Closed gaurav closed 7 years ago
The idea with sending the output to stderr was to ensure that stdout was always well-formed Turtle: I didn't want any error messages getting in there!
The idea with sending the output to stderr was to ensure that stdout was always well-formed Turtle: I didn't want any error messages getting in there!
Right, but announcing success isn't an error message 😄
@hlapp: I got rid of the STDERR stuff and the sudo
, and I managed to set up Maven so that it creates a JAR file containing all the prerequisites it needs. I still need to build testShacl.jar locally, rather than on Travis, but I've filed that issue for another day: https://github.com/phyloref/phylo2owl/issues/14
Great, nice work, @gaurav! There's a lot of commits here now of all the trials and errors. I can just squash the whole PR into a single commit ("squash and merge"), or if you want to retain specific commits, you can do an interactive rebase squashing what can be squashed and keeping commits as individual ones as you would like.
Let me know. By default I would squash into a single commit and merge.
Yup, let's squash and merge it. The largest changes are the source files needed by the SHACL library; the others are mostly small files, and each file should be pretty straightforward to understand.
Provides a Java class that interfaces with TopBraid's SHACL library to validate an RDF/XML file against a file containing SHACL space. Also includes a Python script that can be called by
py.test
to run validations correctly and to report the result.This is pretty complicated, so it might be easier to consider two other options:
rdflib
. If we port our code to Java or Scala and use Apache Jena instead, we would end up with a single code base that doesn't depend on Python at all. This would probably make testing and deployment much simpler.