from flask_appbuilder import Model from flask_babel import lazy_gettext as _ import re from myapp.utils import core class MyappModelBase(): label_columns={ "name": "名称", "name_url": "名称", "name_title": "名称(移动鼠标，查看描述)", 'job_type': '任务类型', "project": "项目组", "project_url":"项目组", "namespace": "命名空间", "namespace_url": "命名空间", "describe": "描述", "describe_url": "描述", "password": '密码', "workdir":"工作目录", "images": "镜像", "repository": "仓库", "args": "启动参数", "args_html": "启动参数", "demo": "参数示例", "demo_html": "参数示例", "entrypoint": '启动命令', "dockerfile": 'Dockerfile', "gitpath": 'git地址', "env": "环境变量", "privileged": "特权模式", "accounts": "k8s账号", "images_url": "镜像", "hostAliases": "host", "overwrite_entrypoint": "覆盖原始启动命令", "command": "启动命令", "working_dir": "启动目录", "volume_mount": "挂载目录", "node_selector": "调度机器", "image_pull_policy": "镜像拉取策略", "resource_memory": "内存申请", "resource_cpu": "cpu申请", "resource_gpu": "gpu申请", "resource":"资源", "timeout": "超时中断", "retry": "重试次数", "outputs": "输出", "version": "版本", "model_name":"模型名称", "model_version": "模型版本", "model_path":"模型地址", "model_name_url": "模型名称", "model_version_url": "模型版本", "model_path_url": "模型地址", "embedding_file_path":"embedding文件地址", "is_fallback":"兜底版本", "check_service":"检查服务", "status": "状态", "model_status":"状态", "status_url":"状态", "final_status":"最终状态", "pipeline": "任务流", "pipeline_url": "任务流", "etl_pipeline_url": "任务流", "service_pipeline_url": "推理编排", "run_id": "kfp运行id", "run_time": "kfp运行时间", "type": "类型", "reset": "重置", "user": "用户", "role": "角色", "dag_json": "流向图", "dag_json_html": "流向图", "username": "用户", "schedule_type": "调度类型", "cron_time": "调度周期", "global_args": "全局参数", "global_env": "全局环境变量", "parallelism": "任务并行数", "run_pipeline": "运行", "status_more": "状态详情", "status_more_html": "状态详情", "execution_date":"执行时间", "base_image":"基础镜像", "tag":"tag", "save":"保存", "history":"历史", "consecutive_build":"连续构建", "need_gpu":"需要gpu", "last_image":"最新镜像", "expired_limit":"有效实例数目", "canary":"分流", "shadow": "流量复制", "log": "日志", "pod": "容器", "ide_type": "IDE类型", "annotations": "注释", "annotations_html": "注释", "spec": "属性", "spec_html": "属性", "info_json": "通知", "info_json_html": "通知", "labels": "标签", "label": "标签", "labels_html": "标签", "label_url": "标签", "add_row_time": "添加时间", "experiment_id": "kfp实验id", "pipeline_file": "workflow yaml", "pipeline_file_html": "workflow yaml", "pipeline_argo_id": "kfp任务流id", "version_id": "kfp版本id", "job_template": "任务模板", "job_template_url": "任务模板", "template": "功能模板", "alert_status": "监控状态", "task_args":"任务参数", "url_html": "网址", "url": "网址", "public":"公开", "alert_user": "通知用户", "experiment": "Experiment yaml", "experiment_html": "Experiment yaml", "train_model": "训练模型", "ip": "ip", "deploy_time": "部署时间", "host": "域名", "host_url": "域名", "inference_host_url":"域名：测试(test.xx)/调试(debug.xx)", "deploy": "部署", "test_deploy": "测试部署", "prod_deploy": "生产部署", "check_test_service": '检测测试服务', "min_replicas": "最低副本数", "max_replicas": "最高副本数", "replicas": "副本数", "ports": "端口", "roll": "滚动发布", "k8s_yaml": "yaml", "service": "服务", "download_url": "下载地址", "metrics": "指标", "metrics_html":"指标", "operate_html":"操作", "md5": "md5", "service_type": "服务类型", "job_args_definition":"模板参数定义示例", "job_describe":"模板描述", "job_args_demo": "模板参数示例", "stop":"停止", "parallel_trial_count": "并行搜索次数", "max_trial_count": "最多搜索次数", "max_failed_trial_count": "最多失败搜索次数", "objective_type": "目标函数类型", "objective_goal": "目标值", "objective_metric_name": "目标度量", "objective_additional_metric_names": "附加目标度量", "algorithm_name": "搜索算法", "algorithm_setting": "搜索算法配置", "parameters": "超参数配置", "parameters_demo": "超参配置示例", "parameters_html": "超参数配置", "job_json": "搜索任务配置", "trial_spec": "任务 yaml", "trial_spec_html": "任务 yaml", "create_experiment": "启动调度", "run_instance":"运行实例", "monitoring": "监控", "monitoring_html":"监控", "monitoring_url":"监控", "link": "链接", "clear": "清理", "expand":"扩展", "expand_html":"扩展", "parameter": "扩展参数", "parameter_html": "扩展参数", "renew":"续期", "api_type":"接口类型", "code_dir":"代码目录", "id_url":"id", "debug": "调试", "run": "运行", "run_url":"运行", "depends_on_past":"过往依赖", "max_active_runs": "最大激活运行数", "des_image":"目标镜像", "target_image":"目标镜像", "image_history":"镜像历史", "elapsed_time":"耗时", "task_status":"任务状态", "hpa":"弹性伸缩容", "health":"健康检查", "transformer":"处理变换", "deploy_history":"部署记录", "model_input": "模型输入", "input_html":"输入", "model_output": "模型输出", "output_html": "输出", "config":"配置", "config_html": "配置", "app":"产品", "field": "领域", "cluster": "集群", "db_name":"数据库", "database":"数据库", "db": "数据库", "metadata_db": "数据库", "security_level":"安全等级", "value_score":"价值评分", "storage_size": "存储大小", "warehouse_level": "数仓类型", "crontab": "调度周期", "cost":"数据成本", "column_type": "列类型", "remark":"备注", "ttl": "保留时长", "sql": "sql", "sql_html": "sql", "primary_part_col_name":"主分区列", "app_group":"应用组", "sql_demo":"sql示例", "create_table_ddl":"建表sql", "insert_sql": "数据导入sql", "metric_type": "指标类型", "metric_data_type":"指标数据类型", "metric_dim": "指标维度", "metric_level": "指标级别", "metric_responsible":"指标责任人", "caliber":"口径", "is_partition":"是否分区列", "partition_type":"分区类型", "c_org_fullname":"组织架构", "lifecycle":"当前生命周期(天)", "rec_lifecycle":"推荐生命周期(天)", "storage_cost":"存储成本(元/月)", "visits_seven":"7日访问次数", "recent_visit":"最近访问日期", "partition_start":"分区开始时间", "partition_end":"分区结束时间", "visits_thirty":"30天访问次数", "product_name":"产品名", "product_id":"产品id", "product_desc":"产品描述", "active":"激活", "job_worker_image":"工作镜像", "job_worker_command": "启动命令", "source_type":"数据源类型", "priority":"优先级", "owner": "责任人", "industry":"行业", "skip":"跳过", "etl_pipeline": "任务流", "etl_pipeline_id": "任务流id", "etl_task": "任务", "etl_task_id": "任务id", "access":"接入", "inference_config": "推理配置", "file_type":"文件类型", "responsible":"责任人", "cycle_unit":"周期单位", "task_type":"任务类型", "help":"帮助", "creator": "创建者", "created_by": "创建者", "changed_by": "修改者", "created_on": "创建时间", "create_time": "创建时间", "update_time":"更新时间", "changed_on": "修改时间", "change_time":"更新时间", "modified": "修改时间", "cronjob_start_time":"补录起点" } # 获取列的中文显示 # @staticmethod def lab(col,label_columns=label_columns): if col in label_columns: return _(label_columns[col]) return _(re.sub("[._]", " ", col).title()) def get_default_node_selector(self,node_selector,resource_gpu,model_type): # prefer already defined selectors if not node_selector: node_selector='' # completely determined by the platform if core.get_gpu(resource_gpu)[0]: node_selector = node_selector.replace('cpu=true', 'gpu=true') + ",gpu=true,%s=true"%model_type else: node_selector = node_selector.replace('gpu=true', 'cpu=true') + ",cpu=true,%s=true"%model_type if 'org' not in node_selector: node_selector += ',org=public' node_selector = re.split(',|;|\n|\t', str(node_selector)) node_selector = [selector.strip() for selector in node_selector if selector.strip()] node_selector = ','.join(list(set(node_selector))) return node_selector