"""Result payload lookup and GPR display filtering."""

from __future__ import annotations

import numpy as np

from python_app.models.dataset_model import ResultCollection, ResultPayload


def collection_payload_by_name(
    collection: ResultCollection,
    name: str,
    kind: int | None = None,
) -> ResultPayload | None:
    """Return the first collection payload matching name and optional kind."""
    for payload in collection.collection_payloads:
        if payload.processing_name != name:
            continue
        if kind is not None and int(payload.kind) != int(kind):
            continue
        return payload
    return None


def collection_payloads_by_prefix(
    collection: ResultCollection,
    prefix: str,
    kind: int | None = None,
) -> list[ResultPayload]:
    """Return collection payloads matching a processing-name prefix."""
    payloads: list[ResultPayload] = []
    for payload in collection.collection_payloads:
        if not str(payload.processing_name).startswith(prefix):
            continue
        if kind is not None and int(payload.kind) != int(kind):
            continue
        payloads.append(payload)
    return payloads


def gpr_object_rows(collection: ResultCollection) -> np.ndarray:
    """Return object rows as `[x_m, z_m, score]` from a GPR collection."""
    points_payload = collection_payload_by_name(collection, "gpr_points", kind=4)
    if points_payload is not None:
        points = np.asarray(points_payload.table, dtype=np.float32)
        if points.ndim == 2 and points.shape[1] >= 3:
            return points[:, :3]

    centers_payload = collection_payload_by_name(collection, "gpr_region_centers", kind=4)
    if centers_payload is not None:
        centers = np.asarray(centers_payload.table, dtype=np.float32)
        if centers.ndim == 2 and centers.shape[1] >= 3:
            return centers[:, :3]

    return np.zeros((0, 3), dtype=np.float32)


def filter_object_rows(
    rows: np.ndarray,
    *,
    min_score: float,
    x_bounds: tuple[float, float],
    z_bounds: tuple[float, float],
) -> np.ndarray:
    """Keep finite objects inside the visible window and above the score threshold."""
    if rows.size == 0:
        return rows
    x_min, x_max = x_bounds
    z_min, z_max = z_bounds
    visible_mask = (
        np.all(np.isfinite(rows[:, :3]), axis=1)
        & (rows[:, 2] >= min_score)
        & (rows[:, 0] >= x_min)
        & (rows[:, 0] <= x_max)
        & (rows[:, 1] >= z_min)
        & (rows[:, 1] <= z_max)
    )
    return rows[visible_mask]