i使用DialogFlow v2(并尝试使用v2Beta1),使用python中的SDK,但问题是detectIntent方法无法识别查询输入中的参数。
因此,在寻找解决方案时,我使用简单的cUrl调用重新创建了问题。
接下来是创建EntityType的cUrl:
curl -X POST -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
"https://dialogflow.googleapis.com/v2beta1/projects/myproject_id/agent/entityTypes?languageCode=es" \
--data "{
'displayName': 'customer',
'kind': 'KIND_MAP',
'autoExpansionMode': 'AUTO_EXPANSION_MODE_DEFAULT',
'entities': [
{
'value': 'one',
'synonyms': [
'one', 'uno'
]
},
{
'value': 'two',
'synonyms': [
'two', 'dos'
]
}
],
'enableFuzzyExtraction': true
}"
那很好,并且创建了EntityType。
现在,创建意图:
curl -X POST -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
"https://dialogflow.googleapis.com/v2beta1/projects/myproject_id/agent/intents" \
--data "{
'displayName': 'ExampleIntent',
'priority': 500000,
'mlDisabled': false,
'trainingPhrases': [
{
'type': 'EXAMPLE',
'parts': [
{
'text': 'start '
},
{
'text': 'one',
'alias': 'customer',
'entityType': '@customer',
'userDefined': true
}
]
},
{
'type': 'EXAMPLE',
'parts': [
{
'text': 'begin '
},
{
'text': 'one',
'alias': 'customer',
'entityType': '@customer',
'userDefined': true
}
]
},
{
'type': 'EXAMPLE',
'parts': [
{
'text': 'do '
},
{
'text': 'one',
'alias': 'customer',
'entityType': '@customer',
'userDefined': true
}
]
}
],
'action': 'start',
'messages': [
{
'text': {
'text': [
'Hi'
]
}
}
],
}"
也可以。
现在,我尝试检测意图:
curl -X POST -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
"https://dialogflow.googleapis.com/v2beta1/projects/myproject_id/agent/sessions/0:detectIntent" \
--data "{
'queryInput': {
'text': {
'text': 'do one',
'languageCode': 'es'
}
}
}"
响应不带参数:
{
"responseId": "de68c5f5-6aa9-4716-ac22-626a22fc5d43-b81332aa",
"queryResult": {
"queryText": "do one",
"action": "start",
"parameters": {},
"allRequiredParamsPresent": true,
"fulfillmentText": "Hi",
"fulfillmentMessages": [
{
"text": {
"text": [
"Hi"
]
}
}
],
"intent": {
"name": "projects/myproject_id/agent/intents/20ab36d6-e8c8-40d6-87dc-78a61e2de600",
"displayName": "ExampleIntent"
},
"intentDetectionConfidence": 1,
"languageCode": "es"
}
}
我试图训练:
curl -X POST -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
"https://dialogflow.googleapis.com/v2beta1/projects/myproject_id/agent:train"
结果是一样的。
但是,如果我转到Web控制台,请输入意图并单击“保存”按钮... detectIntent开始起作用。
{
"responseId": "45c919dc-677d-4ae4-8572-588955cd5414-b81332aa",
"queryResult": {
"queryText": "do one",
"action": "start",
"parameters": {
"customer": [
"one"
]
},
"allRequiredParamsPresent": true,
"fulfillmentText": "Hi",
"fulfillmentMessages": [
{
"text": {
"text": [
"Hi"
]
}
}
],
"intent": {
"name": "projects/myproject_id/agent/intents/20ab36d6-e8c8-40d6-87dc-78a61e2de600",
"displayName": "ExampleIntent"
},
"intentDetectionConfidence": 1,
"languageCode": "es"
}
}
我想念什么?
感谢任何线索使这项工作有效。
答案 0 :(得分:0)
嗯...经过一些研究,我发现了问题,我在createIntent中缺少了 parameters 参数:
curl -X POST -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
"https://dialogflow.googleapis.com/v2beta1/projects/myproject_id/agent/intents" \
--data "{
'displayName': 'ExampleIntent',
'priority': 500000,
'mlDisabled': false,
'parameters': [
{
'displayName': 'customer',
'entityTypeDisplayName': '@customer',
'value': '$customer'
}
],
'trainingPhrases': [
{