How to Integrate IBM Watson with Airtable

Prerequisites

• Create an IBM Cloud account and understand the basics of IBM Watson services.

 

• Create an Airtable account and become familiar with the Airtable API.

 

• Install Node.js and npm on your machine.

 

• Familiarize yourself with basic JavaScript programming.

Set Up IBM Watson

• Log in to your IBM Cloud account.

 

• Navigate to the IBM Watson section and create a new Watson service instance, such as Watson Assistant or Watson Language Translator.

 

• Retrieve your service credentials: API Key and URL, from the IBM Cloud dashboard for your specific Watson service.

Set Up Airtable

• Log in to your Airtable account.

 

• Navigate to your Airtable base and create a new base or select an existing one to integrate with IBM Watson.

 

• Access the Airtable API section and retrieve your API Key.

Install Required Packages

• Initialize a new Node.js project by executing npm init in your terminal.

 

• Install the required packages:

npm install axios airtable ibm-watson

Write the Integration Script

• Create a new JavaScript file, e.g., integrate.js.

 

• Load the necessary modules and configuration:

const Airtable = require('airtable');
const axios = require('axios');
const { IamAuthenticator } = require('ibm-watson/auth');
const { AssistantV2 } = require('ibm-watson/assistant/v2');

const airtableApiKey = 'YOUR_AIRTABLE_API_KEY';
const baseId = 'YOUR_BASE_ID';
const tableName = 'YOUR_TABLE_NAME';

const ibmApiKey = 'YOUR_IBM_API_KEY';
const serviceUrl = 'YOUR_IBM_SERVICE_URL';

const assistant = new AssistantV2({
  version: '2021-06-14',
  authenticator: new IamAuthenticator({
    apikey: ibmApiKey,
  }),
  serviceUrl: serviceUrl,
});

• Connect to Airtable and retrieve data:

const base = new Airtable({ apiKey: airtableApiKey }).base(baseId);

base(tableName).select({
  maxRecords: 5,
  view: "Grid view"
}).eachPage((records, fetchNextPage) => {
  records.forEach(record => {
    console.log('Retrieved record:', record.fields);
  });

  fetchNextPage();
}, err => {
  if (err) { console.error(err); return; }
});

• Implement IBM Watson API call:

const sessionId;

// First, create a session
assistant.createSession({
  assistantId: 'YOUR_ASSISTANT_ID'
})
.then(res => {
  sessionId = res.result.session_id;
  return assistant.message({
    assistantId: 'YOUR_ASSISTANT_ID',
    sessionId: sessionId,
    input: {
      'message_type': 'text',
      'text': 'Hello'
    }
  });
})
.then(res => {
  console.log(JSON.stringify(res.result, null, 2));
})
.catch(err => {
  console.error(err);
});

Integrate IBM Watson and Airtable

• Combine the data retrieval from Airtable with the Watson API call to process the data:

base(tableName).select({
  maxRecords: 5,
  view: "Grid view"
}).eachPage(async (records, fetchNextPage) => {
  for (const record of records) {
    const text = record.fields['TextColumn']; // Replace with your column name
    console.log('Processing record:', text);
    
    try {
      const messageResponse = await assistant.message({
        assistantId: 'YOUR_ASSISTANT_ID',
        sessionId: sessionId,
        input: {
          'message_type': 'text',
          'text': text
        }
      });
      console.log('Watson Response:', JSON.stringify(messageResponse.result, null, 2));
    } catch (err) {
      console.error(err);
    }
  }

  fetchNextPage();
}, err => {
  if (err) { console.error(err); return; }
});

Test and Deploy

• Run your script using node integrate.js to test the integration.

 

• Check both Airtable and the Watson service for expected data exchanges and outputs.

 

• Deploy your script on a server or automation platform if necessary for continuous operation.