This repository has been archived on 2026-10-02. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
NeteaseUrl/tests/test_podcast_worker.py
T
snowy-yang 1c6b17f718 refactor: 播客单集 worker 拆解(140 行 / 圈 35 → 40 行 / 圈 6 / 嵌套 1)
为什么:_podcast_download_worker 是播客单集下载的唯一后台执行路径,
此前零直接测试覆盖;三条 except 分支各写一遍「结算+删任务目录」,
正文散着三处取消检查,改坏任意一处不会有测试报警。

拆出(与单曲 worker 9cbd341 同风格):
- _PodcastTaskSpec(frozen dataclass,program_id 标注 Any——params 历史
  存数字字符串,worker 只透传,int() 是行为变更)
- _PodcastTarget(磁盘名/交付名/任务目录的命名组装)
- _begin_podcast_task / _prepare_podcast_target / _write_podcast_zip /
  _podcast_file_result / _podcast_zip_result / _publish_podcast_task /
  _fail_podcast_task(三处 ×9 行收尾收敛)
- 取消检查复用 downloads._abort_if_canceled(无循环依赖);tempfile
  局部 import 上移模块顶

先写测试后动代码:新增 tests/test_podcast_worker.py 18 例,在**未改动**
实现上一次跑绿 = 忠实编码旧行为,重构后同一套测试即等价性验证。
唯一语义顺序变化:except 分支的日志从「收尾后」挪到「收尾前」(与
单曲 worker 对齐,日志内容不变,无行为影响)。
门禁:ruff 全过 · pyright 0/0 · pytest 312 passed(294→312)。
2026-09-27 09:30:52 +08:00

551 lines
21 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# -*- coding: utf-8 -*-
"""播客单集下载 worker 的行为契约测试(沙箱,不触碰生产 data/)。
为什么:``services/podcast.py::_podcast_download_worker`` 是播客单集下载的**唯一**
后台执行路径(解析 → 音质降级直链 → 流式下载 → 可选 ZIP → /download/result 交付),
此前**零直接测试覆盖**(test_podcast_stream.py 只覆盖到流式下载这一层)。它有三条
except 分支各写一遍「结算 + 删任务目录」,正文里散着三处取消检查——改坏任意一处
都不会有测试报警。
按惯例:**先在未改动的原始实现上跑绿**,这批测试因此忠实编码了旧行为,重构时本身
即等价性验证器。断言只依赖「调用 worker 后世界里发生了什么」(任务终态、磁盘上的
文件、ZIP 内容、账本行、余额),不引用 worker 内部函数名。
替身边界(与真实契约一一对应):
- ``podcast.get_program_detail`` / ``podcast.resolve_episode_url``:可编程节目信息与直链;
- ``_podcast_stream_download``:契约是「成功写 target 并返回字节数;取消抛
DownloadCanceled;失败抛 DownloadException」,替身照此实现,不模拟字节流;
- ``fetch_cover_bytes``:失败返回 None(真实契约内部吞异常),不抛。
账本口径:一律用 ``reserve_and_open`` 真预扣(与 test_single_worker.py 同理由)。
"""
import threading
import time
import zipfile
from pathlib import Path
from typing import Any
import pytest
import auth_system as auth
import core.db as core_db
from core.state import DOWNLOAD_TASKS, DOWNLOAD_TASKS_LOCK
from music_downloader import DownloadCanceled, DownloadException
from services import podcast as podcast_svc
from services.api_service import api_service
PW = 'Passw0rd12'
CREDITS = 50
# ============ 测试替身 ============
class _NullLogger:
"""吞掉 api_service.logger 输出,避免测试往项目根 music_api.log 追加。"""
def debug(self, *a: Any, **k: Any) -> None: ...
def info(self, *a: Any, **k: Any) -> None: ...
def warning(self, *a: Any, **k: Any) -> None: ...
def error(self, *a: Any, **k: Any) -> None: ...
def _program(**over: Any) -> dict[str, Any]:
"""标准节目详情替身(字段与 music_api.podcast.get_program_detail 对齐)。"""
d: dict[str, Any] = {
'main_song_id': 999,
'name': '第一集 测试',
'serial_num': 3,
'cover_url': 'https://cover.example/x.jpg',
'radio_name': '测试电台',
'dj_name': '主播甲',
'duration_text': '10:00',
}
d.update(over)
return d
def _url_info(**over: Any) -> dict[str, Any]:
d: dict[str, Any] = {'url': 'https://cdn.example/a.mp3', 'type': 'mp3',
'size': 11, 'level': 'standard'}
d.update(over)
return d
class _FakeStream:
"""``_podcast_stream_download`` 替身。
``on_call`` 在写入 target 之前被调用(注入取消置位 / 外部终态 / 抛错);
``fail`` 模拟流式下载失败(如完整性不达标)。
"""
def __init__(self, *, fail: Exception | None = None, on_call: Any = None):
self.fail = fail
self.on_call = on_call
self.calls: list[dict[str, Any]] = []
def __call__(self, url: str, target: Path, total_size: int,
cancel_ev: threading.Event, task_id: str = '',
progress_label: str = '') -> int:
self.calls.append({'url': url, 'target': Path(target),
'total_size': total_size})
if self.on_call is not None:
self.on_call()
if cancel_ev.is_set():
raise DownloadCanceled('任务已取消')
if self.fail is not None:
raise self.fail
Path(target).write_bytes(b'audio-bytes')
return 11
# ============ 沙箱 ============
@pytest.fixture()
def sandbox(tmp_path, monkeypatch):
"""临时站点库 + 临时下载目录 + 空任务表 + 播客上游替身。"""
db_path = tmp_path / 'site.db'
monkeypatch.setattr(auth.db, 'DATA_DIR', tmp_path)
monkeypatch.setattr(auth.db, 'DB_PATH', db_path)
monkeypatch.setattr(auth.db, 'BACKUP_DIR', tmp_path / 'backups')
monkeypatch.setattr(auth.db, 'SECRET_KEY_PATH', tmp_path / 'secret.key')
monkeypatch.setattr(core_db, 'DATA_DIR', tmp_path)
monkeypatch.setattr(core_db, 'DB_PATH', db_path)
auth.init_db()
auth.create_user('payer', PW, email='p@example.com', credits=CREDITS)
downloads = tmp_path / 'downloads'
downloads.mkdir()
monkeypatch.setattr(api_service, 'downloads_path', downloads)
monkeypatch.setattr(api_service, 'logger', _NullLogger())
# 操作记录写进程内共享列表——隔离掉,断言改由个例捕获
monkeypatch.setattr(podcast_svc, '_record_log', lambda *a: None)
with DOWNLOAD_TASKS_LOCK:
DOWNLOAD_TASKS.clear()
yield {'root': tmp_path, 'downloads': downloads}
with DOWNLOAD_TASKS_LOCK:
DOWNLOAD_TASKS.clear()
# ============ 工具 ============
def _uid(username: str = 'payer') -> int:
u = auth.get_user_by_name(username)
assert u is not None
return int(u['id'])
def _balance(username: str = 'payer') -> int:
u = auth.get_user_by_name(username)
assert u is not None
return int(u['duration_seconds'] or 0)
def _status(task_id: str) -> str:
with DOWNLOAD_TASKS_LOCK:
return str(DOWNLOAD_TASKS.get(task_id, {}).get('status') or '')
def _register_podcast(sandbox, *, program_id: int = 777, user_id: int,
reserved: int = 1, **params: Any) -> tuple[str, dict[str, Any]]:
tid = 'p' + str(int(time.time() * 1000000) % 10 ** 10)
left = auth.reserve_and_open(user_id, reserved, tid)
assert left is not None, '预扣失败(余额不足,检查 fixture 的 credits)'
task: dict[str, Any] = {
'type': 'single', 'task_id': tid,
'user_id': user_id, 'billing_user_id': user_id, 'billing_reserved': reserved,
'total': 1, 'done': 0, 'failed': 0,
'status': 'queued', 'start': time.time(), 'msg': '排队中...',
'cancel': threading.Event(), 'pause': threading.Event(),
'device_id': 'dev-test', 'result': None,
'params': {
'program_id': program_id, 'quality': 'standard',
'need_zip': False, 'cookies': {}, 'remote_ip': '1.2.3.4',
**params,
},
}
with DOWNLOAD_TASKS_LOCK:
DOWNLOAD_TASKS[tid] = task
return tid, task
def _install_upstream(monkeypatch, *, program: dict[str, Any] | None = None,
url_info: dict[str, Any] | None = None,
stream: Any = None,
cover: Any = lambda url: b'\xff\xd8\xff cover') -> None:
"""装上播客上游替身(节目详情 / 直链 / 流式下载 / 封面)。"""
monkeypatch.setattr(podcast_svc.podcast, 'get_program_detail',
lambda pid, cookies: program if program is not None else _program())
monkeypatch.setattr(podcast_svc.podcast, 'resolve_episode_url',
lambda sid, cookies, quality:
url_info if url_info is not None else _url_info())
if stream is None:
stream = _FakeStream()
monkeypatch.setattr(podcast_svc, '_podcast_stream_download', stream)
monkeypatch.setattr(podcast_svc, 'fetch_cover_bytes', cover)
def _run(task_id: str) -> threading.Thread:
th = threading.Thread(target=podcast_svc._podcast_download_worker,
args=(task_id,), daemon=True)
th.start()
return th
def _zip_names(path: str) -> list[str]:
with zipfile.ZipFile(path) as zf:
return zf.namelist()
# ============ 正常路径 ============
def test_podcast_download_completes(sandbox, monkeypatch):
"""常规单集(不打包):终态 done、音频落在任务目录、账本按交付结算、余额 -1。"""
uid = _uid()
before = _balance()
_install_upstream(monkeypatch)
tid, task = _register_podcast(sandbox, program_id=777, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert not th.is_alive(), 'worker 未在 30s 内结束'
assert _status(tid) == 'done', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['msg'] == '下载完成'
assert task['done'] == 1 and task['completed_at'] > 0
res = task['result']
assert res is not None
# 交付名干净(不带节目 ID),磁盘名带 [节目ID]
assert res['filename'] == '003 - 第一集 测试.mp3'
assert res['mimetype'] == 'audio/mp3'
assert res['audio_files'] == [res['file_path']]
assert 'zip_path' not in res
delivered = Path(res['file_path'])
assert delivered.exists()
assert delivered.name == '003 - 第一集 测试 [777].mp3'
assert delivered.parent == sandbox['downloads'] / 'tasks' / tid
assert delivered.read_bytes() == b'audio-bytes'
row = auth.ledger_get(tid)
assert row is not None and row['status'] != 'open'
assert row['delivered'] == 1
assert _balance() == before - 1, '交付成功:只扣预扣的 1 次,无退款'
def test_no_serial_prefix_when_serial_zero(sandbox, monkeypatch):
"""serial_num 缺失/为 0:文件名不带序号前缀,直接用单集名。"""
uid = _uid()
_install_upstream(monkeypatch, program=_program(serial_num=0))
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done'
assert task['result']['filename'] == '第一集 测试.mp3'
def test_ext_follows_url_info_type(sandbox, monkeypatch):
"""直链 type=m4a:交付名/磁盘名扩展名随之变化。"""
uid = _uid()
_install_upstream(monkeypatch, url_info=_url_info(type='m4a'))
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done'
res = task['result']
assert res['filename'] == '003 - 第一集 测试.m4a'
assert res['mimetype'] == 'audio/m4a'
def test_record_log_gets_episode_name(sandbox, monkeypatch):
"""成功交付后按单集名记一条操作记录(IP 取 params.remote_ip)。"""
uid = _uid()
logs: list[tuple] = []
_install_upstream(monkeypatch)
monkeypatch.setattr(podcast_svc, '_record_log', lambda *a: logs.append(a))
tid, _ = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert logs == [('下载', '1.2.3.4', '[播客] 第一集 测试')], logs
def test_remote_ip_defaults_when_missing(sandbox, monkeypatch):
"""params 无 remote_ip:按 '-' 记录,不因缺字段报错。"""
uid = _uid()
logs: list[tuple] = []
_install_upstream(monkeypatch)
monkeypatch.setattr(podcast_svc, '_record_log', lambda *a: logs.append(a))
tid, task = _register_podcast(sandbox, user_id=uid)
del task['params']['remote_ip']
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done'
assert logs == [('下载', '-', '[播客] 第一集 测试')], logs
def test_zip_contains_audio_cover_and_info(sandbox, monkeypatch):
"""need_zip=true:包内含音频 + 封面 + info.txt,result 走 zip 交付形状。"""
uid = _uid()
_install_upstream(monkeypatch)
tid, task = _register_podcast(sandbox, user_id=uid, need_zip=True)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done', f"终态={_status(tid)} msg={task.get('msg')}"
res = task['result']
assert res is not None
assert res['filename'] == '003 - 第一集 测试.zip'
assert res['mimetype'] == 'application/zip'
assert sorted(_zip_names(res['zip_path'])) == [
'003 - 第一集 测试.jpg', '003 - 第一集 测试.mp3', 'info.txt']
with zipfile.ZipFile(res['zip_path']) as zf:
info = zf.read('info.txt').decode('utf-8')
assert '单集: 第一集 测试' in info
assert '播客: 测试电台' in info
assert '主播: 主播甲' in info
assert '时长: 10:00' in info
assert '音质: ' in info
# 音频正文同时留在任务目录(result.file_path 指向它)
assert Path(res['file_path']).exists()
def test_zip_without_cover_when_cover_url_empty(sandbox, monkeypatch):
"""节目无 cover_url:ZIP 不含封面条目,其余照常(封面非交付必要条件)。"""
uid = _uid()
_install_upstream(monkeypatch, program=_program(cover_url=''))
tid, task = _register_podcast(sandbox, user_id=uid, need_zip=True)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done'
names = _zip_names(task['result']['zip_path'])
assert not [n for n in names if n.endswith('.jpg')], names
assert [n for n in names if n.endswith('.mp3')], names
def test_cover_fetch_failure_still_delivers_zip(sandbox, monkeypatch):
"""取封面抛异常:被吞掉,ZIP 照常交付(历史行为:封面失败不炸任务)。"""
uid = _uid()
def _boom(url: str) -> bytes:
raise RuntimeError('cover cdn down')
_install_upstream(monkeypatch, cover=_boom)
tid, task = _register_podcast(sandbox, user_id=uid, need_zip=True)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done', f"终态={_status(tid)} msg={task.get('msg')}"
names = _zip_names(task['result']['zip_path'])
assert not [n for n in names if n.endswith('.jpg')], names
# ============ 解析失败 ============
def test_missing_song_id_marks_error_and_refunds(sandbox, monkeypatch):
"""单集缺 main_song_id:转 error(文案透出)、退费、任务目录清理。"""
uid = _uid()
before = _balance()
_install_upstream(monkeypatch, program=_program(main_song_id=None))
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'error', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['msg'] == '该单集缺少音频信息(mainSong)'
assert task['result'] is None
assert _balance() == before, '失败应全额退回预扣'
row = auth.ledger_get(tid)
assert row is not None and row['delivered'] == 0 and row['status'] != 'open'
assert not (sandbox['downloads'] / 'tasks' / tid).exists(), '失败应清掉任务目录'
def test_no_url_marks_error_and_refunds(sandbox, monkeypatch):
"""直链解析为空(下架/版权):转 error、退费、目录清理。"""
uid = _uid()
before = _balance()
_install_upstream(monkeypatch, url_info=_url_info(url=''))
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'error'
assert task['msg'] == '未获取到音频直链(单集可能已下架或受版权限制)'
assert _balance() == before
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
def test_stream_failure_marks_error_and_refunds(sandbox, monkeypatch):
"""流式下载失败(如完整性不达标):转 error、退费、目录清理。"""
uid = _uid()
before = _balance()
_install_upstream(monkeypatch,
stream=_FakeStream(fail=DownloadException('下载不完整(5/11 字节)')))
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'error'
assert task['msg'] == '下载不完整(5/11 字节)'
assert _balance() == before
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
def test_unexpected_exception_uses_generic_message(sandbox, monkeypatch):
"""非预期异常:用户可见文案泛化(不泄上游细节),原文进日志、余额退回。"""
uid = _uid()
before = _balance()
def _boom(pid: Any, cookies: Any) -> dict[str, Any]:
raise RuntimeError('upstream host=10.0.0.9 refused')
_install_upstream(monkeypatch)
monkeypatch.setattr(podcast_svc.podcast, 'get_program_detail', _boom)
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'error', f"终态={_status(tid)}"
assert task['msg'] == '下载失败,请稍后重试'
assert '10.0.0.9' not in task['msg'], '上游细节不得出现在用户可见文案里'
assert _balance() == before
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
# ============ 取消 ============
def test_cancel_before_stream_start(sandbox, monkeypatch):
"""流式下载启动前取消已置位:canceled、退费、目录清理。"""
uid = _uid()
before = _balance()
holder: dict[str, Any] = {}
def _cancel_first():
holder['task']['cancel'].set()
_install_upstream(monkeypatch, stream=_FakeStream(on_call=_cancel_first))
tid, task = _register_podcast(sandbox, user_id=uid)
holder['task'] = task
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'canceled', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['result'] is None
assert _balance() == before, '取消应全额退回预扣'
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
def test_cancel_after_download_before_publish(sandbox, monkeypatch):
"""下载已落盘但取消随即置位:不发布交付,转 canceled 并退费。"""
uid = _uid()
before = _balance()
holder: dict[str, Any] = {}
def _cancel_after():
threading.Timer(0.0, lambda: holder['task']['cancel'].set()).start()
_install_upstream(monkeypatch, stream=_FakeStream(on_call=_cancel_after))
tid, task = _register_podcast(sandbox, user_id=uid)
holder['task'] = task
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'canceled', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['result'] is None, '取消后不得发布交付'
assert _balance() == before
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
def test_cancel_during_zip_stage(sandbox, monkeypatch):
"""打包阶段取消置位(封面回调里触发):ZIP 不发布、canceled、退费。"""
uid = _uid()
before = _balance()
holder: dict[str, Any] = {}
def _cancel_in_cover(url: str) -> bytes:
holder['task']['cancel'].set()
return b'\xff\xd8\xff cover'
_install_upstream(monkeypatch, cover=_cancel_in_cover)
tid, task = _register_podcast(sandbox, user_id=uid, need_zip=True)
holder['task'] = task
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'canceled', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['result'] is None
assert _balance() == before
assert not (sandbox['downloads'] / 'tasks' / tid).exists()
def test_terminal_state_not_overwritten_by_late_success(sandbox, monkeypatch):
"""任务已被外部置终态(canceled)时,worker 的成功路径不得覆盖为 done。"""
uid = _uid()
before = _balance()
holder: dict[str, Any] = {}
def _mark_canceled():
with DOWNLOAD_TASKS_LOCK:
holder['task']['status'] = 'canceled'
holder['task']['completed_at'] = time.time()
_install_upstream(monkeypatch, stream=_FakeStream(on_call=_mark_canceled))
tid, task = _register_podcast(sandbox, user_id=uid)
holder['task'] = task
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'canceled', f"终态={_status(tid)}"
assert task['result'] is None, '已终态不得被覆盖为 done'
assert _balance() == before
# ============ 结算兜底 ============
def test_settle_failure_does_not_flip_done_to_error(sandbox, monkeypatch):
"""交付已落定后结算抛错:任务仍为 done(钱由重启对账兜底,不坑已付费用户)。"""
uid = _uid()
def _boom(*a: Any, **k: Any) -> None:
raise RuntimeError('库锁超时')
_install_upstream(monkeypatch)
monkeypatch.setattr(podcast_svc.auth_system, 'ledger_settle', _boom)
tid, task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'done', f"终态={_status(tid)} msg={task.get('msg')}"
assert task['result'] is not None
def test_ledger_open_when_worker_crashes_before_settle(sandbox, monkeypatch):
"""worker 在结算前崩掉:账本行仍为 open(留给重启对账兜底,不是静默丢失)。
注意:光让 ledger_settle 抛错不够——成功路径对结算失败有 try/except 吞掉
(上一条测试验证的行为)。必须让下载本身先抛非预期异常走进通用 except,
连「退款结算」也没走成,账本才停留在 open。
"""
uid = _uid()
def _boom(*a: Any, **k: Any) -> None:
raise RuntimeError('db busy')
_install_upstream(monkeypatch,
stream=_FakeStream(fail=RuntimeError('boom')))
monkeypatch.setattr(podcast_svc.auth_system, 'ledger_settle', _boom)
tid, _task = _register_podcast(sandbox, user_id=uid)
th = _run(tid)
th.join(timeout=30)
assert _status(tid) == 'error'
row = auth.ledger_get(tid)
assert row is not None
assert row['status'] == 'open', f"应留给对账兜底,实际 status={row['status']}"