Adding a Search Collection Input Source

About this task

Once the autocomplete dictionary has been created, you must populate it using Input sources. There are two types of input sources that can be used with autocomplete: an input source from a search collection (discussed in this section), and an input source from a text file (discussed in the next section).

An autocomplete input source from a search collection extracts phrases from content element text nodes in a collection and stores them in the autocomplete dictionary.

To begin adding a search collection input source

Procedure

  1. Navigate to the Inputs tab of the autocomplete-tutorial-dictionary that you have just created
  2. Click the Add Input Source button.
    Note:

    If you navigated away from the dictionary, click on the list icon () next to the Dictionaries link in the left-side navigation bar and click on the name of the dictionary that you created.

    In the window that appears (as shown in Figure 1)

  3. Select Collection for Autocomplete from the list
  4. Click Add.
    Figure 1. Adding an Autocomplete Dictionary Input Source

    The input source has now been added to your autocomplete dictionary.

  5. You must define:
    • Which search collection the phrases will be extracted from
    • Which content elements in that collection will be extracted
    • If any metadata will be collected for any phrase that is collected

    Enter example-metadata in the Collection Name field, which instructs the dictionary to use the example-metadata collection as this input source.

    When creating your own autocomplete dictionary, the content elements that will be used as phrases must be carefully chosen. The phrases that are collected should ones that you consider to be valid in a query. For example, if we added the snippet content element to the autocomplete dictionary, the phrases would be very long (too long for the autocomplete suggestion box to return) and many words will be repeated in multiple phrases, which will make the user work harder to refine their query.

    The example-metadata collection contains two content elements that are appropriate to add to an autocomplete dictionary, title and hero. Add title and hero to the Contents field, each word on a new line.

    Optionally, metadata can also be extracted and stored with the phrases. The metadata can then be used to enhance the results provided to the user. The display automatically identifies some metadata contents and uses them to enhance the result. These contents elements are:

    • url - Turns the suggestion into a link to that URL instead of going to the search results for that phrase.
    • image - Hotlinks the image url in an <img> to the left of the autocomplete suggestion.
    • description - Displays the content of the description below the autocomplete suggestion.

    The example-metadata search collection has a content element named url, so add url to the Metadata field.

    Note: Metadata is only collected if specified in the input source. If the metadata is not collected, it cannot be used to enhance an autocomplete suggestion.
  6. For this tutorial, the other fields can be left at the default values.

    For more information on these fields, see the Building Autocomplete Data from a Search Collection section of the autocomplete documentation.

    When complete, your search collection input source will look like Figure 2.

    Figure 2. Adding a Search Collection Input Source
  7. Click OK to complete the configuration of this input source.
Parent topic: Creating an Autocomplete Dictionary