|
)
|
|
|
|
assert result["confidence"] == "medium"
|
|
assert result["can_apply"] is False
|
|
assert result["winning_segment"]["representative_time"] == "14:20"
|
|
|
|
|
|
def test_result_keeps_only_representative_minute_evidence() -> None:
|
|
rows = [_row("14:20", 10), _row("14:21", 10), _row("14:22", 10), _row("14:23", 5)]
|
|
for row in rows:
|
|
row["evidence"] = [{
|
|
"event_id": "5cb071d6-6d99-46be-85dc-a9bf59ef6ac5",
|
|
"domain": "education",
|
|
"candidate_time": row["time"],
|
|
"rule_ids": ["fixture"],
|
|
"points": row["score"],
|
|
}]
|
|
|
|
result = adjudicate_candidate_rows(
|
|
rows,
|
|
event_count=3,
|
|
domain_count=2,
|
|
request_fingerprint="representative-evidence-fixture",
|
|
)
|
|
|
|
assert [item["candidate_time"] for item in result["evidence"]] == ["14:21"]
|
|
|
|
|
|
def test_missing_mandatory_layer_caps_confidence_at_low() -> None:
|
|
row = _row("14:20", 10)
|
|
row["missing_layers"] = ["D24"]
|
|
|
|
result = adjudicate_candidate_rows(
|
|
[row, _row("14:21", 5)],
|
|
event_count=4,
|
|
domain_count=3,
|
|
request_fingerprint="missing-layer-fixture",
|
|
)
|
|
|
|
assert result["confidence"] == "low"
|
|
assert "missing_mandatory_layers" in result["reasons"]
|
|
assert result["can_apply"] is False
|
|
|
|
|
|
def test_date_precision_weights_are_fixed() -> None:
|
|
assert precision_weight("day") == 1.0
|
|
assert precision_weight("month") == 0.8
|
|
assert precision_weight("year") == 0.5
|
|
|
|
|
|
def test_real_local_scoring_uses_dated_events_and_actual_candidate_minutes() -> None:
|
|
result = score_life_events({
|
|
"birth_date": "1993-04-17",
|
|
"start_time": "14:29",
|
|
"end_time": "14:31",
|
|
"lat": 36.683333,
|
|
"lon": 114.35,
|
|
"tz": 8.0,
|
|
"events": [
|
|
{
|
|
"id": "5cb071d6-6d99-46be-85dc-a9bf59ef6ac5",
|
|
"domain": "education",
|
|
"date": "2011-09",
|
|
"precision": "month",
|
|
},
|
|
{
|
|
"id": "0790866c-ad5e-4a45-b2b4-a5c73f6be6ea",
|
|
"domain": "career",
|
|
"date": "2019-07-01",
|
|
"precision": "day",
|
|
},
|
|
{
|
|
"id": "0ef52e51-ab5f-453b-81e5-adb44a929224",
|
|
"domain": "relationship",
|
|
"date": "2021",
|
|
"precision": "year",
|
|
},
|
|
],
|
|
})
|
|
|
|
assert result["result_id"]
|
|
assert result["event_count"] == 3
|
|
assert result["domain_count"] == 3
|
|
assert result["algorithm_version"] == "birth-time-event-scoring-v1"
|
|
assert result["confidence"] in {"low", "medium"}
|