More details needed to get started please

[pwin]

I would appreciate more information to get going with qEndpoint.

• What are the general principles of operation? It seems that with the loading of successive RDF files the database gets wiped before each is loaded, so how can multiple RDF files be loaded?
• Loading a trig file does not seem to create a named graph.... so what is the way to load data into a named graph?
• What is the URL structure for the HTTP API, and are there any "non-standard" options to this?

[ate47]

For the general principles of this database, you index an RDF graph, and after you apply changes to this graph using SPARQL calls.

The load function will erase the database and put the new graph instead. That's why you have this effect.

For the named graph, this is not technically fully tested yet, so it might not be the safer database to use.
We are using an HDT file to store the original RDF graph, by default this file is configured without named graph capabilities, but implementations exist to support that.
If you run the rdf2hdt -printoptions command in a terminal, you will get all the available versions inside the config dictionary.type. For named graph I suggest to use the config dictionary.type=dictionaryMultiObjLangQuad.

The configs are set in the file qendpoint/repo_model.ttl with the mdlc:hdtSpec property, if you add this line it will use this HDT:

mdlc:main mdlc:hdtSpec "dictionary.type=dictionaryMultiObjLangQuad" .

About the API, you have the /api/endpoint/sparql which is standard, for the others, we don't have a list yet. If you know how Spring is working, the controller is this one.

The different options aren't documented either, if you add #no_optimizer at the start of your query, you'll disable the optimizer and #get_plan to get the query plan inside the result JSON.

[pwin]

Thanks - this is really helpful. What is the recommended procedure for loading many individual RDF files (e.g. Turtle files) please?

[ate47]

With the UI, I don't think you have an option.
Using the CLI you can use the rdf2hdt. The documentation is here, in the configs, the end dictionary should be dictionaryMultiObjLang.
It should accept a directory of RDF files.

Then it'll produce a HDT file, remove everything in your qendpoint/hdt-store directory and then put this file in it with the name index_dev.hdt.