Revert "[serve] Add test for Gradio queues (which use websockets) (ra…

…y-project#36372)" (ray-project#37363)

doc/requirements-doc.txt

@@ -62,11 +62,11 @@ autodoc_pydantic==1.6.1
 sphinx_design==0.4.1
 
 # MyST
-myst-parser==0.18.1
-myst-nb==0.17.2
+myst-parser==0.15.2
+myst-nb==0.13.1
 
 # Jupyter conversion
-jupytext==1.14.6
+jupytext==1.13.6
 
 # Pin urllib to avoid downstream ssl incompatibility issues
-urllib3 < 1.27
+urllib3 < 1.27

python/ray/serve/tests/test_gradio.py

@@ -2,17 +2,27 @@
 import sys
 
 import gradio as gr
-from gradio_client import Client
 import pytest
+import requests
 
+import ray
 from ray._private.test_utils import wait_for_condition
 
 from ray import serve
 from ray.serve.gradio_integrations import GradioIngress, GradioServer
 
 
+@pytest.fixture
+def serve_start_shutdown():
+ ray.init()
+ serve.start()
+ yield
+ serve.shutdown()
+ ray.shutdown()
+
+
 @pytest.mark.parametrize("use_user_defined_class", [False, True])
-def test_gradio_ingress_correctness(serve_instance, use_user_defined_class: bool):
+def test_gradio_ingress_correctness(serve_start_shutdown, use_user_defined_class: bool):
  """
  Ensure a Gradio app deployed to a cluster through GradioIngress still
  produces the correct output.
@@ -38,11 +48,16 @@ def __init__(self):
 
  serve.run(app)
 
- client = Client("http:https://localhost:8000")
- assert client.predict("Alice") == "Good morning Alice!"
+ test_input = "Alice"
+ response = requests.post(
+ "http:https://127.0.0.1:8000/api/predict/", json={"data": [test_input]}
+ )
+ assert response.status_code == 200 and response.json()["data"][0] == greet(
+ test_input
+ )
 
 
-def test_gradio_ingress_scaling(serve_instance):
+def test_gradio_ingress_scaling(serve_start_shutdown):
  """
  Check that a Gradio app that has been deployed to a cluster through
  GradioIngress scales as needed, i.e. separate client requests are served by
@@ -52,15 +67,26 @@ def test_gradio_ingress_scaling(serve_instance):
  def f(*args):
  return os.getpid()
 
- serve.run(
- GradioServer.options(num_replicas=2).bind(
- lambda: gr.Interface(fn=f, inputs="text", outputs="text")
- )
+ app = GradioServer.options(num_replicas=2).bind(
+ lambda: gr.Interface(fn=f, inputs="text", outputs="text")
  )
+ serve.run(app)
+
+ def two_pids_returned():
+ @ray.remote
+ def get_pid_from_request():
+ r = requests.post(
+ "http:https://127.0.0.1:8000/api/predict/", json={"data": ["input"]}
+ )
+ r.raise_for_status()
+ return r.json()["data"][0]
+
+ return (
+ len(set(ray.get([get_pid_from_request.remote() for _ in range(10)]))) == 2
+ )
 
- client = Client("http:https://localhost:8000")
  # Verify that the requests are handled by two separate replicas.
- wait_for_condition(lambda: len({client.predict("input") for _ in range(3)}) == 2)
+ wait_for_condition(two_pids_returned)
 
 
 if __name__ == "__main__":

python/ray/serve/tests/test_gradio_visualization.py

@@ -1,10 +1,9 @@
 import pytest
 from collections import defaultdict
 import asyncio
+import aiohttp
 import random
 
-from gradio_client import Client
-
 from ray.serve.experimental.gradio_visualize_graph import GraphVisualizer
 from ray.dag.utils import _DAGNodeNameGenerator
 from ray import serve
@@ -235,9 +234,28 @@ async def test_gradio_visualization_e2e(graph1):
  visualizer = GraphVisualizer()
  (_, url, _) = visualizer.visualize_with_gradio(handle, _launch=True, _block=False)
 
- client = Client(url)
- client.predict(random.randint(0, 100), 1, 2, fn_index=0)
- assert {client.predict(fn_index=1), client.predict(fn_index=2)} == {1, 2}
+ async with aiohttp.ClientSession() as session:
+
+ async def fetch(data, fn_index):
+ async with session.post(
+ f"{url.strip('/')}/api/predict/",
+ json={
+ "session_hash": "random_hash",
+ "data": data,
+ "fn_index": fn_index,
+ },
+ ) as resp:
+ return (await resp.json())["data"]
+
+ await fetch(
+ [random.randint(0, 100), 1, 2], 0
+ ) # sends request to dag with input (1,2)
+ values = await asyncio.gather(
+ fetch([], 1), # fetches return value for one of the nodes
+ fetch([], 2), # fetches return value for the other node
+ )
+
+ assert [1] in values and [2] in values
 
 
 @pytest.mark.asyncio

python/ray/serve/tests/test_websockets.py

@@ -9,7 +9,6 @@
 from websockets.sync.client import connect
 
 import ray
-from ray._private.test_utils import wait_for_condition
 
 from ray import serve
 from ray.serve._private.constants import RAY_SERVE_ENABLE_EXPERIMENTAL_STREAMING
@@ -153,47 +152,5 @@ async def ws_hi(self, ws: WebSocket):
  assert ws.recv() == "hi"
 
 
-@pytest.mark.skipif(
- not RAY_SERVE_ENABLE_EXPERIMENTAL_STREAMING,
- reason="Streaming feature flag is disabled.",
-)
-@pytest.mark.skipif(sys.platform == "win32", reason="Gradio doesn't work on Windows.")
-@pytest.mark.skipif(sys.version_info.minor < 8, reason="Times out on Python 3.7.")
-def test_gradio_queue(serve_instance):
- """Test the Gradio integration with a Gradio Queue.
-
- Gradio Queues use websockets under the hood.
- """
-
- # Delayed imports because these aren't installed on Windows.
- import gradio as gr
- from gradio_client import Client
- from ray.serve.gradio_integrations import GradioIngress
-
- def counter(num_steps: int = 3):
- for i in range(num_steps):
- yield str(i)
-
- @serve.deployment
- class GradioGenerator(GradioIngress):
- def __init__(self):
- g = gr.Interface(counter, inputs=gr.Slider(1, 10, 3), outputs="text")
- super().__init__(lambda: g.queue())
-
- serve.run(GradioGenerator.bind())
-
- client = Client("http:https://localhost:8000")
- job1 = client.submit(3, api_name="/predict")
- job2 = client.submit(5, api_name="/predict")
-
- wait_for_condition(
- lambda: (
- (job1.done() and job2.done())
- and job1.outputs() == [str(i) for i in range(3)]
- and job2.outputs() == [str(i) for i in range(5)]
- )
- )
-
-
 if __name__ == "__main__":
  sys.exit(pytest.main(["-v", "-s", __file__]))

python/requirements/test-requirements.txt

@@ -27,8 +27,7 @@ flask==2.1.3
 freezegun==1.1.0
 google-api-python-client==2.65.0
 google-cloud-storage==2.5.0
-gradio==3.34; platform_system != "Windows"
-gradio-client==0.2.6; platform_system != "Windows"
+gradio==3.11; platform_system != "Windows"
 websockets==11.0.3
 joblib==1.2.0
 jsonpatch==1.32
@@ -79,10 +78,10 @@ xlrd==2.0.1
 memray; platform_system != "Windows" and sys_platform != "darwin"
 
 # For doc tests
-myst-parser==0.18.1
-myst-nb==0.17.2
+myst-parser==0.15.2
+myst-nb==0.13.1
 sphinx==4.3.2
-jupytext==1.14.6
+jupytext==1.13.6
 jinja2==3.0.3
 pytest-docker-tools==3.1.3
 pytest-forked==1.4.0
@@ -100,7 +99,7 @@ importlib-metadata==4.10.0
 tensorboardX==2.6.0
 starlette==0.27.0
 h11==0.12.0
-markdown-it-py==2.2.0
+markdown-it-py==1.1.0
 attrs==21.4.0
 pytz==2022.7.1
 # Compatibility with spacy 3.5 (model en_core_web_sm)