delete katib

README.md

@@ -10,7 +10,7 @@ cube-studio is a one-stop cloud-native machine learning platform open sourced by
  - 1、data management: feature store, online and offline features; dataset management, structure data and media data, data label platform
  - 2、develop: notebook(vscode/jupyter); docker image management; image build online
  - 3、train: pipeline drag and drop online; open template market; distributed computing/training tasks, example tf/pytorch/mxnet/spark/ray/horovod/kaldi/volcano; batch priority scheduling; resource monitoring/alarm/balancing; cron scheduling
- - 4、automl: nni, katib, ray
+ - 4、automl: nni, ray
  - 5、inference: model manager; serverless traffic control; tf/pytorch/onnx/tensorrt model deploy, tfserving/torchserver/onnxruntime/triton inference; VGPU; load balancing、high availability、elastic scaling
  - 6、infra: multi-user; multi-project; multi-cluster; edge cluster mode; blockchain sharing;
 

README_CN.md

@@ -9,7 +9,7 @@ cube是 腾讯音乐 开源的一站式云原生机器学习平台，目前主
  - 1、数据管理：特征平台，支持在/离线特征；数据源管理，支持结构数据和媒体标注数据管理；
  - 2、在线开发：在线的vscode/jupyter代码开发；在线镜像调试，支持免dockerfile，增量构建；
  - 3、训练编排：任务流编排，在线拖拉拽；开放的模板市场，支持tf/pytorch/mxnet/spark/ray/horovod/kaldi/volcano等分布式计算/训练任务；task的单节点debug，分布式任务的批量优先级调度，聚合日志；任务运行资源监控，报警；定时调度，支持补录，忽略，重试，依赖，并发限制，定时任务算力的智能修正；
- - 4、超参搜索：nni，katib，ray的超参搜索；
+ - 4、超参搜索：nni，ray的超参搜索；
  - 5、推理服务：tf/pytorch/onnx模型的推理服务，serverless流量管控，triton gpu推理加速，依据gpu利用率/qps等指标的hpa能力，虚拟化gpu，虚拟显存等服务化能力；
  - 6、资源统筹：多集群多项目组资源统筹，联邦调度，边缘计算；
 

install/README.md

@@ -6,7 +6,7 @@
 完整的平台包含
  - 1、机器的标准化
  - 2、分布式存储(单机可忽略)、k8s集群、监控体系(prometheus/efk/zipkin)
- - 3、基础能力(tf/pytorch/mxnet/valcano/ray等分布式，nni/katib超参搜索)
+ - 3、基础能力(tf/pytorch/mxnet/valcano/ray等分布式，nni/ray超参搜索)
  - 4、平台web部分(oa/权限/项目组、在线构建镜像、在线开发、pipeline拖拉拽、超参搜索、推理服务管理等)
 
 # 组件说明

install/docker/config.py

@@ -744,8 +744,8 @@ ADMIN_USER='admin'
 PIPELINE_NAMESPACE = 'pipeline'
 # 服务pipeline运行的空间，必填service
 SERVICE_PIPELINE_NAMESPACE='service'
-# 超参搜索命名空间，必填katib
-AUTOML_NAMESPACE = 'katib'
+# 超参搜索命名空间，必填automl
+AUTOML_NAMESPACE = 'automl'
 # notebook必填空间，必填jupyter
 NOTEBOOK_NAMESPACE = 'jupyter'
 # 内部服务命名空间，必填service
@@ -753,12 +753,12 @@ SERVICE_NAMESPACE = 'service'
 # 服务链路追踪地址
 SERVICE_PIPELINE_ZIPKIN='http://xx.xx.xx.xx:9401'
 SERVICE_PIPELINE_JAEGER='tracing.service'
-# katib任务默认镜像
-KATIB_JOB_DEFAULT_IMAGE='ccr.ccs.tencentyun.com/cube-studio/katib'
-# katib的tfjob任务默认镜像
-KATIB_TFJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/tf-mnist-with-summaries:1.0'
-# katib的pytorchjob任务默认镜像
-KATIB_PYTORCHJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/pytorch-dist-mnist-test:v1.0'
+# automl任务默认镜像
+AUTOML_JOB_DEFAULT_IMAGE='ccr.ccs.tencentyun.com/cube-studio/automl'
+# automl的tfjob任务默认镜像
+AUTOML_TFJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/tf-mnist-with-summaries:1.0'
+# automl的pytorchjob任务默认镜像
+AUTOML_PYTORCHJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/pytorch-dist-mnist-test:v1.0'
 # 拉取私有仓库镜像默认携带的k8s hubsecret名称
 HUBSECRET = ['hubsecret']
 # 私有仓库的组织名，用户在线构建的镜像自动推送这个组织下面
@@ -805,7 +805,6 @@ K8S_DASHBOARD_CLUSTER = '/k8s/dashboard/cluster/'  #
 K8S_DASHBOARD_PIPELINE = '/k8s/dashboard/pipeline/'
 
 PIPELINE_URL = '/pipeline/#/'
-KATIB_URL = '/katib/#'
 
 # 这两部分功能需要泛化域名。没有泛化域名此部分功能受限。ISTIO_INGRESS_DOMAIN为泛域名后缀
 ISTIO_INGRESS_DOMAIN = os.getenv('ISTIO_INGRESS_DOMAIN','local.com')  #  泛化域名，尾缀，可以和HOST不一致，没有泛化域名对应的功能没法使用

install/kubernetes/create_ns_secret.sh

@@ -1,5 +1,5 @@
 
-for namespace in 'infra' 'kubeflow' 'istio-system' 'pipeline' 'katib' 'jupyter' 'service' 'monitoring' 'logging' 'kube-system'
+for namespace in 'infra' 'kubeflow' 'istio-system' 'pipeline' 'automl' 'jupyter' 'service' 'monitoring' 'logging' 'kube-system'
 do
     kubectl create ns $namespace
     kubectl delete secret docker-registry hubsecret -n $namespace
@@ -9,7 +9,6 @@ do
 #    kubectl label namespace $namespace istio-inhection=enabled --overwrite
 done
 
-kubectl label ns katib katib-metricscollector-injection=enabled --overwrite
 kubectl label ns service istio-injection-
 
 

install/kubernetes/cube/base/virtual.yaml

@@ -1,4 +1,4 @@
-# 通过vs代理访问其他的服务，包括katib、pipline,argp,minio,grafana等
+# 通过vs代理访问其他的服务，包括pipline,argp,minio,grafana等
 apiVersion: networking.istio.io/v1alpha3
 kind: VirtualService
 metadata:

install/kubernetes/cube/overlays/config/config.py

@@ -567,7 +567,12 @@ class CeleryConfig(object):
             'task': 'task.adjust_node_resource',  # 定时在多项目组间进行资源均衡
             # 'schedule': 10.0,
             'schedule': crontab(minute='*/10'),
-        }
+        },
+        'task_update_aihub': {
+            'task': 'task.update_aihub',  # 更新aihub
+            # 'schedule': 10.0,
+            'schedule': crontab(minute='30', hour='4'),
+        },
     }
 
  # 帮助文档地址，显示在web导航栏
@@ -739,8 +744,8 @@ ADMIN_USER='admin'
 PIPELINE_NAMESPACE = 'pipeline'
 # 服务pipeline运行的空间，必填service
 SERVICE_PIPELINE_NAMESPACE='service'
-# 超参搜索命名空间，必填katib
-AUTOML_NAMESPACE = 'katib'
+# 超参搜索命名空间，必填automl
+AUTOML_NAMESPACE = 'automl'
 # notebook必填空间，必填jupyter
 NOTEBOOK_NAMESPACE = 'jupyter'
 # 内部服务命名空间，必填service
@@ -748,12 +753,12 @@ SERVICE_NAMESPACE = 'service'
 # 服务链路追踪地址
 SERVICE_PIPELINE_ZIPKIN='http://xx.xx.xx.xx:9401'
 SERVICE_PIPELINE_JAEGER='tracing.service'
-# katib任务默认镜像
-KATIB_JOB_DEFAULT_IMAGE='ccr.ccs.tencentyun.com/cube-studio/katib'
-# katib的tfjob任务默认镜像
-KATIB_TFJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/tf-mnist-with-summaries:1.0'
-# katib的pytorchjob任务默认镜像
-KATIB_PYTORCHJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/pytorch-dist-mnist-test:v1.0'
+# automl任务默认镜像
+AUTOML_JOB_DEFAULT_IMAGE='ccr.ccs.tencentyun.com/cube-studio/automl'
+# automl的tfjob任务默认镜像
+AUTOML_TFJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/tf-mnist-with-summaries:1.0'
+# automl的pytorchjob任务默认镜像
+AUTOML_PYTORCHJOB_DEFAULT_IMAGE = 'gcr.io/kubeflow-ci/pytorch-dist-mnist-test:v1.0'
 # 拉取私有仓库镜像默认携带的k8s hubsecret名称
 HUBSECRET = ['hubsecret']
 # 私有仓库的组织名，用户在线构建的镜像自动推送这个组织下面
@@ -800,7 +805,6 @@ K8S_DASHBOARD_CLUSTER = '/k8s/dashboard/cluster/'  #
 K8S_DASHBOARD_PIPELINE = '/k8s/dashboard/pipeline/'
 
 PIPELINE_URL = '/pipeline/#/'
-KATIB_URL = '/katib/#'
 
 # 这两部分功能需要泛化域名。没有泛化域名此部分功能受限。ISTIO_INGRESS_DOMAIN为泛域名后缀
 ISTIO_INGRESS_DOMAIN = os.getenv('ISTIO_INGRESS_DOMAIN','local.com')  #  泛化域名，尾缀，可以和HOST不一致，没有泛化域名对应的功能没法使用

install/kubernetes/dashboard/v2.2.0-user.yaml

@@ -117,8 +117,8 @@ subjects:
 apiVersion: rbac.authorization.k8s.io/v1
 kind: RoleBinding
 metadata:
-  name: kubernetes-dashboard-katib
-  namespace: katib      # 生效空间
+  name: kubernetes-dashboard-automl
+  namespace: automl      # 生效空间
 roleRef:
   apiGroup: rbac.authorization.k8s.io
   kind: ClusterRole

install/kubernetes/frameworkcontroller/readme.md

@@ -2,11 +2,11 @@
 ```bash
 kubectl create serviceaccount frameworkcontroller --namespace service
 kubectl create serviceaccount frameworkcontroller --namespace pipeline
-kubectl create serviceaccount frameworkcontroller --namespace katib
+kubectl create serviceaccount frameworkcontroller --namespace automl
 kubectl create serviceaccount frameworkcontroller --namespace kubeflow
 kubectl create clusterrolebinding frameworkcontroller-service --clusterrole=cluster-admin --user=system:serviceaccount:service:frameworkcontroller
 kubectl create clusterrolebinding frameworkcontroller-pipeline --clusterrole=cluster-admin --user=system:serviceaccount:pipeline:frameworkcontroller
-kubectl create clusterrolebinding frameworkcontroller-katib --clusterrole=cluster-admin --user=system:serviceaccount:katib:frameworkcontroller
+kubectl create clusterrolebinding frameworkcontroller-automl --clusterrole=cluster-admin --user=system:serviceaccount:automl:frameworkcontroller
 kubectl create clusterrolebinding frameworkcontroller-kubeflow --clusterrole=cluster-admin --user=system:serviceaccount:kubeflow:frameworkcontroller
 
 ```
@@ -18,12 +18,12 @@ kubectl create -f frameworkcontroller-with-default-config.yaml
 
 kubectl create serviceaccount frameworkbarrier --namespace service
 kubectl create serviceaccount frameworkbarrier --namespace pipeline
-kubectl create serviceaccount frameworkbarrier --namespace katib
+kubectl create serviceaccount frameworkbarrier --namespace automl
 kubectl create serviceaccount frameworkbarrier --namespace kubeflow
 kubectl create clusterrole frameworkbarrier --verb=get,list,watch --resource=frameworks
 kubectl create clusterrolebinding frameworkbarrier-service --clusterrole=frameworkbarrier  --user=system:serviceaccount:service:frameworkbarrier
 kubectl create clusterrolebinding frameworkbarrier-pipeline --clusterrole=frameworkbarrier  --user=system:serviceaccount:pipeline:frameworkbarrier
-kubectl create clusterrolebinding frameworkbarrier-katib --clusterrole=frameworkbarrier  --user=system:serviceaccount:katib:frameworkbarrier
+kubectl create clusterrolebinding frameworkbarrier-automl --clusterrole=frameworkbarrier  --user=system:serviceaccount:automl:frameworkbarrier
 kubectl create clusterrolebinding frameworkbarrier-kubeflow --clusterrole=frameworkbarrier  --user=system:serviceaccount:kubeflow:frameworkbarrier
 
 ```

install/kubernetes/ingress.yaml

@@ -59,10 +59,6 @@ spec:
   rules:
     - http:
         paths:
-          - path: /katib/
-            backend:
-              serviceName: katib-ui
-              servicePort: 80
           - path: /minio/
             backend:
               serviceName: minio-service

install/kubernetes/juicefs/cube-pv-pvc-with-juicefs/katib-pv-pvc-with-juicefs.yml

@@ -5,9 +5,9 @@
 apiVersion: v1
 kind: PersistentVolume
 metadata:
-  name: katib-kubeflow-user-workspace
+  name: automl-kubeflow-user-workspace
   labels:
-    katib-pvname: katib-kubeflow-user-workspace
+    automl-pvname: automl-kubeflow-user-workspace
 spec:
   capacity:
     storage: 500Gi
@@ -16,7 +16,7 @@ spec:
   persistentVolumeReclaimPolicy: Retain
   csi:
     driver: csi.juicefs.com
-    volumeHandle: katib-kubeflow-user-workspace
+    volumeHandle: automl-kubeflow-user-workspace
     fsType: juicefs
     nodePublishSecretRef:
       name: juicefs-sc-secret
@@ -33,7 +33,7 @@ kind: PersistentVolumeClaim
 apiVersion: v1
 metadata:
   name: kubeflow-user-workspace
-  namespace: katib
+  namespace: automl
 spec:
   accessModes:
     - ReadWriteMany
@@ -42,4 +42,4 @@ spec:
       storage: 500Gi
   selector:
     matchLabels:
-      katib-pvname: katib-kubeflow-user-workspace
+      automl-pvname: automl-kubeflow-user-workspace

install/kubernetes/ns.yaml

@@ -11,7 +11,7 @@ metadata:
 apiVersion: v1
 kind: Namespace
 metadata:
-  name: katib
+  name: automl
 
 ---
 apiVersion: v1

install/kubernetes/prometheus/grafana/dashboard/pod-info.json

@@ -84,7 +84,7 @@
       "targets": [
         {
           "exemplar": true,
-          "expr": "container_spec_memory_limit_bytes{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|katib\"}/1000/1000/1000",
+          "expr": "container_spec_memory_limit_bytes{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|automl\"}/1000/1000/1000",
           "format": "table",
           "hide": false,
           "interval": "",
@@ -94,7 +94,7 @@
         },
         {
           "exemplar": true,
-          "expr": "container_memory_working_set_bytes{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|katib\"}/1000/1000/1000",
+          "expr": "container_memory_working_set_bytes{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|automl\"}/1000/1000/1000",
           "format": "table",
           "hide": false,
           "instant": true,
@@ -105,7 +105,7 @@
         },
         {
           "exemplar": true,
-          "expr": "sum by (node,namespace,pod) (rate(container_cpu_usage_seconds_total{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|katib\"}[2m]))",
+          "expr": "sum by (node,namespace,pod) (rate(container_cpu_usage_seconds_total{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|automl\"}[2m]))",
           "format": "table",
           "hide": false,
           "instant": true,
@@ -115,7 +115,7 @@
         },
         {
           "exemplar": true,
-          "expr": "container_spec_cpu_quota{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|katib\"}/100000",
+          "expr": "container_spec_cpu_quota{job=\"kubelet\", image!=\"\",container_name!=\"POD\",pod!=\"\",namespace=~\"pipeline|jupyter|service|automl\"}/100000",
           "format": "table",
           "hide": false,
           "instant": true,

install/kubernetes/pv-pvc-automl.yaml

@@ -5,9 +5,9 @@
 apiVersion: v1
 kind: PersistentVolume
 metadata:
-  name: katib-kubeflow-user-workspace
+  name: automl-kubeflow-user-workspace
   labels:
-    katib-pvname: katib-kubeflow-user-workspace
+    automl-pvname: automl-kubeflow-user-workspace
 spec:
   capacity:
     storage: 500Gi
@@ -21,7 +21,7 @@ kind: PersistentVolumeClaim
 apiVersion: v1
 metadata:
   name: kubeflow-user-workspace
-  namespace: katib
+  namespace: automl
 spec:
   accessModes:
     - ReadWriteMany
@@ -30,4 +30,4 @@ spec:
       storage: 500Gi
   selector:
     matchLabels:
-      katib-pvname: katib-kubeflow-user-workspace
+      automl-pvname: automl-kubeflow-user-workspace

install/kubernetes/readme.md

@@ -134,8 +134,8 @@ kubectl create configmap kubernetes-config --from-file=kubeconfig -n infra
 kubectl delete configmap kubernetes-config -n pipeline
 kubectl create configmap kubernetes-config --from-file=kubeconfig -n pipeline
 
-kubectl delete configmap kubernetes-config -n katib
-kubectl create configmap kubernetes-config --from-file=kubeconfig -n katib
+kubectl delete configmap kubernetes-config -n automl
+kubectl create configmap kubernetes-config --from-file=kubeconfig -n automl
 ```
 
   
@@ -165,7 +165,7 @@ kubectl apply -k cube/overlays
 ```bash  
 kubectl create -f pv-pvc-infra.yaml  
 kubectl create -f pv-pvc-jupyter.yaml  
-kubectl create -f pv-pvc-katib.yaml  
+kubectl create -f pv-pvc-automl.yaml  
 # kubectl create -f pv-pvc-kubeflow.yaml  
 kubectl create -f pv-pvc-pipeline.yaml  
 kubectl create -f pv-pvc-service.yaml  

install/kubernetes/sa-rbac.yaml

@@ -37,16 +37,16 @@ apiVersion: v1
 kind: ServiceAccount
 metadata:
   name: nni
-  namespace: katib
+  namespace: automl
 ---
 kind: ClusterRoleBinding
 apiVersion: rbac.authorization.k8s.io/v1
 metadata:
-  name: katib-nni-clusterrolebinding
+  name: automl-nni-clusterrolebinding
 subjects:
 - kind: ServiceAccount
   name: nni
-  namespace: katib
+  namespace: automl
 roleRef:
   kind: ClusterRole
   name: kubeflow-clusterrole

install/kubernetes/start.sh

@@ -96,11 +96,11 @@ sleep 5
 kubectl wait crd/frameworks.frameworkcontroller.microsoft.com --for condition=established --timeout=60s
 
 kubectl create serviceaccount frameworkbarrier --namespace pipeline
-kubectl create serviceaccount frameworkbarrier --namespace katib
+kubectl create serviceaccount frameworkbarrier --namespace automl
 kubectl create serviceaccount frameworkbarrier --namespace kubeflow
 kubectl create clusterrole frameworkbarrier --verb=get,list,watch --resource=frameworks
 kubectl create clusterrolebinding frameworkbarrier-pipeline --clusterrole=frameworkbarrier  --user=system:serviceaccount:pipeline:frameworkbarrier
-kubectl create clusterrolebinding frameworkbarrier-katib --clusterrole=frameworkbarrier  --user=system:serviceaccount:katib:frameworkbarrier
+kubectl create clusterrolebinding frameworkbarrier-automl --clusterrole=frameworkbarrier  --user=system:serviceaccount:automl:frameworkbarrier
 kubectl create clusterrolebinding frameworkbarrier-kubeflow --clusterrole=frameworkbarrier  --user=system:serviceaccount:kubeflow:frameworkbarrier
 
 # 部署volcano
@@ -147,12 +147,12 @@ kubectl create configmap kubernetes-config --from-file=kubeconfig -n infra
 kubectl delete configmap kubernetes-config -n pipeline
 kubectl create configmap kubernetes-config --from-file=kubeconfig -n pipeline
 
-kubectl delete configmap kubernetes-config -n katib
-kubectl create configmap kubernetes-config --from-file=kubeconfig -n katib
+kubectl delete configmap kubernetes-config -n automl
+kubectl create configmap kubernetes-config --from-file=kubeconfig -n automl
 
 kubectl create -f pv-pvc-infra.yaml
 kubectl create -f pv-pvc-jupyter.yaml
-kubectl create -f pv-pvc-katib.yaml
+kubectl create -f pv-pvc-automl.yaml
 kubectl create -f pv-pvc-pipeline.yaml
 kubectl create -f pv-pvc-service.yaml
 

install/kubernetes/virtual.yaml

@@ -1,4 +1,4 @@
-## 通过vs代理访问其他的服务，包括katib、pipline,argp,minio,grafana等
+## 通过vs代理访问其他的服务，包括pipline,argp,minio,grafana等
 #apiVersion: networking.istio.io/v1alpha3
 #kind: VirtualService
 #metadata:
@@ -16,7 +16,7 @@
 #        port:
 #          number: 80
 #---
-# 通过vs代理访问其他的服务，包括katib、pipline,argp,minio,grafana等
+# 通过vs代理访问其他的服务，包括pipline,argp,minio,grafana等
 apiVersion: networking.istio.io/v1alpha3
 kind: VirtualService
 metadata:
@@ -47,24 +47,6 @@ spec:
       allowHeaders:
       - "*"
 
-  - match:
-    - uri:
-        prefix: /katib/
-    rewrite:
-      uri: /katib/
-    route:
-    - destination:
-        host: katib-ui.kubeflow.svc.cluster.local
-        port:
-          number: 80
-    corsPolicy:
-      allowOrigin:
-      - "*"
-      allowMethods:
-      - POST
-      - GET
-      allowHeaders:
-      - "*"
   - match:
     - uri:
         prefix: /pipeline/

myapp/models/model_nni.py

@@ -41,7 +41,7 @@ class NNI(Model,AuditMixinNullable,MyappModelBase):
         "Project", foreign_keys=[project_id]
     )
     name = Column(String(200), unique = True, nullable=False)
-    namespace = Column(String(200), nullable=True,default='katib')
+    namespace = Column(String(200), nullable=True,default='automl')
     describe = Column(Text)
     parallel_trial_count = Column(Integer,default=3)
     maxExecDuration =  Column(Integer,default=3600)

myapp/tasks/schedules.py

@@ -1048,7 +1048,7 @@ def watch_gpu(task):
         cluster = clusters[cluster_name]
         k8s_client = K8s(cluster.get('KUBECONFIG',''))
 
-        all_gpu_pods=k8s_client.get_uesd_gpu(namespaces=['pipeline','katib','jupyter','service'])
+        all_gpu_pods=k8s_client.get_uesd_gpu(namespaces=['pipeline','automl','jupyter','service'])
 
         print(all_gpu_pods)
         message = ''
@@ -1089,7 +1089,7 @@ def adjust_node_resource(task):
                     all_node_json[ip]['used_gpu'] = []
 
         # print(all_node_json)
-        for namespace in ['jupyter', 'pipeline', 'katib', 'service']:
+        for namespace in ['jupyter', 'pipeline', 'automl', 'service']:
             all_pods = k8s_client.get_pods(namespace=namespace)
             for pod in all_pods:
                 if pod['host_ip'] not in all_node_json:

myapp/tools/watch_pytorchjob.py

@@ -20,7 +20,6 @@ from myapp.models.model_job import (
 from myapp.utils.celery import session_scope
 from myapp.project import push_admin,push_message
 from myapp.models.model_job import Pipeline,Workflow
-from myapp.models.model_katib import Hyperparameter_Tuning,Experiments
 import pymysql
 conf=app.config
 

myapp/tools/watch_tfjob.py

@@ -20,7 +20,6 @@ from myapp.models.model_job import (
 from myapp.utils.celery import session_scope
 from myapp.project import push_admin,push_message
 from myapp.models.model_job import Pipeline,Workflow
-from myapp.models.model_katib import Hyperparameter_Tuning,Experiments
 import pymysql
 conf=app.config
 

myapp/views/home.py

@@ -763,7 +763,7 @@ class Myapp(BaseMyappView):
                         all_node_json[ip]['user'] = []
 
                 # print(all_node_json)
-                for namespace in ['jupyter', 'pipeline', 'katib', 'service']:
+                for namespace in ['jupyter', 'pipeline', 'automl', 'service']:
                     all_pods = k8s_client.get_pods(namespace=namespace)
                     for pod in all_pods:
                         if pod['status'] == 'Running' and pod['host_ip'] in all_node_json:
@@ -906,7 +906,7 @@ class Myapp(BaseMyappView):
                     # 获取pod的资源占用
                     all_tasks_json[cluster_name]={}
                     # print(all_node_json)
-                    for namespace in ['pipeline', 'katib', 'service']:
+                    for namespace in ['pipeline', 'automl', 'service']:
                         all_tasks_json[cluster_name][namespace]={}
                         all_pods = k8s_client.get_pods(namespace=namespace)
                         for pod in all_pods:

myapp/views/view_nni.py

@@ -351,7 +351,7 @@ class NNI_ModelView_Base():
 
         from myapp.utils.py.py_k8s import K8s
         k8s_client = K8s(nni.project.cluster.get('KUBECONFIG',''))
-        namespace = conf.get('AUTOML_NAMESPACE','katib')
+        namespace = conf.get('AUTOML_NAMESPACE','automl')
         run_id='nni-'+nni.name
 
         try:
@@ -546,7 +546,7 @@ tuner:
 trial:
   codeDir: {nni.working_dir}  
 frameworkcontrollerConfig:
-  namespace: {conf.get('AUTOML_NAMESPACE','katib')}
+  namespace: {conf.get('AUTOML_NAMESPACE','automl')}
   storage: pvc
   configPath: /mnt/{nni.created_by.username}/nni/{nni.name}/trial_template.yaml  
   pvc: