Custom action

To integrate different services that takes several attributes from COM, process it and in return, populate the other offer attributes, complete the following steps:

Procedure

  1. Perform steps 1 through 4 in Creating custom templates.
  2. Select Custom Action
    A dialog appears.
  3. Under Metadata provide appropriate values for the following fields:
    • Action name - Mandatory field
    • Application to be Called - Mandatory field.

      Supports three applications, they are:

      • Snap-CAST
      • GoogleVision
      • GoogleLanguage
      • Azure
      Note: These four are out of the box plugins. However, customers and partners can build new plugins to integrate with any other systems. For more information, see Unica Content Integration Developer Guide for how to build plugins and Unica Content Integration Installation and Configuration Guide to on-board new plugins.
    • Service to be Called - Mandatory field.

      Depending on the application service autopopulates in the field.

      • Snap-CAST - sentiment-analysis
      • GoogleVision - image-analysis
      • GoogleLanguage - text-sentiment-analysis
      • Azure - image analysis
      • Snap-CAST - advertisement learning
    • Attributes to be passed (to service) - It is a Mandatory field.
      Note: In the top right corner of the UI, you will see a brief summary of selected service, which helps you in attribute selection.
      Note: For Snap-cast there is a bug restricting the visibility of help snap-cast leaning. Snap-cast is looking into this limitation (ticket no: HMA-334419) and it will be addressed in future versions. Below information can help to know about Snap-cast leaning.

      Snap-CAST learning offer machine learning services. User can train the machine to provide sentiment analysis based on the provided inputs. User need to provide at least one image url, one image, one text data and one rich text (csv formatted) attribute required for learning API. Rich text attribute to be mapped to 'All tags CSV', which must have comma separated text values. Initial value can be obtained by executing sentiment analysis custom action. Different categories of trainings are supported i,e 'Image','Subject','Keyword','Quantifiable','Bucket','Sentiment'Time sensitive event' and 'Custom grouping'.

      From the dropdown of the field select the attribute to associate this service.

  4. Under Response attributes mapping, map the service response with the COM attributes.
    Note: Refer the help tooltip to know about the list of supported attributes with each service and select only the supported attributes. In case any unsupported attribute is selected, then the API may not give intended result.
  5. Click on Save.