Search code examples
dialogflow-esdialogflow-cxdialogflow-cx-dashboard

Dialogflow CX import entities as CSV

Task:

I want to create a custom entity type to hold a list of 100 product names.

Solution in Dialogflow ES (v2)

In Dialogflow ES, it was easy upload entity or switch to raw mode while creating entity as you can see in images below:

enter image description here enter image description here

Problem in Dialogflow CX (v3)

In Dialogflow CX, I don’t see any of such options (as you can see in images below), and having to do this manually for every custom entity is not only difficult but also time consuming specially when you have it in hundreds.

enter image description here enter image description here

What I am looking for in answer:

I know there is way to upload entity using dialogflow v3 rest api, but I am looking for a way to upload entities from something that is UI based, I also understand that I can make my own UI to upload entities using dialogflow v3 rest apis but that would be re invention of wheels I would say, so now if there is any option that is hidden somewhere to in cx dashboard please let me know, or if someone have contacts in dialogflow dev team please let them know we are having hard time with this

Solution

  • Check the Official Dialogflow CX REST API documentation for this. https://cloud.google.com/dialogflow/cx/docs/reference/rest/v3/projects.locations.agents.entityTypes/create

    Below is a sample code that reads the column City in the sample.csv using pandas and creates an entity. Python

    import requests
import pandas as pd
from google.oauth2 import service_account
from google.cloud.dialogflowcx_v3beta1.types import entity_type
from google.auth.transport.requests import AuthorizedSession

# sample csv data
# Name,City
# John,New York City
# Alice,Philadelphia

data = pd.read_csv("sample.csv") 
cities = data['City'].tolist()
entity_json = []
for each in cities:
    each_entity_value = {}
    each_entity_value['value'] = each
    each_entity_value['synonyms'] = [each]
    entity_json.append(each_entity_value)
print(entity_json)
print('***************************************')

# download the service account json with the required permissions to call the cx agent
credentials = service_account.Credentials.from_service_account_file(
    'credentials.json')
scoped_credentials = credentials.with_scopes(
    ['https://www.googleapis.com/auth/cloud-platform'])
authed_session = AuthorizedSession(scoped_credentials)
kind = entity_type.EntityType.Kind.KIND_MAP

# configure these variables before running the script
project_id = #YOUR-PROJECT-ID
agent_id = #YOUR-CX-AGENT-ID
location = #AGENT-LOCATION-ID
response = authed_session.post('https://dialogflow.googleapis.com/v3/projects/'+ project_id + '/locations/' + location + '/agents/' + agent_id + '/entityTypes',
                                json={
                                        "kind": kind,
                                        "displayName": "city-names",
                                        "entities": entity_json
                                    }
                                )                        

response_txt = response.text
print(response_txt)

    Here are the requirements.

    # requirements.txt
# google-cloud-dialogflow-cx==0.5.0
# pandas==1.1.5
# requests==2.21.0
# requests-oauthlib==1.3.0

    You can always try it on the API explorer first by providing the required parameters.

    parent: projects/<project-id>/locations/<location>/agents/<agent-id>
languageCode: en
Request body: {
  "displayName": "test",
  "entities": [
    {
      "value": "string",
      "synonyms": [
        "string",
        "str"
      ]
    }
  ],
  "kind": "KIND_MAP"
}

    Enable the credentials and execute it. You will get a 200 success response and a new entity will be created in the CX agent console.