dpriskorn / qlever-control

QLever Control

Important: Use the python-qlever branch of this repository. The main branch only works with bash on certain Linux systems and is no longer maintained.

This is a very small repository. Its main contents is a script qlever that can control everything that QLever does. The script is supposed to be very easy to use and pretty much self-explanatory as you use it. If you use Docker, you don't even have to download any QLever code (Docker will pull anything it needs) and the script is all you need.

Directory structure

We recommend that you have a directory qlever for all things QLever on your machine, with subdirectories for the different components, in particular: qlever-control (this repository), qlever-indices (with a subfolder for each of your datasets), and qlever-code (only needed if you want to compile the QLever binaries on your machine instead of using Docker).

Make sure that the qlever-control directory (which contains the qlever script) is in your PATH. If you have compiled QLever binaries, the directory qlever-code/build (which contains these binaries) should also be in your PATH. Note that Docker is easier to use, but typically 10 - 20% slower compared to using the binaries directly.

Quickstart

Create an empty directory as a subdirectory of qlever-indices (see the previous section), go there, and call

eval "$(qlever setup-autocompletion)"
qlever setup-config olympics

The first line will enable autocompletion for the qlever script, which is very useful. You can also put that line in your .bashrc (or similar file if you are using another shell). The second line will create a Qleverfile preconfigured for the 120 Years of Olympics dataset, which is a great dataset to get started because it is small. To see the list of all available configs, type qlever help or just qlever. A dataset that is more interesting and larger, but can still be downloaded and indexed in a matter of minutes is dblp. Have a look at the Qleverfile and see whether the entries make sense (most of them are self-explanatory).

Now you can download the data, build an index for it (which QLever then uses to answer queries efficiently), start the server, and launch a test query as follows:

qlever get-data
qlever index
qlever start
qlever test-query

Each command will not only execute the respective action, but it will also show you the exact command line it uses. That way you can learn, on the side, how QLever works internally. If you just want to know the command used for a particular action, but not execute it, you can append "show" like this:

qlever index show

You can also perform a sequence of actions with a single call, for example:

qlever stop remove-index index start

There are many more actions. If you have enabled the autocompletion as described above, you can just type qlever and then TAB and you will get a list of all the available actions.