在CustomVision.ai API中过帐数据时出错

Question

我在customvision.ai上对模型进行了训练，得到了预测关键字和端点。

下面是我写的内容

import requests
import json
url="https://centralindia.api.cognitive.microsoft.com/customvision/v3.0/Prediction/xxxxxxxxxx/classify/iterations/Iteration1/url"
headers={'content-type':'application/json','Prediction-Key':'24xxxxxxxxxxxxxxxxxxxxxx'}
body={"Url": "https://i.imgur.com/cYzaOkV.jpg"}
r =requests.post(url,data=body,headers=headers)
print(r.content)

但是当我尝试运行这段代码时，我得到了这个错误：b'{"code":"BadRequestImageUrl","message":"Invalid image url"}'

但是，我尝试使用本地图像调用我的模型，在这种情况下，它在为本地img编写的代码中工作得很好：

import requests
import json
url="https://centralindia.api.cognitive.microsoft.com/customvision/v3.0/Prediction/82xxxxxxxxxxxxxxxxxxxx/classify/iterations/Iteration1/image"
headers={'content-type':'application/octet-stream','Prediction-Key':'24xxxxxxxxxxxxxxxxxx'}
r =requests.post(url,data=open("lentigo-adults-1.jpg","rb"),headers=headers)

print(r.content)

我得到了我想要的输出：

b'{"id":"43a1d33b-0ea9-490a-9ae4-d24f8395931a","project":"82462bcc-5616-4f82-98ea-8d4fda55a9e6","iteration":"d0e48a13-e906-4af2-9f4c-15d79e3b4576","created":"2020-10-14T13:48:41.671Z","predictions":[{"probability":0.999876738,"tagId":"b5bd1d3a-dbcd-4b0d-ad0b-3f8d7b12fbea","tagName":"Lentigo"},{"probability":0.0001100349,"tagId":"c69433ef-4630-4e0f-8077-2de485483323","tagName":"Acne Cystic"},{"probability":1.29300688E-05,"tagId":"2102157b-b250-4c3d-bf2a-588cda8e93eb","tagName":"Rosacea"},{"probability":1.10303489E-07,"tagId":"ea68f8e7-b50f-4ca0-88dc-c2c8f3bb6823","tagName":"Herpes"},{"probability":9.26677259E-08,"tagId":"29c5c735-90f7-48ff-b23c-545be3325e9e","tagName":"Acne Pustular"},{"probability":7.811148E-08,"tagId":"579b080c-4ab1-405f-a8bd-5bf32efd9b70","tagName":"Melasma"},{"probability":3.90258634E-08,"tagId":"8340afdd-b758-46b4-9aee-754480437f0e","tagName":"Acne Comedo"},{"probability":1.49268364E-09,"tagId":"b89f964a-1a72-41ec-b8b7-8e0d7e9ef8ff","tagName":"Eczema"}]}'

有人能帮帮我吗？？我想使用img url调用此终结点

感谢您给我宝贵的时间来解决我的问题。

Answer 1

根据我的测试，当我们调用图像URL端点时，主体应该是json string。请将代码r =requests.post(url,data=body,headers=headers)更新为response = requests.post( url, headers=headers, data=json.dumps(body))

例如

import requests
import json
url = "https://testvision03.cognitiveservices.azure.com/customvision/v3.0/Prediction/<>/detect/iterations/Iteration1/url"

body = {"Url": "https://i.imgur.com/cYzaOkV.jpg"}
headers = {
    'Prediction-Key': '',
    'Content-Type': 'application/json'
}

response = requests.post(url, headers=headers, data=json.dumps(body))

print(response.text.encode('utf8'))

​

​