HeXi/tests/test_video_media_naming.py

"""媒体命名 / S3 key (utils.py 纯函数) 单元测试

utils.py 只依赖标准库，故用 importlib 按文件路径裸加载（与
test_video_group_file.py 同法）。media_key_of / media_rel_dir_of 依赖
get_temp_root 的返回值，测试里 monkeypatch 成临时目录。
"""

import importlib.util
import sys
from pathlib import Path

import pytest

_MODULE_PATH = (
    Path(__file__).resolve().parents[1]
    / "hexi"
    / "plugins"
    / "nonebot_plugin_video_analysis"
    / "utils.py"
)


@pytest.fixture(scope="module")
def u():
    """以独立模块名加载 utils.py，避免与插件包 __init__ 冲突"""
    spec = importlib.util.spec_from_file_location("video_utils_under_test", _MODULE_PATH)
    module = importlib.util.module_from_spec(spec)
    sys.modules[spec.name] = module
    spec.loader.exec_module(module)
    yield module
    sys.modules.pop(spec.name, None)


@pytest.fixture
def temp_root(u, tmp_path, monkeypatch):
    """把 get_temp_root 指到 tmp_path（media_key_of 只做路径解析，不需 mkdir）"""
    root = tmp_path / "temp"
    root.mkdir()
    monkeypatch.setattr(u, "get_temp_root", lambda sub="": root / sub if sub else root)
    return root


# ───────────────────────── 作者目录 ─────────────────────────


def test_author_dir_with_id(u):
    """有作者 id 就是稳定的 `{昵称}_{id}`，不带码"""
    assert u.build_author_dir("示例作者", "12345") == "示例作者_12345"
    assert u.build_author_dir("示例作者", "12345", source="x", at=1) == (
        "示例作者_12345"
    )


def test_author_dir_nickname_only_gets_time_code(u):
    """只有昵称 → 追加 4 位时间短码：分不清同名作者，宁可不聚合也不混目录"""
    at = 1_700_000_000
    code = u.short_time_code(at)
    assert u.build_author_dir("示例作者", "", at=at) == f"示例作者_{code}"
    assert u.build_author_dir("示例作者", at=at) == f"示例作者_{code}"
    for junk in ("0", "NA", "none", "None", " null "):
        assert u.build_author_dir("示例作者", junk, at=at) == f"示例作者_{code}"

    # 不同时刻 = 不同目录（同一作者的不同作品不会互相覆盖）
    other = u.build_author_dir("示例作者", at=at + 3600)
    assert other != f"示例作者_{code}"
    assert other.startswith("示例作者_") and len(other.split("_")[-1]) == 4


def test_author_dir_uses_source_code_when_anonymous(u):
    """连昵称都没有 → 未知作者 + 来源短码（同一来源稳定，不同来源不撞）"""
    url = "https://www.douyin.com/video/7412345678901234567"
    first = u.build_author_dir(None, None, source=url)
    assert first.startswith("未知作者_")
    assert first == u.build_author_dir("", "", source=url)  # 同来源稳定

    second = u.build_author_dir(None, None, source=url + "8")
    assert second != first

    # 没有来源时才退化成时间短码
    at = 1_700_000_000
    assert u.build_author_dir(None, None, at=at) == (
        f"未知作者_{u.short_time_code(at)}"
    )


def test_author_dir_treats_placeholder_nicknames_as_anonymous(u):
    """抓取层拿不到昵称时的兜底串不算作者名，走来源码"""
    url = "https://www.bilibili.com/opus/123"
    expected = u.build_author_dir(None, None, source=url)
    for placeholder in ("未知作者", "小红书用户", "B站用户", "b站用户"):
        assert u.build_author_dir(placeholder, "", source=url) == expected
    # 纯 emoji 昵称 slugify 后为空 → 同样按匿名处理
    assert u.build_author_dir("🎬🎬", "", source=url) == expected


def test_short_source_code_is_stable_base36(u):
    url = "https://www.douyin.com/video/7412345678901234567"
    code = u.short_source_code(url)
    assert len(code) == 4 and code.isalnum()
    assert code == u.short_source_code(url)  # 可复现（不能用带盐的 hash()）
    assert code != u.short_source_code(url + "8")


def test_author_dir_sanitizes_illegal_chars(u):
    """Windows 非法字符不能进目录名（slugify 直接删掉，空白转 -）"""
    assert u.build_author_dir("a/b:c*d?", "12 3") == "abcd_12-3"
    assert "/" not in u.build_author_dir("a/b", "1/2")


def test_author_dir_dedupes_same_handle(u):
    """X 这类站点上传者名就是 handle，别产出 someone_someone"""
    assert u.build_author_dir("@Someone", "@Someone") == "someone"


def test_author_dir_keeps_long_sec_uid_distinct(u):
    """sec_uid 公共前缀就有 17 字符，截断到 20 会让不同作者撞同一目录"""
    first = "MS4wLjABAAAA" + "x" * 45
    second = "MS4wLjABAAAA" + "y" * 45
    assert u.build_author_dir("n", first) != u.build_author_dir("n", second)


# ───────────────────────── 作品名 ─────────────────────────


def test_work_stem_strips_hashtags_and_spaces(u):
    assert u.build_work_stem("旅行 #随手拍 vlog") == "旅行-vlog"


def test_work_stem_empty_falls_back(u):
    assert u.build_work_stem("") == "作品"
    assert u.build_work_stem(None) == "作品"
    assert u.build_work_stem("🎬") == "作品"


def test_work_stem_truncates(u):
    assert len(u.build_work_stem("标题" * 20)) <= 15


# ───────────────────────── 短码与重名 ─────────────────────────


def test_short_time_code_is_four_char_base36(u):
    assert u.short_time_code(0) == "0000"
    code = u.short_time_code(1_700_000_000)
    assert len(code) == 4 and code.isalnum()
    # 相邻秒不同码
    assert code != u.short_time_code(1_700_000_001)


def test_unique_media_path_without_collision(u, tmp_path):
    target = tmp_path / "作品.mp4"
    assert u.unique_media_path(target) == target


def test_unique_media_path_adds_code_then_index(u, tmp_path):
    target = tmp_path / "作品.mp4"
    target.write_bytes(b"x")
    code = u.short_time_code(1_700_000_000)

    second = u.unique_media_path(target, at=1_700_000_000)
    assert second.name == f"作品_{code}.mp4"

    second.write_bytes(b"x")
    third = u.unique_media_path(target, at=1_700_000_000)
    assert third.name == f"作品_{code}_2.mp4"


def test_unique_media_path_creates_author_dir(u, tmp_path):
    """落盘前才建作者目录：命名函数顺带确保父目录存在"""
    target = tmp_path / "作者_123" / "作品.mp4"
    assert u.unique_media_path(target) == target
    assert target.parent.is_dir()


def test_unique_media_path_works_for_directories(u, tmp_path):
    """多图作品目录同名时同样加短码"""
    note_dir = tmp_path / "作者_123" / "作品"
    note_dir.mkdir(parents=True)
    renamed = u.unique_media_path(note_dir, at=1_700_000_000)
    assert renamed.name == f"作品_{u.short_time_code(1_700_000_000)}"


# ───────────────────────── S3 key ─────────────────────────


def test_media_key_keeps_author_dir_and_drops_platform(u, temp_root):
    f = temp_root / "douyin" / "作者_123" / "作品.mp4"
    assert u.media_key_of(f) == "作者_123/作品.mp4"
    assert u.media_rel_dir_of(f) == "作者_123"


def test_media_key_of_multi_image_work(u, temp_root):
    f = temp_root / "bilibili" / "作者_9" / "作品_ab12" / "001.jpg"
    assert u.media_key_of(f) == "作者_9/作品_ab12/001.jpg"
    assert u.media_rel_dir_of(f) == "作者_9/作品_ab12"


def test_media_key_of_archive_follows_same_rule(u, temp_root):
    """群文件 zip 落在 archive/{作者目录} 下，key 规则与媒体一致"""
    f = temp_root / "archive" / "作者_9" / "作品_群文件_120606.zip"
    assert u.media_key_of(f) == "作者_9/作品_群文件_120606.zip"


def test_media_key_of_outside_temp_falls_back_to_name(u, temp_root, tmp_path):
    f = tmp_path / "elsewhere" / "x.mp4"
    assert u.media_key_of(f) == "x.mp4"
    assert u.media_rel_dir_of(f) == ""


def test_media_key_of_flat_file_under_platform(u, temp_root):
    """老数据/第三方产物：平台层下没有作者层 → 只用文件名"""
    f = temp_root / "douyin" / "x.mp4"
    assert u.media_key_of(f) == "x.mp4"
    assert u.media_rel_dir_of(f) == ""