Using the test assistant
Use the Test assistant to ask questions about your data and verify that your data has been correctly transferred and is up-to-date. It also lets verify that you are not sharing any wrong or confidential data.
Asking questions
Start the test assistant from the Data task page:
- Run the data task.
- When it is completed, click Test assistant in the menu bar. The Test assistant opens in the right panel.
To ask a question, enter it and send it by:
- Pressing ENTER on your keyboard.
- Clicking Enter.
The data can be used by the test assistant only when they are defined as Metadata.
- Navigate to the Datasets tab.
- Select the Metadata check box.
- Run the task. Once it is completed, you can ask questions about the selected data.
The questions history
You can view, copy, delete, and re-ask previous questions. You can also copy the answers.
Conversations cannot be saved.
Click 
and hover over the question and select the action: Copy, Resend, or Delete.
Deleting a question from the history cannot be undone.
Settings
| Settings | Description |
| Number of documents in context | The number of relevant documents that will be passed to the model as context. |
| Prompt template | Enter the template the AI must follow to filter the documents to be included. |
| Filter | Enter the expression to filter the documents to be included.
As the filter is based on the metadata and the file-based knowledge marts do not have metadata, think carefully of the filter you are configuring. It might be more relevant to exclude data instead of including them. For more information, see Using the test assistant. |
| Document retrieval | Select the option from the drop-down list:
|
| Answers generation | Select the option from the drop-down list:
|
Examples of filters
The test assistant can help you define the Filter to use in the Settings but the filter format depends on the vector databases.
Examples for knowledge marts
The table below shows some examples of questions you can ask and the filters returned by the test assistant.
| Vector database | Questions | Answer to use as a filter | Format | Documentation |
|---|---|---|---|---|
| Elasticsearch | Write a filter for documents where metadata.UnitsInStock > 39 | { "range": { "metadata.UnitsInStock": { "gt": 39 } } }
|
JSON | |
| OpenSearch | Write a filter for documents where metadata.UnitsInStock > 39 | { "range": { "metadata.UnitsInStock": { "gt": 39 } } }
|
JSON | Full-text queries |
| Pinecone | Mongo-style filter to get vectors where UnitsInStock > 39 | { "UnitsInStock": { "$gt": 39 } }
|
JSON | Filter by metadata |
| Snowflake Cortex | Write a Cortex SQL query to filter rows where metadata.UnitsInStock > 39 |
|
SQL | Filter syntax |
Remember that the answers depend on your data, those are only examples.
Examples for file-based knowledge marts
Use the Filter to limit the search to specific files. The table below shows some examples:
| Vector database | Filter | Format | Documentation |
|---|---|---|---|
| Elasticsearch | {"terms": {"source_id.keyword": ["gs://ai-ready/test/hello.txt"]}}
|
JSON | |
| OpenSearch | {"terms": {"source_id.keyword": ["gs://ai-ready/test/hello.txt"]}}
|
JSON | Full-text queries |
| Pinecone | {"source_id": "s3://username-filebased/docxfiles/newF/cv.docx"}
|
JSON | Filter by metadata |
| Snowflake Cortex | "source_id" like "s3://username-filebased/pdf/small_pdf/%"
|
SQL | Filter syntax |
Remember that the answers depend on your data, those are only examples.