python - 如何使用 python 查询本地镜像的 Google Cloud Vision API？

Question

我想使用本地存储的图像向 Google 视觉 API 发送请求。我遵循“Python Client for Google Cloud Vision¶ ”指南，特别是“注释图像”下的代码示例。为了构建我的请求，我遵循“Base64 Encoding”中的示例。

这是我编写的代码:

from google.cloud.vision import ImageAnnotatorClient
import os
import base64
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = 'keys/sc_google.json'

def encode_image(image_path):
    with open(image_path, "rb") as f:
        image_content = f.read()
        return base64.b64encode(image_content)

client = ImageAnnotatorClient()
image_path = 'images/happy_baby3.png'
query = {
    "requests": [
        {
            "image": {
                "content": encode_image(image_path)
            },
            "features": [
                {
                    "type": "FACE_DETECTION",
                    "maxResults": 10
                }
            ]
        }
    ]
}
response = client.annotate_image(query)
print(response.annotations)

当我运行代码时，我收到以下错误消息:

Traceback (most recent call last):
  File "/mnt/c/Users/nomis/PycharmProjects/hello_web_app/emotion_detection.py", line 29, in <module>
    response = client.annotate_image(query)
  File "/usr/local/lib/python3.6/dist-packages/google/cloud/vision_helpers/__init__.py", line 55, in annotate_image
    image = protobuf.get(request, "image")
  File "/usr/local/lib/python3.6/dist-packages/google/api_core/protobuf_helpers.py", line 192, in get
    raise KeyError(key)
KeyError: 'image'

我也尝试过this Example但出现错误“在‘types.py’中找不到引用‘图像’。

您能否解释一下我做错了什么，并提供一个关于如何使用本地镜像查询 Google Vision API 的最小工作示例？

Answer 1

在您链接的示例中，没有请求数组。试试这个:

query = {
            "image": {
                "content": encode_image(image_path)
            },
            "features": [
                {
                    "type": "FACE_DETECTION",
                    "maxResults": 10
                }
            ]
}