xiaoxia-saas/tests/unit/test_1970_narrative_match.py

"""#1970 PR3 叙事模式文案标签匹配纯函数测试。"""

from __future__ import annotations

import random
from dataclasses import dataclass, field

import pytest

from packages.domain.narrative_match import (
    build_asset_tag_name_index,
    match_assets_by_script_tags,
    normalize_tag,
    pick_narrative_assets,
)


@dataclass
class FakeAsset:
    id: str
    tag_ids: list[str] = field(default_factory=list)
    tags: list[str] = field(default_factory=list)
    status: str = "ready"
    file_type: str = "video"
    duration: float = 10.0
    quality_score: float | None = None
    created_at: object = None
    metadata: dict = field(default_factory=dict)


# ── normalize_tag ──────────────────────────────────────────────────────────


class TestNormalizeTag:
    def test_strip_and_lower(self):
        assert normalize_tag("  带货 ") == "带货"
        assert normalize_tag("Factory") == "factory"

    def test_none_and_non_string(self):
        assert normalize_tag(None) == ""
        assert normalize_tag(123) == "123"

    def test_short_tag_filtered_by_normalize_set(self):
        # 单字噪声标签不参与匹配（_normalize_tags 层过滤）
        from packages.domain.narrative_match import _normalize_tags

        assert _normalize_tags(["的", " a ", "工厂"]) == {"工厂"}


# ── match_assets_by_script_tags ────────────────────────────────────────────


class TestMatchSplit:
    def test_split_by_tag_names(self):
        assets = [
            FakeAsset("a1", tags=["工厂"]),
            FakeAsset("a2", tags=["旅游"]),
            FakeAsset("a3", tags=["工厂", "车间"]),
        ]
        matched, unmatched = match_assets_by_script_tags(assets, script_tags=["工厂"])
        assert [a.id for a in matched] == ["a1", "a3"]
        assert [a.id for a in unmatched] == ["a2"]

    def test_case_insensitive(self):
        assets = [FakeAsset("a1", tags=["Factory"])]
        matched, unmatched = match_assets_by_script_tags(assets, script_tags=["FACTORY"])
        assert [a.id for a in matched] == ["a1"]
        assert unmatched == []

    def test_tag_ids_via_name_index(self):
        assets = [FakeAsset("a1", tag_ids=["t1"]), FakeAsset("a2", tag_ids=["t2"])]
        index = {"a1": ["测评"], "a2": ["vlog"]}
        matched, unmatched = match_assets_by_script_tags(assets, script_tags=["测评"], tag_names_by_id=index)
        assert [a.id for a in matched] == ["a1"]
        assert [a.id for a in unmatched] == ["a2"]

    def test_empty_script_tags_degrades_all_unmatched(self):
        assets = [FakeAsset("a1", tags=["工厂"])]
        matched, unmatched = match_assets_by_script_tags(assets, script_tags=[])
        assert matched == []
        assert [a.id for a in unmatched] == ["a1"]

    def test_no_match_degrades(self):
        assets = [FakeAsset("a1", tags=["工厂"]), FakeAsset("a2", tags=["车间"])]
        matched, unmatched = match_assets_by_script_tags(assets, script_tags=["美食"])
        assert matched == []
        assert {a.id for a in unmatched} == {"a1", "a2"}

    def test_order_preserved(self):
        assets = [FakeAsset(f"a{i}", tags=["x" if i % 2 else "工厂"]) for i in range(6)]
        matched, _ = match_assets_by_script_tags(assets, script_tags=["工厂"])
        assert [a.id for a in matched] == ["a0", "a2", "a4"]

    def test_build_index_ignores_blank(self):
        # 空白/None/单字符噪声标签均不参与匹配
        idx = build_asset_tag_name_index({"a1": [" 工厂 ", "", None, "A"]})
        assert idx == {"a1": {"工厂"}}


# ── pick_narrative_assets ──────────────────────────────────────────────────


class TestPickNarrativeAssets:
    def _assets(self):
        # smart_match 需要 created_at（None 走 recency 兜底）
        import datetime as dt

        old = dt.datetime(2020, 1, 1, tzinfo=dt.UTC)
        return [
            FakeAsset("match1", tags=["工厂"], created_at=old),
            FakeAsset("nomatch1", tags=["旅游"], created_at=old),
            FakeAsset("match2", tags=["工厂"], created_at=old),
            FakeAsset("nomatch2", tags=["美食"], created_at=old),
        ]

    def test_matched_pool_prioritized(self):
        picked = pick_narrative_assets(self._assets(), script_tags=["工厂"], limit=2, rng=random.Random(0))
        assert {a.id for a in picked} <= {"match1", "match2"}
        assert all(a.id.startswith("match") for a in picked)

    def test_fallback_fills_from_unmatched(self):
        picked = pick_narrative_assets(self._assets(), script_tags=["工厂"], limit=4, rng=random.Random(0))
        ids = {a.id for a in picked}
        assert ids == {"match1", "match2", "nomatch1", "nomatch2"}
        # 命中池排在前面
        assert picked[0].id.startswith("match")
        assert picked[1].id.startswith("match")

    def test_no_tag_match_equals_random_selection(self):
        assets = self._assets()
        picked = pick_narrative_assets(assets, script_tags=["不存在"], limit=3, rng=random.Random(42))
        assert len(picked) == 3

    def test_empty_tags_selects_all_pool(self):
        assets = self._assets()
        picked = pick_narrative_assets(assets, script_tags=[], limit=None, rng=random.Random(1))
        assert len(picked) == 4

    def test_limit_none_returns_all_with_matched_first(self):
        picked = pick_narrative_assets(self._assets(), script_tags=["工厂"], limit=None, rng=random.Random(1))
        assert len(picked) == 4
        assert {a.id for a in picked[:2]} == {"match1", "match2"}

    def test_tag_ids_index_path(self):
        assets = [FakeAsset("a1", tag_ids=["t1"]), FakeAsset("a2", tag_ids=["t2"])]
        # 补 created_at
        import datetime as dt

        for a in assets:
            a.created_at = dt.datetime(2020, 1, 1, tzinfo=dt.UTC)
        picked = pick_narrative_assets(
            assets,
            script_tags=["教程"],
            tag_names_by_id={"a1": ["教程"], "a2": ["旅游"]},
            limit=1,
            rng=random.Random(0),
        )
        assert [a.id for a in picked] == ["a1"]

    def test_deterministic_with_seed(self):
        r1 = pick_narrative_assets(self._assets(), script_tags=["工厂"], limit=4, rng=random.Random(7))
        r2 = pick_narrative_assets(self._assets(), script_tags=["工厂"], limit=4, rng=random.Random(7))
        assert [a.id for a in r1] == [a.id for a in r2]


if __name__ == "__main__":
    pytest.main([__file__, "-q"])