Create a Custom Parse Connector


🕐 5 min read

The custom parse connector ingests the fetched binary data using parse utils for custom parsing.

Use cases

Instructions

To create your custom parse connector, we recommend that you first create a sample custom parse connector and edit the generated files. It’s easier to get all the files and directory structure your connector needs by just using an existing example.

Notes:
  • Custom parsing currently doesn’t support datasets larger than 100 MB.
  • If your data is complex and needs preprocessing, use the TACO Toolkit library to prepare your data.

To create your custom parse connector, do the following steps.

Step 1: Create a boilerplate custom parse connector

  1. Enter the following command to create the connector:

    taco create my-custom-parse-connector --boilerplate custom-parse

    This creates a directory with the custom parse boilerplate code, which is included with the toolkit.

  2. Change directories to the my-custom-parse-connector directory. 
    cd my-custom-parse-connector

  3. Build the connector by entering the following command:

    taco build

    This command clears any earlier or existing build caches. Then the command installs the dependencies and builds both the front-end code and the back-end code (handlers). Finally, the command copies the connector.json file (the configuration file) to your directory.

Step 2: Configure your connector’s properties

In your new custom parse connector directory, find and open the connector.json file.

{
  "name": "my-custom-parse-connector",
  "version": "1.0.0",
  "tableau-version": {
    "min": "2023.3"
  },
  "vendor": {
    "name": "vendor-name",
    "support-link": "https://vendor-name.com",
    "email": "support@vendor-name.com"
  },
  "permission": {
    "api": {
      "https://*.example.com/": [
        "GET",
        "POST"
      ]
    }
  },
  "auth": {
    "type": "none"
  },
  "window": {
    "height": 800,
    "width": 600
  }
}

Make the following changes:

  1. Change the general properties.

    Name Value
    name Your connector’s name
    version Your connector’s version
    min The earliest Tableau version your connector supports

  2. Change the company properties.

    Name Value
    vendor.name Your company name
    vendor.support-link Your company’s URL
    vendor.email Your company’s email

  3. Change the permissions.

    Name Value
    permission.api The URI for the API that the connector is allowed to access, along with the methods (POST, GET, PUT, PATCH, DELETE) that the connector is allowed to use.

  4. Verify the authentication value

    Name Value
    auth.type Set to none

    For more information about authentication, see the Authentication section in the Considerations for Building Your Connector topic.

  5. Change the HTML pane size.

    Name Value
    window.height The height of the connector HTML pane
    window.width The width of the connector HTML pane

Step 3: Edit the user interface

When you open a web data connector in Tableau, the connector displays an HTML page that links to your code and to your connector’s handlers. Optionally, this page can also display a user interface (UI) for your users to select the data that they want to download.

The /app/components/ConnectorView.tsx file is the main component for the connector UI. You can modify or replace this file based on your connector UI requirements.

Step 4: Edit the connector object

The /app/components/useConnector.ts file contains a custom React hook that abstracts common connector object operations. The hook creates a connector object. A component can:

The following is the code in userConnector.ts file for the connector creation.

const [connector] = useState<Connector>(
    () =>
      new Connector(
        (_: Connector) => {
          Logger.info('Connector initialized.')

          setConnectorState({ ...connectorState, isInitializing: false })
        },
        (_: Connector, error: Error) => {
          Logger.error(`Connector Initialized Error: ${error.message}`)
          setConnectorState({ ...connectorState, errorMessage: error.message, isInitializing: false })
        }
      )
  )

Step 5: Update the fetcher file

Replace the example URL(s) in the handlers/DataFetcher.ts file to reflect the URL(s) where your data is stored.

  urls: Record<string, string> = {
    csv: 'https://www.example.com/user.csv',
    excel: 'https://www.example.com/user.xslx',
  } as const

Step 6: Configure how the data is presented

Now you must define how you want to map the data to one or more or tables. This mapping of data is done in the schema.

To decide how to map your data, look at your data source. When you’re done looking at the summary of the JSON data source, make the necessary edits to structure the returned data.

The custom parse connector has two default parsers.

The handlers/CsvDataParser.ts file:

import { CsvUtils, DataContainer, DataType, ParseOptions, AsyncParser } from '@tableau/taco-toolkit/handlers'

export default class CsvDataParser extends AsyncParser<Uint8Array> {
  async parse(fetcherResult: Uint8Array, { dataContainer }: ParseOptions): Promise<DataContainer> {
    const tableName = 'csv-data-table'

    const containerBuilder = AsyncParser.createContainerBuilder(dataContainer)
    const { tableBuilder } = containerBuilder.getTable(tableName)

    // fetcherResult contains the CSV content yielded from Fetcher
    const { headers, rows } = await CsvUtils.parse(fetcherResult, { hasHeader: true })
    if (!headers) {
      throw new Error('The parsing result does not include the expected headers.')
    }

    tableBuilder.addColumnHeaders(headers.map((id: string) => ({ id, dataType: DataType.String })))

    const dataRows = await CsvUtils.createDataRows(rows, headers)
    tableBuilder.addRows(dataRows)

    return containerBuilder.getDataContainer()
  }
}

The handlers/ExcelDataParser.ts file:

import { ExcelUtils, DataContainer, ParseOptions, AsyncParser } from '@tableau/taco-toolkit/handlers'

export default class ExcelDataParser extends AsyncParser<Uint8Array> {
  async parse(fetcherResult: Uint8Array, { dataContainer }: ParseOptions): Promise<DataContainer> {
    // fetcherResult contains the Excel content yielded from Fetcher
    const tables = await ExcelUtils.parse(fetcherResult)
    const containerBuilder = AsyncParser.createContainerBuilder(dataContainer)

    containerBuilder.appendTables(tables)

    return containerBuilder.getDataContainer()
  }
}

Step 7: Build your connector

Enter these commands to build, pack, and run your new connector:

taco build

taco pack

taco run Desktop