Adding new utterances to specific Intents in Dialogflow Agent

Question

The questions is quite simple, I have developed an agent in Dialogflow-es, it has been trained with N intents and M utterances by intent. Now, I want to automate the training process, or part of it, using a python script or a notebook. I want to add new utterances to the N intents of the agent using the Dialogflow API and training again after that, but the documentation is not clear at all. I have seen the method batchUpdate but it's not clear for me whether this method is intendet to do what I want neither how to implement it.

Here is the documentation --> GCP-intent batchUpdate

Somebody has done something similar? Or is there an API method to do that?

Thanks in advance.

Answer 1

Finally I have figure it out. it is achieved by following two steps:

1. Retrieve the current intents and training phrases
2. Add new intents and training phrases to previous one.

Although there is a GCP library to do this, I wanted to implement the two steps using requests to keep the code as agnostic as possible.

here is the code of step one and two.

Step one:

url = "https://ValidURL/agent/intents"
headers = {
    "Authorization": token,
    "Content-Type": "application/json",
    "x-goog-user-project": "valid-project-name"
}

params = {'intentView': "INTENT_VIEW_FULL"}

while has_next_page:
    if next_page_token:
        params['pageToken'] = next_page_token
        
    response = requests.get(url, headers=headers, params=params)
    response_json = response.json()

    if response.status_code == 200:
        intents = response_json.get('intents', [])
        all_display_names.extend([intent['displayName'] for intent in intents])
        all_responses.extend([intent for intent in intents])
        
        next_page_token = response_json.get('nextPageToken')  
        has_next_page = bool(next_page_token)
    else:
        print("Error", response.status_code)
        break

Step two:

url = "https://validUrL/agents/intents:batchUpdate"
headers = {
    "Authorization": token,
    "x-goog-user-project": "project_name"
}

# Building body part
body = {
    "updateMask":{"paths": ["training_phrases"]},
    "intentView": "INTENT_VIEW_FULL",
    "intentBatchInline": {
        "intents": [
            {
                "name": f"Project_name/agent/intents/{intent_id}",
                "display_name": display_name,
                "trainingPhrases": new_training_phrases,
                "languageCode":"es"
            }
        ]
    }
}

# post the new training phrases
response = requests.post(url, headers=headers, json=body)


if response.status_code == 200:
    print("Success.")
else:
    print("Error:", response.status_code)
    print(response.text)