def test_v2_errors():
context = init_ctx(context=NuclioContext(logger=logger))
event = MockEvent('{"test": "ok"}', path="/v2/models/m1/xx")
resp = context.mlrun_handler(context, event)
# expected: 400, 'illegal model operation xx, method=None'
assert resp.status_code == 400, f"didnt get proper handler error {resp.body}"
event = MockEvent('{"test": "ok"}', path="/v2/models/m5/xx")
resp = context.mlrun_handler(context, event)
# expected: 400, 'model m5 doesnt exist, available models: m1 | m2 | m3:v1 | m3:v2'
assert resp.status_code == 400, f"didnt get proper model error {resp.body}"
def test_v2_health():
context = init_ctx()
event = MockEvent(None, path="/", method="GET")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
# expected: {'name': 'ModelRouter', 'version': 'v2', 'extensions': []}
assert data["version"] == "v2", f"wrong health response {resp.body}"
event = MockEvent("", path="/v2/health", method="GET")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data["version"] == "v2", f"wrong health response {resp.body}"
def test_v2_stream_mode():
# model and operation are specified inside the message body
context = init_ctx()
event = MockEvent('{"model": "m2", "inputs": [5]}')
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data["outputs"] == 1000, f"wrong model response {resp.body}"
event = MockEvent(
'{"model": "m3:v2", "operation": "explain", "inputs": [5]}', path="")
resp = context.mlrun_handler(context, event)
logger.info(f"resp: {resp.body}")
data = json.loads(resp.body)
assert data["outputs"]["explained"] == 5, f"wrong model response {data}"
def test_v2_explain():
context = init_ctx()
event = MockEvent(testdata, path="/v2/models/m1/explain")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data["outputs"][
"explained"] == 5, f"wrong explain response {resp.body}"
def run_model(url, expected):
url = f"/v2/models/{url}/infer" if url else "/v2/models/infer"
event = MockEvent(testdata, path=url)
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data["outputs"] == {
"predictions": [expected]
}, f"wrong model response {data['outputs']}"
def test_ensemble_get_models():
context = init_ctx(ensemble_spec)
event = MockEvent("", path="/v2/models/", method="GET")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
# expected: {"models": ["m1", "m2", "m3:v1", "m3:v2", "VotingEnsemble"]}
assert len(data["models"]) == 5, f"wrong get models response {resp.body}"
def test_v2_async_mode():
# model loading is async
os.environ["SERVING_SPEC_ENV"] = json.dumps(asyncspec)
context = GraphContext()
nuclio_init_hook(context, globals(), serving_subkind)
context.logger.info("model initialized")
context.logger.info("test not ready, should return err 408")
event = MockEvent("", path="/v2/models/m5/ready", method="GET")
resp = context.mlrun_handler(context, event)
assert (
resp.status_code == 408
), f"didnt get proper ready resp, expected 408, got {resp.status_code}"
event = MockEvent(testdata, path="/v2/models/m5/infer")
resp = context.mlrun_handler(context, event)
context.logger.info("model responded")
logger.info(resp)
assert (resp.status_code !=
200), f"expected failure, got {resp.status_code} {resp.body}"
event = MockEvent('{"model": "m5", "inputs": [5]}')
event.trigger = "stream"
resp = context.mlrun_handler(context, event)
context.logger.info("model responded")
logger.info(resp)
data = json.loads(resp.body)
assert data["outputs"] == 5, f"wrong model response {data}"
def get_model(name, version, url):
event = MockEvent("", path=f"/v2/models/{url}", method="GET")
resp = context.mlrun_handler(context, event)
logger.info(f"resp: {resp}")
data = json.loads(resp.body)
# expected: {"name": "m3", "version": "v2", "inputs": [], "outputs": []}
assert (data["name"] == name and data["version"]
== version), f"wrong get model meta response {resp.body}"
def test_v2_raised_err():
os.environ["SERVING_SPEC_ENV"] = json.dumps(raiser_spec)
context = GraphContext()
nuclio_init_hook(context, globals(), serving_subkind)
event = MockEvent(testdata, path="/v2/models/m6/infer")
resp = context.mlrun_handler(context, event)
context.logger.info(f"model responded: {resp}")
assert resp.status_code == 400, "expecting Response() with status_code 400"
def test_v2_get_models():
context = init_ctx()
e = MockEvent("", path="/v2/models/", method="GET")
resp = context.mlrun_handler(context, e)
data = json.loads(resp.body)
# expected: {"models": ["m1", "m2", "m3:v1", "m3:v2"]}
assert len(data["models"]) == 4, f"wrong get models response {resp.body}"
def test_v2_model_ready():
context = init_ctx()
event = MockEvent("", path="/v2/models/m1/ready", method="GET")
resp = context.mlrun_handler(context, event)
assert resp.status_code == 200, f"didnt get proper ready resp {resp.body}"
def test_v2_custom_handler():
context = init_ctx()
event = MockEvent('{"test": "ok"}', path="/v2/models/m1/myop")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data["test"] == "ok", f"wrong custom op response {resp.body}"
def run_model(url, expected):
event = MockEvent(testdata, path=f"/v2/models/{url}/infer")
resp = context.mlrun_handler(context, event)
data = json.loads(resp.body)
assert data[
"outputs"] == expected, f"wrong model response {data['outputs']}"