Module codeflare_sdk.utils.generate_yaml
This sub-module exists primarily to be used internally by the Cluster object (in the cluster sub-module) for AppWrapper generation.
Expand source code
# Copyright 2022 IBM, Red Hat
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This sub-module exists primarily to be used internally by the Cluster object
(in the cluster sub-module) for AppWrapper generation.
"""
import json
from typing import Optional
import typing
import yaml
import sys
import os
import argparse
import uuid
from kubernetes import client, config
from .kube_api_helpers import _kube_api_error_handling
from ..cluster.auth import get_api_client, config_check
from os import urandom
from base64 import b64encode
from urllib3.util import parse_url
from kubernetes.client.exceptions import ApiException
import codeflare_sdk
def read_template(template):
with open(template, "r") as stream:
try:
return yaml.safe_load(stream)
except yaml.YAMLError as exc:
print(exc)
def gen_names(name):
if not name:
gen_id = str(uuid.uuid4())
appwrapper_name = "appwrapper-" + gen_id
cluster_name = "cluster-" + gen_id
return appwrapper_name, cluster_name
else:
return name, name
# Check if the routes api exists
def is_openshift_cluster():
try:
config_check()
for api in client.ApisApi(get_api_client()).get_api_versions().groups:
for v in api.versions:
if "route.openshift.io/v1" in v.group_version:
return True
else:
return False
except Exception as e: # pragma: no cover
return _kube_api_error_handling(e)
def is_kind_cluster():
try:
config_check()
v1 = client.CoreV1Api()
label_selector = "kubernetes.io/hostname=kind-control-plane"
nodes = v1.list_node(label_selector=label_selector)
# If we find one or more nodes with the label, assume it's a KinD cluster
return len(nodes.items) > 0
except Exception as e:
print(f"Error checking if cluster is KinD: {e}")
return False
def update_names(
cluster_yaml: dict,
cluster: "codeflare_sdk.cluster.Cluster",
):
metadata = cluster_yaml.get("metadata")
metadata["name"] = cluster.config.name
metadata["namespace"] = cluster.config.namespace
def update_image(spec, image):
containers = spec.get("containers")
if image != "":
for container in containers:
container["image"] = image
def update_image_pull_secrets(spec, image_pull_secrets):
template_secrets = spec.get("imagePullSecrets", [])
spec["imagePullSecrets"] = template_secrets + [
{"name": x} for x in image_pull_secrets
]
def update_env(spec, env):
containers = spec.get("containers")
for container in containers:
if env:
if "env" in container:
container["env"].extend(env)
else:
container["env"] = env
def update_resources(
spec,
cpu_requests,
cpu_limits,
memory_requests,
memory_limits,
custom_resources,
):
container = spec.get("containers")
for resource in container:
requests = resource.get("resources").get("requests")
if requests is not None:
requests["cpu"] = cpu_requests
requests["memory"] = memory_requests
limits = resource.get("resources").get("limits")
if limits is not None:
limits["cpu"] = cpu_limits
limits["memory"] = memory_limits
for k in custom_resources.keys():
limits[k] = custom_resources[k]
requests[k] = custom_resources[k]
def head_worker_gpu_count_from_cluster(
cluster: "codeflare_sdk.cluster.Cluster",
) -> typing.Tuple[int, int]:
head_gpus = 0
worker_gpus = 0
for k in cluster.config.head_extended_resource_requests.keys():
resource_type = cluster.config.extended_resource_mapping[k]
if resource_type == "GPU":
head_gpus += int(cluster.config.head_extended_resource_requests[k])
for k in cluster.config.worker_extended_resource_requests.keys():
resource_type = cluster.config.extended_resource_mapping[k]
if resource_type == "GPU":
worker_gpus += int(cluster.config.worker_extended_resource_requests[k])
return head_gpus, worker_gpus
FORBIDDEN_CUSTOM_RESOURCE_TYPES = ["GPU", "CPU", "memory"]
def head_worker_resources_from_cluster(
cluster: "codeflare_sdk.cluster.Cluster",
) -> typing.Tuple[dict, dict]:
to_return = {}, {}
for k in cluster.config.head_extended_resource_requests.keys():
resource_type = cluster.config.extended_resource_mapping[k]
if resource_type in FORBIDDEN_CUSTOM_RESOURCE_TYPES:
continue
to_return[0][resource_type] = cluster.config.head_extended_resource_requests[
k
] + to_return[0].get(resource_type, 0)
for k in cluster.config.worker_extended_resource_requests.keys():
resource_type = cluster.config.extended_resource_mapping[k]
if resource_type in FORBIDDEN_CUSTOM_RESOURCE_TYPES:
continue
to_return[1][resource_type] = cluster.config.worker_extended_resource_requests[
k
] + to_return[1].get(resource_type, 0)
return to_return
def update_nodes(
ray_cluster_dict: dict,
cluster: "codeflare_sdk.cluster.Cluster",
):
head = ray_cluster_dict.get("spec").get("headGroupSpec")
worker = ray_cluster_dict.get("spec").get("workerGroupSpecs")[0]
head_gpus, worker_gpus = head_worker_gpu_count_from_cluster(cluster)
head_resources, worker_resources = head_worker_resources_from_cluster(cluster)
head_resources = json.dumps(head_resources).replace('"', '\\"')
head_resources = f'"{head_resources}"'
worker_resources = json.dumps(worker_resources).replace('"', '\\"')
worker_resources = f'"{worker_resources}"'
head["rayStartParams"]["num-gpus"] = str(head_gpus)
head["rayStartParams"]["resources"] = head_resources
# Head counts as first worker
worker["replicas"] = cluster.config.num_workers
worker["minReplicas"] = cluster.config.num_workers
worker["maxReplicas"] = cluster.config.num_workers
worker["groupName"] = "small-group-" + cluster.config.name
worker["rayStartParams"]["num-gpus"] = str(worker_gpus)
worker["rayStartParams"]["resources"] = worker_resources
for comp in [head, worker]:
spec = comp.get("template").get("spec")
update_image_pull_secrets(spec, cluster.config.image_pull_secrets)
update_image(spec, cluster.config.image)
update_env(spec, cluster.config.envs)
if comp == head:
# TODO: Eventually add head node configuration outside of template
update_resources(
spec,
cluster.config.head_cpu_requests,
cluster.config.head_cpu_limits,
cluster.config.head_memory_requests,
cluster.config.head_memory_limits,
cluster.config.head_extended_resource_requests,
)
else:
update_resources(
spec,
cluster.config.worker_cpu_requests,
cluster.config.worker_cpu_limits,
cluster.config.worker_memory_requests,
cluster.config.worker_memory_limits,
cluster.config.worker_extended_resource_requests,
)
def del_from_list_by_name(l: list, target: typing.List[str]) -> list:
return [x for x in l if x["name"] not in target]
def get_default_kueue_name(namespace: str):
# If the local queue is set, use it. Otherwise, try to use the default queue.
try:
config_check()
api_instance = client.CustomObjectsApi(get_api_client())
local_queues = api_instance.list_namespaced_custom_object(
group="kueue.x-k8s.io",
version="v1beta1",
namespace=namespace,
plural="localqueues",
)
except ApiException as e: # pragma: no cover
if e.status == 404 or e.status == 403:
return
else:
return _kube_api_error_handling(e)
for lq in local_queues["items"]:
if (
"annotations" in lq["metadata"]
and "kueue.x-k8s.io/default-queue" in lq["metadata"]["annotations"]
and lq["metadata"]["annotations"]["kueue.x-k8s.io/default-queue"].lower()
== "true"
):
return lq["metadata"]["name"]
def local_queue_exists(namespace: str, local_queue_name: str):
# get all local queues in the namespace
try:
config_check()
api_instance = client.CustomObjectsApi(get_api_client())
local_queues = api_instance.list_namespaced_custom_object(
group="kueue.x-k8s.io",
version="v1beta1",
namespace=namespace,
plural="localqueues",
)
except Exception as e: # pragma: no cover
return _kube_api_error_handling(e)
# check if local queue with the name provided in cluster config exists
for lq in local_queues["items"]:
if lq["metadata"]["name"] == local_queue_name:
return True
return False
def add_queue_label(item: dict, namespace: str, local_queue: Optional[str]):
lq_name = local_queue or get_default_kueue_name(namespace)
if lq_name == None:
return
elif not local_queue_exists(namespace, lq_name):
raise ValueError(
"local_queue provided does not exist or is not in this namespace. Please provide the correct local_queue name in Cluster Configuration"
)
if not "labels" in item["metadata"]:
item["metadata"]["labels"] = {}
item["metadata"]["labels"].update({"kueue.x-k8s.io/queue-name": lq_name})
def augment_labels(item: dict, labels: dict):
if not "labels" in item["metadata"]:
item["metadata"]["labels"] = {}
item["metadata"]["labels"].update(labels)
def notebook_annotations(item: dict):
nb_prefix = os.environ.get("NB_PREFIX")
if nb_prefix:
if not "annotations" in item["metadata"]:
item["metadata"]["annotations"] = {}
item["metadata"]["annotations"].update(
{"app.kubernetes.io/managed-by": nb_prefix}
)
def wrap_cluster(cluster_yaml: dict, appwrapper_name: str, namespace: str):
return {
"apiVersion": "workload.codeflare.dev/v1beta2",
"kind": "AppWrapper",
"metadata": {"name": appwrapper_name, "namespace": namespace},
"spec": {"components": [{"template": cluster_yaml}]},
}
def write_user_yaml(user_yaml, output_file_name):
# Create the directory if it doesn't exist
directory_path = os.path.dirname(output_file_name)
if not os.path.exists(directory_path):
os.makedirs(directory_path)
with open(output_file_name, "w") as outfile:
yaml.dump(user_yaml, outfile, default_flow_style=False)
print(f"Written to: {output_file_name}")
def generate_appwrapper(cluster: "codeflare_sdk.cluster.Cluster"):
cluster_yaml = read_template(cluster.config.template)
appwrapper_name, _ = gen_names(cluster.config.name)
update_names(
cluster_yaml,
cluster,
)
update_nodes(cluster_yaml, cluster)
augment_labels(cluster_yaml, cluster.config.labels)
notebook_annotations(cluster_yaml)
user_yaml = (
wrap_cluster(cluster_yaml, appwrapper_name, cluster.config.namespace)
if cluster.config.appwrapper
else cluster_yaml
)
add_queue_label(user_yaml, cluster.config.namespace, cluster.config.local_queue)
if cluster.config.write_to_file:
directory_path = os.path.expanduser("~/.codeflare/resources/")
outfile = os.path.join(directory_path, appwrapper_name + ".yaml")
write_user_yaml(user_yaml, outfile)
return outfile
else:
user_yaml = yaml.dump(user_yaml)
print(f"Yaml resources loaded for {cluster.config.name}")
return user_yamlFunctions
def add_queue_label(item: dict, namespace: str, local_queue: Optional[str])-
Expand source code
def add_queue_label(item: dict, namespace: str, local_queue: Optional[str]): lq_name = local_queue or get_default_kueue_name(namespace) if lq_name == None: return elif not local_queue_exists(namespace, lq_name): raise ValueError( "local_queue provided does not exist or is not in this namespace. Please provide the correct local_queue name in Cluster Configuration" ) if not "labels" in item["metadata"]: item["metadata"]["labels"] = {} item["metadata"]["labels"].update({"kueue.x-k8s.io/queue-name": lq_name}) def augment_labels(item: dict, labels: dict)-
Expand source code
def augment_labels(item: dict, labels: dict): if not "labels" in item["metadata"]: item["metadata"]["labels"] = {} item["metadata"]["labels"].update(labels) def del_from_list_by_name(l: list, target: List[str]) ‑> list-
Expand source code
def del_from_list_by_name(l: list, target: typing.List[str]) -> list: return [x for x in l if x["name"] not in target] def gen_names(name)-
Expand source code
def gen_names(name): if not name: gen_id = str(uuid.uuid4()) appwrapper_name = "appwrapper-" + gen_id cluster_name = "cluster-" + gen_id return appwrapper_name, cluster_name else: return name, name def generate_appwrapper(cluster: codeflare_sdk.cluster.Cluster)-
Expand source code
def generate_appwrapper(cluster: "codeflare_sdk.cluster.Cluster"): cluster_yaml = read_template(cluster.config.template) appwrapper_name, _ = gen_names(cluster.config.name) update_names( cluster_yaml, cluster, ) update_nodes(cluster_yaml, cluster) augment_labels(cluster_yaml, cluster.config.labels) notebook_annotations(cluster_yaml) user_yaml = ( wrap_cluster(cluster_yaml, appwrapper_name, cluster.config.namespace) if cluster.config.appwrapper else cluster_yaml ) add_queue_label(user_yaml, cluster.config.namespace, cluster.config.local_queue) if cluster.config.write_to_file: directory_path = os.path.expanduser("~/.codeflare/resources/") outfile = os.path.join(directory_path, appwrapper_name + ".yaml") write_user_yaml(user_yaml, outfile) return outfile else: user_yaml = yaml.dump(user_yaml) print(f"Yaml resources loaded for {cluster.config.name}") return user_yaml def get_default_kueue_name(namespace: str)-
Expand source code
def get_default_kueue_name(namespace: str): # If the local queue is set, use it. Otherwise, try to use the default queue. try: config_check() api_instance = client.CustomObjectsApi(get_api_client()) local_queues = api_instance.list_namespaced_custom_object( group="kueue.x-k8s.io", version="v1beta1", namespace=namespace, plural="localqueues", ) except ApiException as e: # pragma: no cover if e.status == 404 or e.status == 403: return else: return _kube_api_error_handling(e) for lq in local_queues["items"]: if ( "annotations" in lq["metadata"] and "kueue.x-k8s.io/default-queue" in lq["metadata"]["annotations"] and lq["metadata"]["annotations"]["kueue.x-k8s.io/default-queue"].lower() == "true" ): return lq["metadata"]["name"] def head_worker_gpu_count_from_cluster(cluster: codeflare_sdk.cluster.Cluster) ‑> Tuple[int, int]-
Expand source code
def head_worker_gpu_count_from_cluster( cluster: "codeflare_sdk.cluster.Cluster", ) -> typing.Tuple[int, int]: head_gpus = 0 worker_gpus = 0 for k in cluster.config.head_extended_resource_requests.keys(): resource_type = cluster.config.extended_resource_mapping[k] if resource_type == "GPU": head_gpus += int(cluster.config.head_extended_resource_requests[k]) for k in cluster.config.worker_extended_resource_requests.keys(): resource_type = cluster.config.extended_resource_mapping[k] if resource_type == "GPU": worker_gpus += int(cluster.config.worker_extended_resource_requests[k]) return head_gpus, worker_gpus def head_worker_resources_from_cluster(cluster: codeflare_sdk.cluster.Cluster) ‑> Tuple[dict, dict]-
Expand source code
def head_worker_resources_from_cluster( cluster: "codeflare_sdk.cluster.Cluster", ) -> typing.Tuple[dict, dict]: to_return = {}, {} for k in cluster.config.head_extended_resource_requests.keys(): resource_type = cluster.config.extended_resource_mapping[k] if resource_type in FORBIDDEN_CUSTOM_RESOURCE_TYPES: continue to_return[0][resource_type] = cluster.config.head_extended_resource_requests[ k ] + to_return[0].get(resource_type, 0) for k in cluster.config.worker_extended_resource_requests.keys(): resource_type = cluster.config.extended_resource_mapping[k] if resource_type in FORBIDDEN_CUSTOM_RESOURCE_TYPES: continue to_return[1][resource_type] = cluster.config.worker_extended_resource_requests[ k ] + to_return[1].get(resource_type, 0) return to_return def is_kind_cluster()-
Expand source code
def is_kind_cluster(): try: config_check() v1 = client.CoreV1Api() label_selector = "kubernetes.io/hostname=kind-control-plane" nodes = v1.list_node(label_selector=label_selector) # If we find one or more nodes with the label, assume it's a KinD cluster return len(nodes.items) > 0 except Exception as e: print(f"Error checking if cluster is KinD: {e}") return False def is_openshift_cluster()-
Expand source code
def is_openshift_cluster(): try: config_check() for api in client.ApisApi(get_api_client()).get_api_versions().groups: for v in api.versions: if "route.openshift.io/v1" in v.group_version: return True else: return False except Exception as e: # pragma: no cover return _kube_api_error_handling(e) def local_queue_exists(namespace: str, local_queue_name: str)-
Expand source code
def local_queue_exists(namespace: str, local_queue_name: str): # get all local queues in the namespace try: config_check() api_instance = client.CustomObjectsApi(get_api_client()) local_queues = api_instance.list_namespaced_custom_object( group="kueue.x-k8s.io", version="v1beta1", namespace=namespace, plural="localqueues", ) except Exception as e: # pragma: no cover return _kube_api_error_handling(e) # check if local queue with the name provided in cluster config exists for lq in local_queues["items"]: if lq["metadata"]["name"] == local_queue_name: return True return False def notebook_annotations(item: dict)-
Expand source code
def notebook_annotations(item: dict): nb_prefix = os.environ.get("NB_PREFIX") if nb_prefix: if not "annotations" in item["metadata"]: item["metadata"]["annotations"] = {} item["metadata"]["annotations"].update( {"app.kubernetes.io/managed-by": nb_prefix} ) def read_template(template)-
Expand source code
def read_template(template): with open(template, "r") as stream: try: return yaml.safe_load(stream) except yaml.YAMLError as exc: print(exc) def update_env(spec, env)-
Expand source code
def update_env(spec, env): containers = spec.get("containers") for container in containers: if env: if "env" in container: container["env"].extend(env) else: container["env"] = env def update_image(spec, image)-
Expand source code
def update_image(spec, image): containers = spec.get("containers") if image != "": for container in containers: container["image"] = image def update_image_pull_secrets(spec, image_pull_secrets)-
Expand source code
def update_image_pull_secrets(spec, image_pull_secrets): template_secrets = spec.get("imagePullSecrets", []) spec["imagePullSecrets"] = template_secrets + [ {"name": x} for x in image_pull_secrets ] def update_names(cluster_yaml: dict, cluster: codeflare_sdk.cluster.Cluster)-
Expand source code
def update_names( cluster_yaml: dict, cluster: "codeflare_sdk.cluster.Cluster", ): metadata = cluster_yaml.get("metadata") metadata["name"] = cluster.config.name metadata["namespace"] = cluster.config.namespace def update_nodes(ray_cluster_dict: dict, cluster: codeflare_sdk.cluster.Cluster)-
Expand source code
def update_nodes( ray_cluster_dict: dict, cluster: "codeflare_sdk.cluster.Cluster", ): head = ray_cluster_dict.get("spec").get("headGroupSpec") worker = ray_cluster_dict.get("spec").get("workerGroupSpecs")[0] head_gpus, worker_gpus = head_worker_gpu_count_from_cluster(cluster) head_resources, worker_resources = head_worker_resources_from_cluster(cluster) head_resources = json.dumps(head_resources).replace('"', '\\"') head_resources = f'"{head_resources}"' worker_resources = json.dumps(worker_resources).replace('"', '\\"') worker_resources = f'"{worker_resources}"' head["rayStartParams"]["num-gpus"] = str(head_gpus) head["rayStartParams"]["resources"] = head_resources # Head counts as first worker worker["replicas"] = cluster.config.num_workers worker["minReplicas"] = cluster.config.num_workers worker["maxReplicas"] = cluster.config.num_workers worker["groupName"] = "small-group-" + cluster.config.name worker["rayStartParams"]["num-gpus"] = str(worker_gpus) worker["rayStartParams"]["resources"] = worker_resources for comp in [head, worker]: spec = comp.get("template").get("spec") update_image_pull_secrets(spec, cluster.config.image_pull_secrets) update_image(spec, cluster.config.image) update_env(spec, cluster.config.envs) if comp == head: # TODO: Eventually add head node configuration outside of template update_resources( spec, cluster.config.head_cpu_requests, cluster.config.head_cpu_limits, cluster.config.head_memory_requests, cluster.config.head_memory_limits, cluster.config.head_extended_resource_requests, ) else: update_resources( spec, cluster.config.worker_cpu_requests, cluster.config.worker_cpu_limits, cluster.config.worker_memory_requests, cluster.config.worker_memory_limits, cluster.config.worker_extended_resource_requests, ) def update_resources(spec, cpu_requests, cpu_limits, memory_requests, memory_limits, custom_resources)-
Expand source code
def update_resources( spec, cpu_requests, cpu_limits, memory_requests, memory_limits, custom_resources, ): container = spec.get("containers") for resource in container: requests = resource.get("resources").get("requests") if requests is not None: requests["cpu"] = cpu_requests requests["memory"] = memory_requests limits = resource.get("resources").get("limits") if limits is not None: limits["cpu"] = cpu_limits limits["memory"] = memory_limits for k in custom_resources.keys(): limits[k] = custom_resources[k] requests[k] = custom_resources[k] def wrap_cluster(cluster_yaml: dict, appwrapper_name: str, namespace: str)-
Expand source code
def wrap_cluster(cluster_yaml: dict, appwrapper_name: str, namespace: str): return { "apiVersion": "workload.codeflare.dev/v1beta2", "kind": "AppWrapper", "metadata": {"name": appwrapper_name, "namespace": namespace}, "spec": {"components": [{"template": cluster_yaml}]}, } def write_user_yaml(user_yaml, output_file_name)-
Expand source code
def write_user_yaml(user_yaml, output_file_name): # Create the directory if it doesn't exist directory_path = os.path.dirname(output_file_name) if not os.path.exists(directory_path): os.makedirs(directory_path) with open(output_file_name, "w") as outfile: yaml.dump(user_yaml, outfile, default_flow_style=False) print(f"Written to: {output_file_name}")