[Cherry-Pick] [Feature] support v1 update/clear api for RL (#6761) (#6974)

* [Feature] support v1 update/clear api for RL

* [fix] fix stale control responses when control method timed out

* [chore] remove unused code

* [chore] optimize tags and key_prefix

* [test] fix ci

* [chore] fix code style

* [fix] fix ep control

* [fix] fix ep control for engine cache queue
This commit is contained in:
Yonghua Li
2026-03-25 19:18:35 +08:00
committed by GitHub
parent 49c2310854
commit 35034f91fa
25 changed files with 1665 additions and 328 deletions
@@ -87,15 +87,19 @@ class TestCacheTransferManager(unittest.TestCase):
# --------------------------
# mock IPCSignal
# --------------------------
patcher2 = patch("fastdeploy.cache_manager.cache_transfer_manager.IPCSignal", new=MagicMock())
class DummyIPCSignal:
def __init__(self, name, array, dtype, suffix, create=False):
self.value = array
patcher2 = patch("fastdeploy.cache_manager.cache_transfer_manager.IPCSignal", new=DummyIPCSignal)
patcher2.start()
self.addCleanup(patcher2.stop)
# --------------------------
# mock _init_cpu_cache 和 _init_gpu_cache
# --------------------------
patcher3 = patch.object(CacheTransferManager, "_init_cpu_cache", lambda self, args: None)
patcher4 = patch.object(CacheTransferManager, "_init_gpu_cache", lambda self, args: None)
patcher3 = patch.object(CacheTransferManager, "_init_cpu_cache", lambda self: None)
patcher4 = patch.object(CacheTransferManager, "_init_gpu_cache", lambda self: None)
patcher3.start()
patcher4.start()
self.addCleanup(patcher3.stop)