GoogleAPICallError：无意外状态：长时间运行的操作既没有响应也没有错误集

【问题标题】：GoogleAPICallError: None Unexpected state: Long-running operation had neither response nor error setGoogleAPICallError：无意外状态：长时间运行的操作既没有响应也没有错误集
【发布时间】：2021-12-11 16:57:30
【问题描述】：

我是 Google Cloud Platform 的新手，我正在尝试创建一个 Feature Store 来填充来自 Google Cloud Storage 的 csv 文件中的值。目的是从 Python 的本地笔记本中做到这一点。 我基本上遵循代码here，进行适当的更改，因为我正在使用信用卡公共数据集。 运行代码时出现的错误如下：

GoogleAPICallError: None Unexpected state: Long-running operation had neither response nor error set.

它发生在从 csv 文件中提取数据的过程中。

这是我正在处理的代码：

import os
from datetime import datetime
from google.cloud import bigquery
from google.cloud import aiplatform
from google.cloud.aiplatform_v1.types import feature as feature_pb2
from google.cloud.aiplatform_v1.types import featurestore as featurestore_pb2
from google.cloud.aiplatform_v1.types import \
    featurestore_service as featurestore_service_pb2
from google.cloud.aiplatform_v1.types import entity_type as entity_type_pb2
from google.cloud.aiplatform_v1.types import FeatureSelector, IdMatcher

credential_path = r"C:\Users\...\.json"
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = credential_path

## Constants
PROJECT_ID = "my-project-ID"
REGION = "us-central1"
API_ENDPOINT = "us-central1-aiplatform.googleapis.com"
INPUT_CSV_FILE = "my-input-file.csv"
FEATURESTORE_ID = "fraud_detection"

## Output dataset
DESTINATION_DATA_SET = "fraud_predictions"
TIMESTAMP = datetime.now().strftime("%Y%m%d%H%M%S")
DESTINATION_DATA_SET = "{prefix}_{timestamp}".format(
    prefix=DESTINATION_DATA_SET, timestamp=TIMESTAMP
)

## Output table. Make sure that the table does NOT already exist; 
## the BatchReadFeatureValues API cannot overwrite an existing table
DESTINATION_TABLE_NAME = "training_data"

DESTINATION_PATTERN = "bq://{project}.{dataset}.{table}"
DESTINATION_TABLE_URI = DESTINATION_PATTERN.format(
    project=PROJECT_ID, dataset=DESTINATION_DATA_SET, 
    table=DESTINATION_TABLE_NAME
)

## Create dataset
client = bigquery.Client(project=PROJECT_ID)
dataset_id = "{}.{}".format(client.project, DESTINATION_DATA_SET)
dataset = bigquery.Dataset(dataset_id)
dataset.location = REGION
dataset = client.create_dataset(dataset)
print("Created dataset {}.{}".format(client.project, dataset.dataset_id))

## Create client for CRUD and data_client for reading feature values.
client = aiplatform.gapic.FeaturestoreServiceClient(
    client_options={"api_endpoint": API_ENDPOINT})
data_client = aiplatform.gapic.FeaturestoreOnlineServingServiceClient(
    client_options={"api_endpoint": API_ENDPOINT})
BASE_RESOURCE_PATH = client.common_location_path(PROJECT_ID, REGION)

## Create featurestore (only the first time)
create_lro = client.create_featurestore(
    featurestore_service_pb2.CreateFeaturestoreRequest(
        parent=BASE_RESOURCE_PATH,
        featurestore_id=FEATURESTORE_ID,
        featurestore=featurestore_pb2.Featurestore(
            online_serving_config=featurestore_pb2.Featurestore.OnlineServingConfig(
                fixed_node_count=1
            ),
        ),
    )
)

## Wait for LRO to finish and get the LRO result.
print(create_lro.result())

client.get_featurestore(
    name=client.featurestore_path(PROJECT_ID, REGION, FEATURESTORE_ID)
)

## Create credit card entity type (only the first time)
cc_entity_type_lro = client.create_entity_type(
    featurestore_service_pb2.CreateEntityTypeRequest(
        parent=client.featurestore_path(PROJECT_ID, REGION, FEATURESTORE_ID),
        entity_type_id="creditcards",
        entity_type=entity_type_pb2.EntityType(
            description="Credit card entity",
        ),
    )
)

## Create fraud entity type (only the first time)
fraud_entity_type_lro = client.create_entity_type(
    featurestore_service_pb2.CreateEntityTypeRequest(
        parent=client.featurestore_path(PROJECT_ID, REGION, FEATURESTORE_ID),
        entity_type_id="frauds",
        entity_type=entity_type_pb2.EntityType(
            description="Fraud entity",
        ),
    )
)

## Create features for credit card type (only the first time)
client.batch_create_features(
    parent=client.entity_type_path(PROJECT_ID, REGION, FEATURESTORE_ID, "creditcards"),
    requests=[
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v1",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v2",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v3",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v4",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v5",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v6",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v7",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v8",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v9",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v10",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v11",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v12",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v13",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v14",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v15",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v16",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v17",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v18",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v19",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v20",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v21",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v22",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v23",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v24",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v25",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v26",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v27",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="v28",
        ),
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="amount",
        ),
    ],
).result()

## Create features for fraud type (only the first time)
client.batch_create_features(
    parent=client.entity_type_path(PROJECT_ID, REGION, FEATURESTORE_ID, "frauds"),
    requests=[
        featurestore_service_pb2.CreateFeatureRequest(
            feature=feature_pb2.Feature(
                value_type=feature_pb2.Feature.ValueType.DOUBLE, description="",
            ),
            feature_id="class",
        ),
    ],
).result()

## Import features values for credit cards
import_cc_request = aiplatform.gapic.ImportFeatureValuesRequest(
    entity_type=client.entity_type_path(
        PROJECT_ID, REGION, FEATURESTORE_ID, "creditcards"),
    csv_source=aiplatform.gapic.CsvSource(gcs_source=aiplatform.gapic.GcsSource(
        uris=["gs://fraud-detection-19102021/dataset/cc_details_train.csv"])),
    entity_id_field="cc_id",
    feature_specs=[
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v1"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v2"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v3"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v4"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v5"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v6"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v7"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v8"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v9"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v10"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v11"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v12"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v13"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v14"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v15"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v16"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v17"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v18"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v19"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v20"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v21"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v22"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v23"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v24"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v25"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v26"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v27"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="v28"),
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="amount"),
    ],
    feature_time_field='time',
    worker_count=1,
)

## Start to import
ingestion_lro = client.import_feature_values(import_cc_request)

## Polls for the LRO status and prints when the LRO has completed
ingestion_lro.result()

## Import features values for frauds
import_fraud_request = aiplatform.gapic.ImportFeatureValuesRequest(
    entity_type=client.entity_type_path(
        PROJECT_ID, REGION, FEATURESTORE_ID, "frauds"),
    csv_source=aiplatform.gapic.CsvSource(gcs_source=aiplatform.gapic.GcsSource(
        uris=["gs://fraud-detection-19102021/dataset/data_fraud_train.csv"])),
    entity_id_field="fraud_id",
    feature_specs=[
        aiplatform.gapic.ImportFeatureValuesRequest.FeatureSpec(id="class"),
    ],
    feature_time_field='time',
    worker_count=1,
)

## Start to import
ingestion_lro = client.import_feature_values(import_fraud_request)

## Polls for the LRO status and prints when the LRO has completed
ingestion_lro.result()

当我从 Google Cloud Console 的 Feature 部分检查 Ingestion Jobs 时，我看到工作已完成，但没有向我的功能添加任何值。

任何建议都非常宝贵。

谢谢大家。

编辑 1 在下图中，有一个我用作输入的 csv 文件的第一行示例 (cc_details_train.csv)。所有看不见的特征都是相似的，特征class 可以假设为 0 或 1 值。 注入作业持续大约 5 分钟以导入（理想情况下）3000 行，但它没有错误地结束，也没有导入任何值。

【问题讨论】：

  • 我有一些问题，你所有的字段都是数字吗？您将所有设置为双值类型，您可以确认吗？您的处理文件的大小是多少？ （示例数据使用电影数据）您可能需要检查处理时间表，here。另外，您能否在代码中准确指出错误点的位置？
  • @Betjens 感谢您的评论。 1）是的，我要摄取的所有特征值都是双精度类型。 2）我的文件大小是 3000 行。从 GCP 的“功能”部分，我看到摄取作业已完成，但功能中添加了 0 个值。 3) 当我执行 ingestion_lro.result() 时出现错误。
  • 所以我挖掘了一下，发现 cc_details_train.csv 是你的实际输入文件......这需要澄清。如果您还可以提供其他信息，这将有所帮助。即：输入文件的示例行。我认为您应该使用更少的功能或尽可能增加您的工人数量。对于我的测试运行，我只使用一行，第一次需要 10-15 分钟。我无法想象 3000 行需要多少时间。
  • @Betjens 在我完成的测试中，注入工作持续大约 5 分钟以导入（理想情况下）3000 行，但它没有错误地结束并且没有导入任何值。我添加了 csv 文件第一行的图像。
  • 你的 entity_id 是 cc_id???我看到你的文件还包含一个名为类的列，它包含什么？？？。我也复制了 3000 行并添加了 0 个值。根据official documentation，每次运行都会有延迟。另外，您是否尝试过使用顶点用户界面？ Vertex-AI > 特征

标签： google-cloud-platform google-api-python-client google-cloud-ml google-cloud-vertex-ai


【解决方案1】：

使用 CSV 到 ImportValues / 使用 ImportFeatureValuesRequest 时的 VERTEX AI 建议

使用此功能时，您可能根本无法导入任何数据。您必须注意您使用的时间字段，因为它必须符合谷歌时间格式。

  1. feature_time_field，必须遵循google设置的时间约束规则，即RFC3339，即：'2021-04-15T08:28:14Z'。您可以查看here字段的详细信息，有关时间戳格式的详细信息可以找到here。
  2. 其他列、字段必须匹配设计值。一个例外是字段 entity_id_field，因为它可以是任何值。

注意：我的测试发现，如果我没有将时间字段正确设置为谷歌推荐的日期格式，它将根本不会上传任何特征值。

test.csv

cc_id,time,v1,v2,v3,v4,v5,v6,v7,v8,v9,v10,v11,v12,v13,v14,v15,v16,v17,v18,v19,v20,v21,v22,v23,v24,v25,v26,v27,v28,amount
100,2021-04-15T08:28:14Z,-1.359807,-0.072781,2.534897,1.872351,2.596267,0.465238,0.923123,0.347986,0.987354,1.234657,2.128645,1.958237,0.876123,-1.712984,-0.876436,1.74699,-1.645877,-0.936121,1.456327,0.087623,1.900872,2.876234,1.874123,0.923451,0.123432,0.000012,1.212121,0.010203,1000

输出：

imported_entity_count: 1
imported_feature_value_count: 29

关于优化和使用功能

您可以查看官方文档here 以查看建议处理的最小和最大记录数量。作为一条建议，您应该只使用实际运行的功能和建议的值来运行它。

查看正在运行的摄取作业

如果您使用 VertexUI 或代码来生成摄取的作业。您可以通过进入 UI 到此路径来跟踪它的运行：

VertexAI > Features > View Ingested Jobs

【讨论】：

  • 感谢您的回答！现在我可以在我的特征库中导入特征值了。我有最后一个疑问。即使注入作业说所有值都已导入，如果我在 UI 中打开我的功能之一，在“统计”中标签“计数”仍然等于“-”。这样可以吗？
  • 对于我在 UI 中看到的内容，它看起来像是功能监控预览的一部分。嗯，它是一个持续的实现，我认为你应该启用并等待间隔完成，看看它是否真的有效。根据功能监控中的描述：“提供描述性统计和分布形状。”
  • 您好，Marco，我刚刚证实了我的假设。我能够获得该信息。仅供参考
猜你喜欢
  • 2021-12-23
  • 2023-01-31
  • 1970-01-01
  • 1970-01-01
  • 1970-01-01
  • 1970-01-01
  • 2013-01-21
  • 1970-01-01
  • 1970-01-01
相关资源
最近更新 更多
热门标签
Java Python linux javascript Mysql C# Docker 算法 前端 SpringBoot Redis Vue spring 设计模式 .net core .net kubernetes c++ 数据库 数据结构 大数据 js 机器学习 微服务 Android Go 程序员 面试 JVM ASP.net core 云原生 人工智能 后端 PHP git CSS golang k8s Nginx Django mybatis 深度学习 多线程 React 架构 devops 爬虫 云计算 Spring Boot LeetCode