How to Connect Your Content Source to AutoClassifier

The following information describes how to connect your content source to AutoClassifier.

AutoClassifier Integration Prerequisites

  • Minimum Connectivity Hub supported version: v1.1

  • Minimum Connector Framework supported version: v4.0.5

  • Network connectivity (HTTP) from Connectivity Hub or Connector Framework to the AutoClassifier engine.

  • If you are using AutoClassifier with Windows Authentication, you must configure the Web.config file to consider requests made through Connectivity Hub or the Connector Framework as anonymous authentication. To do so, add one of the following to the Web.config file, depending on the service you are calling:

    Copy
    If WCF service is being called
    <!-- If WCF service is being called-->
    <location path="Enrichment.svc">
        <system.web>
            <authorization>
                <allow users="*" />
            </authorization>
        </system.web>
    </location>

    Copy
    If Rest API is being called
    <!-- IF Rest API is being called -->
    <location path="/API/ContentProcessing/ProcessItem">
        <system.web>
            <authorization>
                <allow users="*" />
            </authorization>
        </system.web>
    </location>

How to Integrate AutoClassifier with Connectivity Hub

AutoClassifier adds metadata for the items your Connectivity Hub crawls. Connectivity Hub must be configured to read the metadata and include them in the index.

To configure AutoClassifier with ConnectivityHub you configure Pipeline stage(s) to return metadata (properties). The AutoClassifier Pipeline stages must be configured to return at least one property.

To start, you identify the name, the type of the property, and whether they are multivalued. For example, category, keyword, (both text multi-valued).

By creating an AutoClassifier Pipeline stage, you tell your content source what data to expect as well as instruct your content source what to do with the data it receives.

Configure Your Content Source to Handle Incoming Data

To begin, you instruct your content source what data to expect.

  1. In Connectivity Hub edit your content source and select the Advanced tab.

  2. To configure the Enrichment pipeline integration settings, scroll down to the Enrichment pipeline integration section.

  3. Select Enrichment Web API or Enrichment Web Service and specify the following:

    1. If you selected Enrichment Web API, in the Service URL field, enter the AutoClassifier engine URL with enrichment service. For example: http://<AutoClassifierEngineURL>/api/ContentEnrichment/ProcessItem.

      1. If you want to call a specific pipeline, you can append the pipline ID to the Service URL. For more information, see Use your pipeline ID to call a specific pipeline in Connectivity Hub in the AutoClassifier documentation.

    2. If you selected Enrichment Web Service, in the Service URL field, enter the AutoClassifier engine URL to the enrichment service endpoint. For example: http://<AutoClassifierEngineURL>/enrichment.svc

    3. Authentication Mode: Select an authentication mode from the drop-down list:

      1. Use Service Account: Select this option to use the service account that was configured in Connectivity Hub for authentication.

      2. Specify User Account: If you select this option, enter the login credentials for the user account in the text fields that appear.

    4. Properties returned: enter a semicolon separated list of triplets:

      1. Each triplet consists of the property name, property type and whether there are multiple values - in true/false format.

      2. Include each property returned by AutoClassifier that is applicable to the content source being modified (edited).

    Connectivity Hub 3.0 and later will also send deleted and skipped items to AutoClassifier. If you do not want include deleted items, you configure a trigger script in AutoClassifier. For more information, see Add Triggers to Determine When Your Pipelines Run in the AutoClassifier documentation.

  4. Select Save.

Configure Your Content Source to Process Incoming Data

Next you instruct your content source what to do with the data it will receive:

  1. Select your content source and select Metadata from the drop-down menu in the Actions column.

  2. Create a new property.

  3. The type of your property must be the same type that was set in the Properties returned field above.

  4. Enter a Title and Description of your property.

  5. Select The value is calculated by an enrichment pipeline. The options available from the drop-down menu are the properties you set in the Enrichment pipeline integration above.

  6. Select the metadata property you want to include in your index.