RAG Document Hub
RAG Document Hub serves as a library for RAG documents. You can browse existing files using various filtering options or upload new ones.
Step 1. Open the RAG Document Hub application in the AI Model group:
Here is the main view of the application :
Step 2. Switch RAG Destination
RAG Destinations are declared in ML Models Hub. Multiple destinations (connections) may be available, each containing a different set of files.
To switch the RAG destination:
(1) Click the Switch RAG Destination button in the top-right corner.
(2) Select the RAG destination.
(3) Press the Switch button.
Step 3. Filtering options
Press the icon on the right side of each filter to see possible options or type in a value for search. You can also start typing in, and matching options will appear.
Press the Go button or press Enter on keyboard to see the search results:
Click Adapt Filters to choose which filters should be visible.
Tick or Untick Fields to show/hide them.
Use up/down arrows to change order of filters.
Finally, press OK to save changes.
Step 4. Documents Table & View Settings
Yellow columns are standard. Green columns are destination-specific, corresponding to attributes defined for each destination. Attribute settings are managed in the ML Models Hub application.
To customize the table view:
(1) Click the settings icon in the top-right corner of the table. The View Settings panel will open.
(2) Hover over a column name and use the drag controls to change the order.
(3) Select columns to show or hide.
(4) Click OK to confirm changes.
Step 5. Uploading Documents to RAG Document HUB
You can upload your files to RAG Document HUB using two methods: Manual Upload or Automated Link-Based Retrieval.
In order to add a file, click on the Add button in the top right corner of the RAG Document Hub:
Option 1. Upload File from local Storage
This method allows you to upload a document directly from your local storage. If the document is updated, you must reupload it to ensure the latest version is available.
Best for: Static documents that do not change frequently.
How to Upload a Document Manually:
(1) Select file – Select the document file from its location on your computer.
(2) URL for Source Document – Add a URL link if you want the document accessible in AI Model scenarios via a web link (Optional).
(3) Document Attributes – enter values for any document attributes.
NOTE: Attribute configurations are managed at the RAG Connection level in the ML Models Hub
Click Upload to send the document to the RAG system.
Option 2. Add Link from Shared Location
With this method, the system automatically retrieves and synchronizes the latest version of a document from a shared storage location (e.g., SharePoint). This eliminates the need for manual uploads and ensures the document remains up to date.
Best for: Dynamic documents that require frequent updates without manual intervention.
How to Set Up Automated Link-Based Retrieval:
(1) Synchronized With – Choose the destination where your document will be synchronized. This must be configured in the ML Models Hub
(2) URL for Document from Shared Location – Open the document in its shared location, copy the link from the browser, and paste it into this field.
(3) Document Name – The system will automatically retrieve the document name from the shared location.
(4) Synchronization Frequency – The system applies a default sync frequency, but you can adjust it to match your document’s update cycle.
(5) Document Attributes – Add any required custom attributes.
NOTE: These are managed in the ML Models Hub for your selected RAG destination.
Possible errors during document upload:
While uploading documents to the RAG Document Hub, it is possible encounter the following types of errors:
- Unauthorized – Occurs when the user does not have the correct permissions or their session has expired.
- Internal Server Error – Indicates a problem with the backend system, such as a database failure. This error is typically temporary and may require admin support.
- Error While Parsing – Happens when the system is unable to read or interpret the uploaded file, often due to an unsupported or corrupted file format.
- RAG Error – Arises when the system has trouble breaking the document into smaller parts for processing, possibly due to file structure or content issues.
Step 6. Preview File
Click on the row containing the document you want to preview.
Use the controls to resize the preview window (arrow on the left side) or enter full screen (top-right corner icon).
Click the X icon in the top-right corner to close the preview.
Step 7. Search within document
You can search information by direct quote or logical meaning, using RAG Search functionality. Press the RAG Search Simulation button.
Use filters to pre-filter source documents. Only visible documents (documents that are result of filtering) will be used for retrieving the RAG Search results.
Write a search phrase in the search field, and press Enter. Or press the arrow icon.
Press the window resize button (1) to set RAG Search Window to fullscreen.
Press the blue document name in results (2) to open source document in preview mode.
You can modify number of results and search accuracy ( strictness) by pressing the settings button:
Access the help dialogue by pressing the “i” Icon.
Edit the values, and press Save.
Step 8. Edit File
Select the file you want to edit and click the Edit button:
Step 9. Delete File
Select one or multiple files and press the Bin button. Confirm deletion when prompted.
