python - 属性错误: Dict has no object class_name

Question

我正在尝试解析一个网站，对其进行标记并将不同的句子存储在一个数组中，因此这变成了一个字符串数组。我需要访问从 json 返回的类。例如，我必须上课:nhate和hate。如果该类别令人讨厌并且该类别的置信度 > 0.50，则执行某些操作。但是我无法访问这些类(class)。

words = text.split(".")
c=0
for i in words:

  if not words[c]:
      words[c] = "this was empty before." 
  classes = natural_language_classifier.classify('90e7b4x199-nlc-36073',words[c])
  result = json.dumps(classes, indent=2)
  if (classes.class_name == 'hate' and classes.confidence > 0.50):
    print(json.dumps(classes, indent=2)) 
  c=c+1

我得到的错误是:

Traceback (most recent call last):                        
File "parse.py", line 45, in <module>                        
if (classes.class_name == 'hate' and classes.confidence > 0.50) 
AttributeError: 'dict' object has no attribute 'class_name'

编辑:我得到的json是这样的:

{
  "classifier_id": "10D41B-nlc-1",
  "url": "https://gateway.watsonplatform.net/natural-language-classifier    /api/v1/classifiers/10D41B-nlc-1/classify?text=How%20hot%20wil/10D41B-nlc-1",
  "text": "How hot will it be today?",
  "top_class": "nhate",
  "classes": [
    {
      "class_name": "nhate",
      "confidence": 0.9998201258549781
    },
    {
      "class_name": "hate",
      "confidence": 0.00017987414502176904
    }
  ]
}

已编辑

打印(类)给我:

    {u'url': u'https://gateway.watsonplatform.net/natural-language-classifier/api/v1
    /classifiers/90e7b4x199-nlc-36073', 
    u'text': u' A Partnership With Abu Shaklak Printing House', 
    u'classes': 
    [{u'class_name': u'nhate.', u'confidence': 0.9398546
    187612434}, {u'class_name': u'hate.', u'confidence':   0.0449277873541271}, {u'cla
    ss_name': u'Feels good man', u'confidence': 0.015217593884629425}],     u'classifier
    _id': u'90e7b4x199-nlc-36073', u'top_class': u'nhate.'}

Answer 1

你也许可以做这样的事情:

import json

# For demonstration purposes, use the data from the question rather than calling the api:

# data = natural_language_classifier.classify('90e7b4x199-nlc-36073',words[c])

data = json.loads("""{
  "classifier_id": "10D41B-nlc-1",
  "url": "https://gateway.watsonplatform.net/natural-language-classifier    /api/v1/classifiers/10D41B-nlc-1/classify?text=How%20hot%20wil/10D41B-nlc-1",
  "text": "How hot will it be today?",
  "top_class": "nhate",
  "classes": [
    {
      "class_name": "nhate",
      "confidence": 0.9998201258549781
    },
    {
      "class_name": "hate",
      "confidence": 0.00017987414502176904
    }
  ]
}""")

def hate_gt(data, confidence):
  if 'classes' in data:
    for cls in data['classes']:
      if cls['class_name'] == 'hate' and cls['confidence'] > confidence:
          return True
  return False

print(hate_gt(data, 0.00001))  # True
print(hate_gt(data, 0.5))      # False

在这里试试:https://repl.it/Hl9b/1

我不确定这个 API 合约是什么样的，所以我不想做出任何假设“讨厌”将始终是数组中的第二项。

另请注意，我将 classes 变量名称更改为 data，因为 classes['classes'] 令人困惑。