SPARQL kernel and mybinder?

Binder mybinder.org ops
1

I would love to run this ipynb-file:

But mybinder reports that a SPARQL kernel is missing.

A few years ago an issue was raised indicating that a SPARQL kernel was available at that time: https requests from Sparql kernel · Issue #694 · jupyterhub/mybinder.org-deploy · GitHub

1 Like
2

I had raised this issue but it was closed and I was adviced to raise the question here:

3

Please share how are you launching a MyBinder session and seeing this information?

MyBinder uses accompanying configuration files to specify the environment in the specific session that gets launched. In other words, the configuration files specify what is installed and available in the active session. Searching ‘github mybinder sparql’ at Google leads to GitHub - betatim/sparql-notebooks: Sparql notebooks on binder where requirements.txt specifies sparqlkernel for the environment that gets made from sessions launched from that specific repo and the postBuild configuration file installs it into Jupyter for the user of that session.

Launch using the following URL (or click here) and run the code in the first cell and you’ll see sparql code working:

https://mybinder.org/v2/gh/betatim/sparql-notebooks/master?filepath=wikidata.ipynb

You’ll see that the author of the repo you referenced in the issue #694 purposefully(?) didn’t include sparql in the enviornment.yml file configuration file and didn’t install it in the postBuild. Note that the line jupyter sparqlkernel install in the postBuild is commented out so that it doesn’t run. (It probably wouldn’t have worked anyway because sqarql also not installed in the environment.) I suspect the author of that repo stopped tinkering with it around that time when they found the port needed was blocked at the time, and they left sparql as not getting installed for sessions launched from that repo. That’s speculation. However, the point remains that the configuration files in that repo are not clearly set to install sparql currently in MyBinder sessions launched from that repo.

3 Likes
4

Thanks a lot for the very detailed and helpful reply :smiley:

I will also inform Bob DuCharme, the author of the repo.

1 Like
5

similar missing coming from Wikidata:PAWS - Wikidata - it was documented before but now seems not adapted.

1 Like
6

Digging a bit more with the posts hereabove I solved the problem. If the kernel SPARQL is missing, just install it ! Open a Linux shell console and execute both commands (pip + jupiter) as said in the .md file of → GitHub - paulovn/sparql-kernel: A Jupyter kernel to launch queries against SPARQL endpoints .
Then under mybinder interface form generate the link YourUserName/YourNBdirectory.
Under paws it is similar. So that my both configurations are now operational and kernel SPARQL appears and is selectable in the menus. Note that the Notebook file extension is still .ipynb although you work with SPARQL.

https://mybinder.org/v2/gh/myGitUserName/sparql-notebooks/HEAD  

https://hub-paws.wmcloud.org/user/myMediaWikiUserName/lab

Really magic and powerful ! Thanks all ! :+1:

1 Like
7

Going deeper, if you have Jupyter on your local, you can also run the two commands above (pip + jupyter) to install SPARQL kernel and create your own notebooks. I already have Python + R and can now run the SPARQL files as the basic → http://localhost:8888/notebooks/notebooks/wikidata.ipynb . Wonderful again ! :+1:

1 Like