data={}
data['intents']=[]
data['intents'].append({
'tag': tag,
'patterns': patterns,
'response': response
})
with open('training.json', 'a') as training:
json.dump(data, training)
我正在尝试将值附加到关键意图上。但是当我尝试附加值时,我得到的输出如下:
{"intents": [{"response": "customize", "patterns": "erp", "tag": "purchase"}]}{"intents": [{"response": "kjj", "tag": "sales", "patterns": "jjkj"}]}
我希望我的输出格式如下:
{"intents":[
{"tag":"sale",
"patterns":["ptr1","ptr2"],
"responses":["resp1","resp2"]
},
{"tag":"purchase",
"patterns":["abc","def"],
"responses":["xyz","zzz"]
}
]
}
答案 0 :(得分:1)
您无法添加新数据,否则它将破坏json,因此必须替换数据。而且不要专注于缩进,这是没有必要的 试试这个:
import json
data={}
data['intents']=[]
data['intents'].append({
'tag': 'tag',
'patterns': 'patterns',
'response': 'response'
})
try:
with open('training.json', 'r') as training:
old_data = training.readlines()
if old_data:
old_data = json.loads(old_data[0])
for intents in data['intents']:
old_data['intents'].append(intents)
data = old_data
old_data = None
with open('training.json', 'w') as training:
json.dump(data, training)
except:
with open('training.json', 'w') as training:
json.dump(data, training)
答案 1 :(得分:0)
由于要将项目追加到对象的列表中(而不是将文本追加到json
文件本身),需要首先读取json
,然后追加并编写:< / p>
import os
data = {'intents': []}
if os.path.exists('training.json'):
with open('training.json', 'r') as f:
data = json.load(f)
data['intents'].append({
'tag': tag,
'patterns': patterns,
'response': response
})
with open('training.json', 'w') as training:
json.dump(data, training)
答案 2 :(得分:0)
您可以添加它而无需轻松复制旧JSOF文件中的内容。
import json
data={}
data['intents']=[]
data['intents'].append({
'tag': 'tag',
'patterns': 'patterns',
'responses': 'responses'
})
tag = request.json['tag']
new_response = request.json['response']
for intent in intents['intents']:
if intent['tag'] in tag:
if new_response in intent['responses']:
response = {"success":False,"message":"Response already found in the training data"}
return response
else:
print("updating file")
intent['responses'].append(new_response)
data = intents
with open('intents.json', 'w') as training:
json.dump(data, training)
这会将新的响应附加在同一标签下。另外,如果其中存在相同的响应,则它将跳过。