Bugfix-#175

After last release, sheXer could brake when extracting shapes using a property both to point to literals and IRIs. That is fixed in this release.

Supprot to extraction involving BNodes in TURTLE_ITER and NT

• Extraction works properly even if there are BNodes as target instances or in the neighborhood of those whenever a non-rdflib parser (TURTLE_ITER, NT) is used. In the rest of cases, BNodes are ignored.
• No @ shape will appear in any constraint. The macro BNode is now properly used.
• The macro NONLITERAL has been introduced to generalize situations in which both BNodes and IRIs play the same role w.r.t some target nodes.
• Bug related to cardinality in OR constraints has been detected and corrected.

Let specify the instantiation property using prefixes

2.5.7

Merge branch 'develop'

Corrected a bug with blank nodes and endpoint mining

2.5.6

Merge branch 'master' of https://github.com/DaniFdezAlvarez/shexer

Compliant rdfs annotations at class level and constraint level

2.5.5

Merge branch 'develop'

log msg to std_err

Now the messages outputed with the verbose mode are written to std_err, so they are not mixed with resulting schemes being written in std_out

Fix a bug related to shapes names in UML generation

2.5.3

Fix release conflict

Fix dependence issue

Fix a building issue with a dependence non declared in setup.py

Support to XZ files

Added support to xz compressed files. If you want to process this content, set compression_mode to CONST.XZ when building the Shaper object

Support for examples as inline comments

Now you can use the parameter examples_mode of the class Shaper (with ShEx outputs) for getting examples of instances matching a shape or cnode constraints matching a triple constraint as comments in the obtained schemes. See README for more information.