利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Python ai-video-generatorcontent-creationffmpeginstagram-reelsllmpythonshort-videosubtitlestext-to-speechtiktokvideo-automationvideo-workflowworkflow-automationyoutube-shorts
import asyncio
import edge_tts
from unittest.mock import patch
import threading
import unittest
from app.services import voice
class TestEdgeProducerShutdown(unittest.TestCase):
def test_callback_failure_stops_producer_and_closes_generator(self):
resume = threading.Event()
closed = threading.Event()
consumed = []
class Stream:
def stream_sync(self):
try:
yield {"type": "audio", "data": b"first"}
resume.wait(2)
for index in range(100):
consumed.append(index)
yield {"type": "audio", "data": b"tail"}
finally:
closed.set()
def fail(chunk):
raise OSError("disk full")
with self.assertRaisesRegex(OSError, "disk full"):
voice._stream_edge_tts_sync_with_timeout(Stream(), fail, 2)
resume.set()
self.assertTrue(closed.wait(2), "producer did not finish")
self.assertLessEqual(len(consumed), 2, "abandoned producer consumed the entire stream")
def test_normal_stream_and_producer_error_are_forwarded(self):
class Stream:
def stream_sync(self):
yield {"type": "audio", "data": b"first"}
raise ValueError("broken stream")
chunks = []
with self.assertRaisesRegex(ValueError, "broken stream"):
voice._stream_edge_tts_sync_with_timeout(Stream(), chunks.append, 2)
self.assertEqual(chunks, [{"type": "audio", "data": b"first"}])
def test_timed_out_stream_stops_when_blocked_read_returns(self):
resume = threading.Event()
closed = threading.Event()
consumed = []
class Stream:
def stream_sync(self):
try:
resume.wait(2)
for index in range(100):
consumed.append(index)
yield {"type": "audio", "data": b"late"}
finally:
closed.set()
try:
with self.assertRaises(TimeoutError):
voice._stream_edge_tts_sync_with_timeout(Stream(), lambda chunk: None, 0.01)
finally:
resume.set()
self.assertTrue(closed.wait(2))
self.assertLessEqual(len(consumed), 2)
def test_real_sdk_stream_cancelled_without_draining_tail_after_callback_failure(self):
resume = threading.Event()
closed = threading.Event()
fetched = []
communicate = edge_tts.Communicate("Hello.", "en-US-AriaNeural")
async def network_stream():
try:
yield {"type": "audio", "data": b"first"}
while not resume.is_set():
await asyncio.sleep(0.001)
for index in range(100):
fetched.append(index)
yield {"type": "audio", "data": b"tail"}
finally:
closed.set()
def fail(chunk):
raise OSError("disk full")
with patch.object(communicate, "stream", new=network_stream):
try:
with self.assertRaisesRegex(OSError, "disk full"):
voice._stream_edge_tts_sync_with_timeout(communicate, fail, 2)
finally:
resume.set()
self.assertTrue(closed.wait(2))
self.assertLessEqual(len(fetched), 2)
def test_real_sdk_blocked_async_read_is_cancelled_after_timeout(self):
closed = threading.Event()
started = threading.Event()
communicate = edge_tts.Communicate("Hello.", "en-US-AriaNeural")
async def network_stream():
try:
started.set()
await asyncio.sleep(0.2)
yield {"type": "audio", "data": b"late"}
finally:
closed.set()
with patch.object(communicate, "stream", new=network_stream):
with self.assertRaises(TimeoutError):
voice._stream_edge_tts_sync_with_timeout(communicate, lambda chunk: None, 0.05)
self.assertTrue(started.is_set())
self.assertTrue(closed.wait(0.1), "SDK read was not cancelled promptly")