Debug search issues

You’ve built your search solution but have noticed that there are some warnings on the indexer.

In this exercise, you’ll create an Azure AI Search solution, import some sample data, and then resolve a warning on the indexer.

Note To complete this exercise, you will need a Microsoft Azure subscription. If you don’t already have one, you can sign up for a free trial at https://azure.com/free.

Create your search solution

Before you can begin using a Debug Session, you need to create an Azure Cognitive Search service.

select this button to deploy all the resources you need in the Azure portal.
Under Resource Group, select Create new.
Type acs-cognitive-search-exercise.
Select the closest Region to you.
For Resource Prefix, enter acslearnex and add a random combination of numbers or characters to ensure the storage name is unique.
For the Location, select the same region you used above.
At the bottom of the pane, select Review + create.
Wait until the resource is deployed, then select Go to resource group.

Import sample data

With your resources created, you can now import your source data.

In the listed resources, select the search service.
On the Overview pane, select Import data.
On the import data pane, for the Data Source, select Samples.
In the list of samples, select hotels-sample.
Select Next:Add cognitive skills (Optional).
Expand the Add enrichments section.
Select Text Cognitive Skills.
Select Next:Customize target index.
Leave the defaults, then select Next:Create an indexer.
Select Submit.

Use a debug session to resolve warnings on your indexer

The indexer will now begin to ingest 50 documents. However, if you check the status of the indexer you’ll find that there’s warnings.

A screenshot showing 150 warnings on the indexer.

Select Debug sessions in the left pane.
Select + Add Debug Session.
Select Choose an existing connection for Storage connection string, then select your storage account.
Select + Container to add a new container. Name it acs-debug-storage.
Set its Anonymous access level to Container(anonymous read access for containers and blobs).

Note: You may need to enable blob anonymous in order to select this option. To do so, in the storage account go to Configuration , set Allow Blob anonymous access to Enabled and then select Save.
Select Create.
Select your new container in the list, then select Select.
Select hotel-sample-indexer for the Indexer Template.
Select Save Session.

The dependency graph shows you that for each document there’s an error on three skills.
Select V3.
On the skills details pane, select Errors/Warnings(1).
Expand the Message column so you can see details.

The details are:

Invalid language code ‘(Unknown)’. Supported languages: ar,cs,da,de,en,es,fi,fr,hu,it,ja,ko,nl,no,pl,pt-BR,pt-PT,ru,sv,tr,zh-Hans. For additional details see https://aka.ms/language-service/language-support.

If you look back at the dependency graph, the Language detection skill has outputs to the three skills with warnings. Also the skill input causing the error is languageCode.
In the dependency graph, select Language detection.

Looking at the skill settings JSON, note the field being used to deduce the language is the HotelId.

This field will be causing the error as the skill can’t work out the language based on an ID.

Resolve the warning on the indexer

Select source under inputs, and change the field to /document/Description.
Select Save.
Select Run.

The indexer should no longer have any errors or warnings. The skillset can now be updated.
Select Commit changes…
Select OK.
Now you need to make sure that your skillset is attached to an Azure AI Services resource, otherwise you’ll hit the basic quote and the indexer will timeout. To do this, select Skillsets in the left pane, then select your hotels-sample-skillset.
Select the AI Services tab, then select the AI services resource in the list.
Select Save.
Now run your indexer to update the documents with the fixed AI enrichments. To do this select Indexers in the left pane, select hotels-sample-indexer, then select Run. When it has finished running, you should see that the warnings are now zero.

[!TIP] Now you’ve completed the exercise, if you’ve finished exploring Azure AI Search services, delete the Azure resources that you created during the exercise. The easiest way to do this is delete the acs-cognitive-search-exercise resource group.