All the import components can be created from a single dialog, but the steps below create them one at a time:

1. First, create a container for the data to be imported. This will be the import's target container.
2. Upload an RDF file. Click on the Create expand_more dropdown at the bottom of the page and select File. This will open a new form.
3. Fill out the Item fields — the document's Title (required) and optional Description, the same fields used when creating a document.
4. Fill out the mandatory File fields:

   Filename

   Click on Choose file and select an RDF file from your computer

   Title

   File title

5. Once all the fields have been filled in, click Save.
6. The browser will open your uploaded file.
7. If you are importing with a transformation, create a transformation query that will transform the uploaded RDF to a new RDF structure. Click on the Create expand_more dropdown at the bottom of the page and select CONSTRUCT. This will open a new form. Fill out the Item fields and the mandatory CONSTRUCT fields:

   Text

   SPARQL CONSTRUCT query string

   You will not be able to save the query if its syntax is not valid SPARQL 1.1.

   Title

   Query title

   Once all the fields have been filled in, click Save. You can find the query saved in the Queries container, which you can open from the document tree in the left sidebar.
8. Click on the Create expand_more dropdown button at the bottom of the page and select RDF Import. This will open a new form.
9. Fill out the Item fields.
10. Fill out the mandatory RDF Import fields:

    File

    Start typing the name of your RDF file and select it from the typeahead

    Title

    Import title

11. Select either the target graph or the transformation query — but not both:

    Name

    Select the target graph (document) using the typeahead

    Query

    Select the transformation query using the typeahead

12. Click Save and you will be redirected to the newly created import document.
13. Reload the document after a few moments to see the import progress. On success, the amount of data imported is displayed; on failure, the constraint violations or other import errors.
14. Go back to the target document to see the newly imported data items.

The import is described by a document, which the command takes as its last argument, so create that document first. With the CLI installed and the LDH_* variables set:

doc=$(ldh create item \
    --title "UNESCO category mappings" \
    --container "${LDH_BASE}categories/")

ldh import rdf \
    --title "UNESCO category mappings" \
    --rdf-file categories/unesco-mappings.ttl \
    --content-type "text/turtle" \
    --graph "${LDH_BASE}categories/mappings/" \
    "$doc"

Pass either --graph with the target document's URI or --query-file with the path to a transformation query, but not both. In the Northwind demo, the transformation query routes each mapping into its category's own document.

The command prints the URI of the import it created, a fragment of $doc. The import runs in the background; add --wait to return only once it has ended, with exit code 1 and the error if it failed.

Source: categories/unesco-mappings.ttl — the mapping file that connects Northwind categories to UNESCO Thesaurus concepts, shown in full in the RDF imports reference