from __future__ import annotations

import pandas as pd

from apps.tools.build_ranking_model import build_model


def test_build_model_normalizes_reaction_date_strings(tmp_path) -> None:
    frame = pd.DataFrame(
        [
            {
                "event_date": "2024-12-30",
                "reaction_date": "2024-12-30",
                "event_type": "earnings_release",
                "reaction_day_return": 0.08,
                "fwd_return_20d": 0.12,
                "event_direction": "bullish",
                "guidance_status": "raised",
                "close_location": 0.72,
                "volume_ratio_20d": 2.2,
                "gap_size": 0.03,
                "document_quality_score": 0.81,
                "parse_confidence_overall": 0.82,
            },
            {
                "event_date": "2025-01-02",
                "reaction_date": "2025-01-02",
                "event_type": "earnings_release",
                "reaction_day_return": 0.09,
                "fwd_return_20d": 0.10,
                "event_direction": "bullish",
                "guidance_status": "raised",
                "close_location": 0.75,
                "volume_ratio_20d": 2.4,
                "gap_size": 0.02,
                "document_quality_score": 0.83,
                "parse_confidence_overall": 0.84,
            },
        ]
    )
    snapshot_path = tmp_path / "train.parquet"
    frame.to_parquet(snapshot_path)

    model = build_model(snapshot_path, cutoff_event_date="2024-12-31", min_bucket_count=1)

    assert model["training_rows"] == 1
    assert model["global_mean"] == 0.12
    event_type_feature = next(feature for feature in model["features"] if feature["name"] == "event_type")
    assert event_type_feature["values"]["earnings_release"] == 0.12