def create_kubernetes_connection(cleanup=True,
start_clipper=True,
connect=True,
with_proxy=False,
num_frontend_replicas=1):
logger.info("Creating KubernetesContainerManager")
if with_proxy:
cm = KubernetesContainerManager(kubernetes_proxy_addr="127.0.0.1:8080")
else:
cm = KubernetesContainerManager()
cl = ClipperConnection(cm)
if cleanup:
cl.stop_all()
# Give kubernetes some time to clean up
time.sleep(20)
logger.info("Done cleaning up clipper")
if start_clipper:
logger.info("Starting Clipper")
cl.start_clipper(
query_frontend_image=
"568959175238.dkr.ecr.us-west-1.amazonaws.com/clipper/query_frontend:{}".
format(clipper_version),
mgmt_frontend_image=
"568959175238.dkr.ecr.us-west-1.amazonaws.com/clipper/management_frontend:{}".
format(clipper_version),
num_frontend_replicas=num_frontend_replicas)
time.sleep(1)
if connect:
try:
cl.connect()
except Exception:
pass
except ClipperException:
pass
return cl
def create_kubernetes_connection(cleanup=False,
start_clipper=False,
connect=False,
with_proxy=False,
num_frontend_replicas=1,
cleanup_name='default-cluster',
new_name='default-cluster',
connect_name='default-cluster',
service_types=None,
namespace='default'):
logger.info("Creating KubernetesContainerManager")
cl = None
assert cleanup or start_clipper or connect, "You must set at least one of {cleanup, start_clipper, connect} to be true."
if with_proxy:
kubernetes_proxy_addr = "127.0.0.1:8080"
else:
kubernetes_proxy_addr = None
if cleanup:
logger.info("Cleaning up Kubernetes Cluster {}".format(cleanup_name))
cm = KubernetesContainerManager(
cluster_name=cleanup_name,
useInternalIP=USE_MINIKUBE,
service_types=service_types,
kubernetes_proxy_addr=kubernetes_proxy_addr)
cl = ClipperConnection(cm)
cl.stop_all()
logger.info("Done cleaning up clipper")
if start_clipper:
logger.info("Starting up Kubernetes Cluster {}".format(new_name))
cm = KubernetesContainerManager(
cluster_name=new_name,
kubernetes_proxy_addr=kubernetes_proxy_addr,
namespace=namespace,
useInternalIP=USE_MINIKUBE,
service_types=service_types,
create_namespace_if_not_exists=True)
cl = ClipperConnection(cm)
cl.start_clipper(num_frontend_replicas=num_frontend_replicas)
if connect:
try:
cm = KubernetesContainerManager(
cluster_name=connect_name,
useInternalIP=USE_MINIKUBE,
service_types=service_types,
kubernetes_proxy_addr=kubernetes_proxy_addr)
cl = ClipperConnection(cm)
cl.connect()
except Exception:
pass
return cl
def signal_handler(signal, frame):
print("Stopping Clipper...")
clipper_conn = ClipperConnection(
KubernetesContainerManager(useInternalIP=True))
# clipper_conn = ClipperConnection(DockerContainerManager())
clipper_conn.stop_all()
sys.exit(0)
def main(version, label):
from clipper_admin import ClipperConnection, KubernetesContainerManager
clipper_conn = ClipperConnection(
KubernetesContainerManager(useInternalIP=True))
clipper_conn.connect()
from clipper_admin.deployers import python as python_deployer
registry = 'localhost:5000'
python_deployer.deploy_python_closure(clipper_conn,
name="sum-model",
version=version,
input_type="doubles",
func=feature_sum,
labels=[label],
registry=registry)
# turn into predictions
with open('submission.csv') as fh:
lines = fh.readlines()[1:] # ignore first line
preds = [line.strip().split(',')[1] for line in lines]
return preds
# pickle function and write to appropriate location
s = six.StringIO()
c = CloudPickler(s, 2)
c.dump(libffm)
serialized_prediction_function = s.getvalue()
filepath = 'docker/lib/func.pkl'
with open(filepath, 'w') as fh:
fh.write(serialized_prediction_function)
# refresh creds
os.system('gcloud container clusters get-credentials redis-cluster')
os.system('kubectl cluster-info')
clipper_conn = ClipperConnection(KubernetesContainerManager(clipper_ip, useInternalIP=True))
clipper_conn.connect()
# Build model and deploy to clipper
version = int(time.time())
clipper_conn.build_and_deploy_model('ffm', version, 'strings', 'docker/lib', 'clipper/python-closure-container:develop', container_registry='ryanhoque')
# Uncomment the following if first time
#clipper_conn.link_model_to_app(app_name="testbed", model_name="ffm")
# finally deploy new version of model to clipper (set version as timestamp)
print('Successfully deployed model ffm version ' + str(version) + ' to Clipper.')
saver.restore(sess, './model.ckpt')
load_infer_op = graph.get_tensor_by_name('probabilities:0')
accuracy_op = graph.get_tensor_by_name('Mean_1:0')
oX = graph.get_tensor_by_name('Placeholder:0')
oY = graph.get_tensor_by_name('Placeholder_1:0')
def predict(X):
print("inputs {}".format(X))
result = sess.run(load_infer_op, feed_dict={oX: X})
ret = [str(i) for i in result]
print("return is {}".format(ret))
return ret
manager = KubernetesContainerManager(kubernetes_proxy_addr=K8S_ADDR,
namespace=K8S_NS)
clipper_conn = ClipperConnection(manager)
clipper_conn.connect()
# clipper_conn.delete_application(APP_NAME)
# clipper_conn.register_application(
# name = APP_NAME, input_type = 'doubles', default_output = '0', slo_micros = 100000000)
deploy_tensorflow_model(clipper_conn,
name=PREDICT_NAME,
version=VERSION,
input_type="doubles",
func=predict,
tf_sess_or_saved_model_path=sess,
registry=REGISTRY,
pkgs_to_install=['tensorflow'])
import argparse
from clipper_admin import ClipperConnection, KubernetesContainerManager
manager = KubernetesContainerManager(
kubernetes_proxy_addr='127.0.0.1:8001', namespace='mdt')
clipper_conn = ClipperConnection(manager)
def deploy():
clipper_conn.start_clipper()
def undeploy():
clipper_conn.stop_all()
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='Deploy clippers')
parser.add_argument('--op', help='opertations support d|u|r')
args = parser.parse_args()
if args.op == 'd':
deploy()
elif args.op == 'u':
undeploy()
else:
print("unsupported operation {}".format(args.op))
from clipper_admin import ClipperConnection, KubernetesContainerManager
from clipper_admin.deployers import python as python_deployer
import requests, json, numpy as np
import time
clipper_conn = ClipperConnection(
KubernetesContainerManager("https://35.197.66.133", useInternalIP=True))
clipper_conn.connect()
addr = clipper_conn.get_query_addr()
headers = {"Content-type": "application/json"}
# sample from 5 arbitrary valid feature vectors
possible_data = [
"27,17,45,28,2,28,27,29,28,1,1,,23,68fd1e64,960c983b,9fbfbfd5,38c11726,25c83c98,7e0ccccf,fe06fd10,062b5529,a73ee510,ca53fc84,67360210,895d8bbb,4f8e2224,f862f261,b4cc2435,4c0041e5,e5ba7672,b4abdd09,21ddcdc9,5840adea,36a7ab86,,32c7478e,85e4d73f,010f6491,ee63dd9b",
"1,1,19,7,1,3,1,7,7,1,1,,2,09ca0b81,8947f767,a87e61f7,c4ba2a67,25c83c98,7e0ccccf,ce6020cc,062b5529,a73ee510,b04d3cfe,70dcd184,899eb56b,aca22cf9,b28479f6,a473257f,88f592e4,d4bb7bd8,bd17c3da,1d04f4a4,a458ea53,82bdc0bb,,32c7478e,5bdcd9c4,010f6491,cca57dcc",
"8,11,38,9,316,25,8,11,10,1,1,,9,05db9164,09e68b86,aa8c1539,85dd697c,25c83c98,7e0ccccf,bc252bd0,5b392875,a73ee510,ef5c0d3c,0bd0c3b3,d8c29807,c0e6befc,8ceecbc8,d2f03b75,c64d548f,e5ba7672,63cdbb21,cf99e5de,5840adea,5f957280,,55dd3565,1793a828,e8b83407,b7d9c3bc",
",4,13,20,17700,,0,20,1,,0,,20,68fd1e64,08d6d899,9143c832,f56b7dd5,0942e0a7,7e0ccccf,e88f1cec,0b153874,a73ee510,3b08e48b,8f410860,ae1bb660,b8eec0b1,b28479f6,bffbd637,bad5ee18,776ce399,bbf70d82,,,0429f84b,,be7c41b4,c0d61a5c,,",
"16,18,5203,8,0,0,4,49,10,0,1,,0,05db9164,9f7e1d07,0253bbf5,d6420627,4cf72387,,0db090eb,0b153874,a73ee510,3b08e48b,10e6a64f,31adfaee,38b5339a,07d13a8f,3e25e5f5,1621c7f4,e5ba7672,6a58e423,21ddcdc9,5840adea,bcc7a461,,32c7478e,3214afd4,ea9a246c,e7ecb821"
]
fh = open('predictions.log', 'a')
while (True):
# poisson process with rate 0.1
interarrival_time = np.random.exponential(scale=10)
time.sleep(interarrival_time)
num_requests = int(np.random.random() * 10) + 1 # uniform between 1 and 10
input = list()
for _ in range(num_requests):
input.append(possible_data[int(np.random.random() * 5)])
headers = {'Content-type': 'application/json'}
r = requests.post(url, headers=headers, data=req_json)
# print(json.loads(r.text))
str_r = json.loads(r.text)["output"].replace("[", "").replace("]",
"").split()
vals = [float(i) for i in str_r]
new_lineage_objs = [
Lineage.add_node(input_lin[i], model_name, vals[i])
for i in range(len(input_lin))
]
return new_lineage_objs
#setup Clipper connection
clipper_conn = ClipperConnection(
KubernetesContainerManager(useInternalIP=True))
clipper_conn.connect()
batch_size = 2
# batches = np.array([])
# times = np.array([])
for i in range(20):
# batch_size = np.random.randint(5, high=50)
print("request " + str(i))
if batch_size > 1:
input_list = [
Lineage(np.random.random_sample()) for i in range(batch_size)
]
out_lin_1 = predict(clipper_conn.get_query_addr(),
"lineage1",
input_list,
X = ckd.iloc[:300, ]
X = X.loc[:, X.columns != 'classification']
X.fillna(0, inplace=True)
y = ckd.iloc[:300, -1]
# train a classifier
model = GradientBoostingClassifier(random_state=2019)
model.fit(X, y)
# First we need to import Clipper
from clipper_admin import ClipperConnection, KubernetesContainerManager
from clipper_admin.deployers.python import deploy_python_closure
# Create a Clipper connection
clipper_conn = ClipperConnection(
KubernetesContainerManager(useInternalIP=True,
kubernetes_proxy_addr="127.0.0.1:8080"))
# Start a Clipper cluster or connect to a running one
clipper_conn.start_clipper()
# Register an app called 'kddtutorial'. This would create a REST endpoint
clipper_conn.register_application(name="kddtutorial",
input_type="doubles",
default_output="-1.0",
slo_micros=10000000)
# Access the trained model via closure capture
def predict(inputs):
global model
pred = model.predict(inputs)
#!/usr/bin/env python
# Start long-lived Clipper cluster
import os, sys
from clipper_admin import ClipperConnection, KubernetesContainerManager
from clipper_admin.deployers import python as python_deployer
if len(sys.argv) != 3:
print('Usage: python start-clipper.py clipper-cluster-IP redis-service-IP')
print("For example, python start-clipper.py 35.197.66.133 10.59.247.82")
sys.exit(1)
clipper_ip = 'https://' + sys.argv[1]
redis_ip = 'https://' + sys.argv[2]
clipper_conn = ClipperConnection(
KubernetesContainerManager(clipper_ip, useInternalIP=True))
try:
clipper_conn.stop_all()
clipper_conn.stop_all_model_containers()
clipper_conn.start_clipper()
clipper_conn.register_application(name="testbed",
input_type="strings",
default_output="-1.0",
slo_micros=100000000)
clipper_conn.get_all_apps()
except Exception as e:
print(e)