类 Service
服务
azureml.deploy.server.service.Service(service, http_client)
基于服务元数据属性的服务使用和批处理的动态对象。
批处理
batch(records, parallel_count=10)
为此服务上的批处理执行注册一组输入记录。
参数
records
要执行的输入记录的 data.frame 或列表。
parallel_count
用于处理批次中的项的线程数。 默认值为 10。 请确保不要使用过高的数字,因为这可能会对性能产生负面影响。
返回
用于控制此服务的批处理生命周期的 Batch
实例。
capabilities
capabilities()
提供以下信息,描述此服务的所有内容:
api - API REST 终结点。
name - 服务名称。
version - 服务版本。
published_by - 服务发布作者。
runtime - 服务运行时上下文 R|Python。
description - 服务说明。
creation_time - 服务发布时间戳。
snapshot_id - 此服务绑定的快照标识符。
inputs - 输入架构名称/类型定义。
outputs - 输出架构名称/类型定义。
inputs_encoded - 已编码为 Python 的输入架构名称/类型。
outputs_encoded - 已编码为 Python 的输出架构名称/类型。
artifacts - 支持的生成文件。
operation_id - 函数
alias
。swagger - 此服务的 swagger.json 文档的 API REST 终结点。
返回
描述服务的键/值的 dict
。
get_batch
get_batch(execution_id)
基于执行标识符检索服务批次。
参数
execution_id
批处理执行的标识符。
返回
用于控制此服务的批处理生命周期的 Batch
实例。
list_executions
list_batch_executions()
获取此服务的当前排队的所有批处理执行标识符。
返回
执行标识符的 list
。
swagger
swagger()
检索此服务的 swagger.json(参阅 http://swagger.io/)。
返回
此服务的 swagger 文档,作为 json str
。
反馈
https://aka.ms/ContentUserFeedback。
即将发布:在整个 2024 年,我们将逐步淘汰作为内容反馈机制的“GitHub 问题”,并将其取代为新的反馈系统。 有关详细信息,请参阅:提交和查看相关反馈