add schema module

README.rst

@@ -1,25 +1,69 @@
-shaclgen
-========
+SHACLGEN
+===============
 
-shaclgen generates shacl templates based on the properties and classes
-present in a graph. This module uses the rdflib library for working with
-rdf.
+Shaclgen takes either a data graph(s) or schema(s) as input and generates a basic shape file based on the classes and properties present.
 
-From the command line:
-~~~~~~~~~~~~~~~~~~~~~~
+**Shape files from data graphs:**
+By default, the input graph is processed as a data graph (instance triples). Three formats are possible for data graphs: simple, nested, and extended.
 
+- Simple: Each class and property generate individual Node- and PropertyShapes.
+
+- Nested: Property shapes will be nested in nodeshapes iif they occur with one class.
+
+- Extended: Expands nested shapes to create individual property shapes for each property, in addition to nesting them when appropriate.
+
+**Shape files from ontologies:**
+If the input is a schema or ontology, shaclgen generates a nested shape file: properties with rdfs:domain defined in the ontology will be nested within the appropriate NodeShape. rdfs:range definitions for XML and rdfs datatypes are included.
+
+Added support for OWL constructions is planned.
+
+***************
+
+
+
+
+Installation
+***************
+Using pip:
+::
+
+ pip install shaclgen
+
+From source:
+
+https://github.com/alexiskeely/shaclgen
+
+
+Command line use:
+*****************
 ::
 
-    $ shaclgen [uri to data] [serialization]
+       $ shaclgen graph [optional arguments]
+
+Example usage:
+::
 
-Supported serializations include: - ``ttl`` for turtle - ``xml`` for
-rdf/xml - ``nt`` for ntriples
+  $ shaclgen https://www.lib.washington.edu/static/public/cams/data/datasets/uwSemWebParts/webResource-1-0-0.nt
 
-example:
 
+
+Command line arguments:
 ::
 
-    $ shaclgen https://www.lib.washington.edu/static/public/cams/data/datasets/uwSemWebParts/aggregation-1-0-0.ttl ttl
+  positional arguments:
+  graph                 The data graph(s).
+
+::
+
+  optional arguments:
+  -h, --help            show this help message and exit
+  -nf, --nested         generates a nested shape file
+  -ef, --extended       generates an expanded shape file
+  -o, --ontology        input file(s) or URL(s) is a schema or ontology
+  -s SERIALIZATION, --serialization SERIALIZATION
+                        result graph serialization, default is turtle
+
+***************
 
 This project is still in development. Comments, questions, and issues
 are welcome!

setup.py

@@ -10,7 +10,7 @@
 
 setup(
     name = 'shaclgen',
-    version = '0.1.3',
+    version = '0.1.4',
     packages = ['shaclgen'],
     description='Shacl graph generator',
     long_description=l_description,

shaclgen/__init__.py

@@ -0,0 +1,7 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Sat Nov 30 13:30:18 2019
+
+@author: alexi
+"""
+

shaclgen/__main__.py

@@ -2,38 +2,69 @@
 #!/usr/bin/env python
 
 #%%
-from .shaclgen import generate_groups, generate_triples, generate_shacl,generate_merged
+from .shaclgen import data_graph
+from .schema import schema
+
 import argparse
+from argparse import RawDescriptionHelpFormatter
+from argparse import RawTextHelpFormatter
 
-parser = argparse.ArgumentParser(description="""  
-                                 Shacl file generator.
-                                 Shaclgen will create a simple shape file by default: 
-                                 every class and property will get their own shape.
-                                 Nested and extended shape files are possible.""")
+parser = argparse.ArgumentParser(
+     formatter_class=RawDescriptionHelpFormatter,
+     description=("""
+    ---------------------------Shaclgen------------------------------------------
+   
+    Shaclgen takes either a data graph(s) or schema(s) as input and generates 
+    a basic shape file based on the classes and properties present. 
+    
+    Shape files from data graphs:
+    By default, the input graph is processed as a data graph (instance triples).
+    Three formats are possible for data grapghs: simple, nested, and extended.
+    
+    Simple: Each class and property generate individual Node- and PropertyShapes        
+    Nested: Property shapes will be nested in nodeshapes iif 
+            they occur with one class.
+    Extended: Expands nested shapes to create individual property shapes for each 
+            property, in addition to nesting them when appropriate. 
+        
+    Shape files from ontologies:
+    If the input is a schema or ontology (-o), shaclgen will generate 
+    a nested shape file: properties with rdfs:domain defined in the ontology 
+    will be nested within the appropriate NodeShape. rdfs:range definitions
+    for XML and rdfs datatypes are included."""))
 
-parser.add_argument("graph", nargs='+',type=str, help="the data graph(s)")
-parser.add_argument("serialization", type=str,help="the data graph rdf serialization")
+parser.add_argument("graph", nargs='+',type=str, help="The data graph(s).")
+
 group = parser.add_mutually_exclusive_group()
-group.add_argument("-nf", "--nested", action="store_true", help='Property shapes will be nested in nodeshapes iif they occur with one class.')
-group.add_argument("-ef", "--extended", action="store_true", help='Expands nested shapes to create individual property shapes for each property, in addition to nesting them when appropriate.')
+group.add_argument("-nf", "--nested", action="store_true", help='generates a nested shape file')
+group.add_argument("-ef", "--extended", action="store_true", help='generates an expanded shape file')
+parser.add_argument("-o", "--ontology", action="store_true", help='input file(s) or URL(s) is a schema or ontology')
+parser.add_argument("-s", "--serialization", help='result graph serialization, default is turtle')
 
 args = parser.parse_args()
-#
-#print(args.graph[0])
-#print(args.serialization)
+
 
 def main():
-    output = generate_groups(args.graph, args.serialization)
-    if args.nested:
-        triples = generate_triples(output, 'nf')
-    elif args.extended:
-        triples = generate_triples(output, 'ef')
+    if args.ontology:      
+        g = schema(args.graph)
+        if args.serialization:
+            print('...generating schema shape file...\n')
+            g.gen_graph(args.serialization)
+        else:
+            print('...generating schema shape file...\n')
+            g.gen_graph('turtle')
     else:
-        triples = generate_triples(output, 'sf')
+        kwargs = {'serial': 'turtle'}
+        g = data_graph(args.graph)
+        if args.nested:
+            kwargs['graph_format'] = 'nf'
+        elif args.extended:
+            kwargs['graph_format'] = 'ef'
+        if args.serialization:
+            kwargs['serial'] = args.serialization
+        print('...generating data shape file...\n')
+        g.gen_graph(**kwargs)
 
-    graph = generate_shacl(triples)
-    print(graph)
-#
 if __name__ == '__main__':
     main()
 

shaclgen/schema.py

@@ -7,9 +7,10 @@
 import collections
 
 class schema():
-    def __init__(self, graph=None,):
+    def __init__(self, args:list):
         self.G = Graph()
-        self.G.load(graph,format=guess_format(graph))
+        for graph in args:
+            self.G.parse(graph,format=guess_format(graph))
         self.CLASSES = collections.OrderedDict()
         self.PROPS = collections.OrderedDict()    
         self.REST = collections.OrderedDict()    
@@ -39,7 +40,9 @@ def extract_props(self):
 
 
         #gather property values
+        count = 0
         for prop in self.PROPS.keys():
+            count = count +1
             s = URIRef(prop)           
             self.PROPS[prop]['domain']= None
             self.PROPS[prop]['range']= None
@@ -57,7 +60,7 @@ def extract_props(self):
                 self.PROPS[prop]['e_prop'] = o
 
             for o in self.G.objects(subject=s, predicate=RDFS.label):
-                self.PROPS[prop]['label'] = o
+                self.PROPS[prop]['label'] = self.gen_shape_labels(prop)+str(count)
 
 
 
@@ -75,11 +78,12 @@ def extract_classes(self):
                 classes.append(s)
             else:
                 pass
-
+        count = 0    
         for c in sorted(classes):
             self.CLASSES[c] = {}
-            for c in self.CLASSES.keys():
-                self.CLASSES[c]['label'] = self.gen_shape_labels(c)
+        for c in self.CLASSES.keys():
+            count = count +1
+            self.CLASSES[c]['label'] = self.gen_shape_labels(c)+str(count)
 
     def extract_restrictions(self):
         # does not handle nested restrictions within other class descriptions
@@ -132,29 +136,30 @@ def gen_shape_labels(self, URI):
             label = URI.split("#")[-1]
         else:
             label = URI.split("/")[-1]
-        return label
+        return label+'_'
 
-    def gen_graph(self):
+    def gen_graph(self, serial='turtle'):
+
         self.extract_props()
         self.extract_classes()
         self.extract_restrictions()
         ng = Graph()
         SH = Namespace('http://www.w3.org/ns/shacl#')
-        ng.bind('SH', SH) 
+        ng.bind('sh', SH) 
 
         EX = Namespace('http://www.example.org/')
-        ng.bind('EX', EX)
+        ng.bind('ex', EX)
 
         # add class Node Shapes
         for c in self.CLASSES.keys():
-            label = self.gen_shape_labels(c)+'_ClassShape'
+            label = self.CLASSES[c]['label']
             ng.add((EX[label], RDF.type, SH.NodeShape))
             ng.add((EX[label], SH.targetClass, c))
         for p in self.PROPS.keys():
             if self.PROPS[p]['domain'] is not None:
                 blank = BNode()
                 if self.PROPS[p]['domain'] in self.CLASSES.keys():
-                    label = self.gen_shape_labels(self.PROPS[p]['domain'])+'_ClassShape'
+                    label = self.CLASSES[self.PROPS[p]['domain']]['label']
                     ng.add((EX[label], SH.property, blank))
                     ng.add((blank, SH.path, p))
                     if self.PROPS[p]['range'] is not None:
@@ -177,7 +182,7 @@ def gen_graph(self):
                         else:
                             pass
                 else:
-                    label = self.gen_shape_labels(self.PROPS[p])+'_PropShape'
+                    label = self.PROPS[p]['label']
                     ng.add((EX[label], RDF.type, SH.NodeShape))
                     ng.add((EX[label], SH.targetSubjectsOf, p))
                     ng.add((EX[label], SH.nodeKind, SH.BlankNodeOrIRI))
@@ -191,7 +196,7 @@ def gen_graph(self):
                                 ng.add((blank, SH['class'], rang ))
             else:
                 blank = BNode()
-                label = self.gen_shape_labels(p)+'_PropShape'
+                label = self.PROPS[p]['label']
                 ng.add((EX[label], RDF.type, SH.NodeShape))
                 ng.add((EX[label], SH.targetSubjectsOf, p))
                 ng.add((EX[label], SH.nodeKind, SH.BlankNodeOrIRI))
@@ -204,11 +209,8 @@ def gen_graph(self):
                         else:
                             ng.add((blank, SH['class'], rang ))
 
-        print(ng.serialize(format='turtle').decode())        
-        return ng        
+        print(ng.serialize(format=serial).decode())        
 
 
 
-    def save_graph(self, path):
-        ng = self.gen_graph()
-        ng.serialize(path, format='turtle')    
+