Merge pull request #292 from mpsonntag/bagToSeq

[RDF] 'Bag' to 'Seq'

.travis.yml

@@ -12,18 +12,21 @@ matrix:
       python: "3.4"
     - os: linux
       python: "3.5"
+    - os: linux
+      python: "3.6"
 
     - os: osx
       language: generic
       env:
-        - OSXENV=3.5
-#    Keep only one osx branch active for now
-#    since currently osx builds on travis
-#    are frequently stalled or indefinitely delayed.
-#    - os: osx
-#      language: generic
-#      env:
-#        - OSXENV=2.7
+        - OSXENV=3.5.0
+    - os: osx
+      language: generic
+      env:
+        - OSXENV=3.6.0
+    - os: osx
+      language: generic
+      env:
+        - OSXENV=2.7.14
 
 install:
   - export PYVER=${TRAVIS_PYTHON_VERSION:0:1}
@@ -40,18 +43,20 @@ install:
     fi;
 
   - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
-      if [[ "$OSXENV" == "2.7" ]]; then
-        brew install python@2;
-        virtualenv venv -p python;
-        source venv/bin/activate;
-        export PYCMD=python;
-        export PIPCMD=pip;
-      else
-        brew upgrade python;
-        source venv/bin/activate;
-        export PYCMD=python3;
-        export PIPCMD=pip3;
-      fi;
+      brew update;
+      brew install pyenv;
+      brew upgrade pyenv;
+      brew install pyenv-virtualenv;
+      eval "$(pyenv init -)";
+      eval "$(pyenv virtualenv-init -)";
+      pyenv install $OSXENV;
+      pyenv virtualenv $OSXENV venv;
+      pyenv activate venv;
+      which python;
+      python --version;
+      which pip;
+      export PYCMD=python;
+      export PIPCMD=pip;
     fi;
 
   - $PIPCMD install lxml enum34 pyyaml rdflib

appveyor.yml

@@ -4,10 +4,14 @@ environment:
       PYVER: "2.7"
     - PYTHON: "C:\\Python35"
       PYVER: "3.5"
+    - PYTHON: "C:\\Python36"
+      PYVER: "3.6"
     - PYTHON: "C:\\Python27-x64"
       PYVER: "2.7"
     - PYTHON: "C:\\Python35-x64"
       PYVER: "3.5"
+    - PYTHON: "C:\\Python36-x64"
+      PYVER: "3.6"
 
 build: false
 

doc/odml_ontology/root-ontology.ttl

@@ -56,7 +56,7 @@ xsd:date rdf:type rdfs:Datatype ;
 ###  https://g-node.org/projects/odml-rdf#hasValue
 :hasValue rdf:type owl:ObjectProperty ;
           rdfs:domain :Property ;
-          rdfs:range rdf:Bag ;
+          rdfs:range rdf:Seq ;
           rdfs:label "hasValue"^^xsd:string .
 
 
@@ -94,7 +94,7 @@ xsd:date rdf:type rdfs:Datatype ;
 
 ###  https://g-node.org/projects/odml-rdf#isValueOf
 :isValueOf rdf:type owl:ObjectProperty ;
-           rdfs:domain rdf:Bag ;
+           rdfs:domain rdf:Seq ;
            rdfs:range :Property ;
            rdfs:label "isValueOf"^^xsd:string .
 
@@ -211,15 +211,15 @@ xsd:date rdf:type rdfs:Datatype ;
 #    Classes
 #################################################################
 
-###  http://www.w3.org/1999/02/22-rdf-syntax-ns#Bag
-rdf:Bag rdf:type owl:Class ;
+###  http://www.w3.org/1999/02/22-rdf-syntax-ns#Seq
+rdf:Seq rdf:type owl:Class ;
         rdfs:subClassOf [ rdf:type owl:Restriction ;
                           owl:onProperty :isValueOf ;
                           owl:someValuesFrom :Property
                         ] ;
-        rdfs:comment "The class of unordered containers." ;
+        rdfs:comment "The class of ordered containers." ;
         rdfs:isDefinedBy rdf: ;
-        rdfs:label "Bag" .
+        rdfs:label "Seq" .
 
 
 ###  https://g-node.org/projects/odml-rdf#Cell
@@ -346,7 +346,7 @@ rdf:Bag rdf:type owl:Class ;
           rdfs:subClassOf owl:Thing ,
                           [ rdf:type owl:Restriction ;
                             owl:onProperty :hasValue ;
-                            owl:someValuesFrom rdf:Bag
+                            owl:someValuesFrom rdf:Seq
                           ] ,
                           [ rdf:type owl:Restriction ;
                             owl:onProperty :isPropertyOf ;

odml/tools/rdf_converter.py

@@ -5,6 +5,7 @@
 from io import StringIO
 from os.path import dirname, abspath
 from rdflib import Graph, Literal, URIRef
+from rdflib.graph import Seq
 from rdflib.namespace import XSD, RDF
 
 import odml
@@ -123,11 +124,27 @@ def save_element(self, e, node=None):
             elif isinstance(fmt, Property.__class__) and \
                     k == 'value' and len(getattr(e, k)) > 0:
                 values = getattr(e, k)
-                bag = URIRef(odmlns + str(uuid.uuid4()))
-                self.g.add((bag, RDF.type, RDF.Bag))
-                self.g.add((curr_node, fmt.rdf_map(k), bag))
+                seq = URIRef(odmlns + str(uuid.uuid4()))
+                self.g.add((seq, RDF.type, RDF.Seq))
+                self.g.add((curr_node, fmt.rdf_map(k), seq))
+                # rdflib so far does not respect RDF:li item order
+                # in RDF:Seq on loading so we have to use custom
+                # numbered Node elements for now. Once rdflib upgrades
+                # this should be reversed to RDF:li again!
+                # see https://github.com/RDFLib/rdflib/issues/280
+                # -- keep until supported
+                # bag = URIRef(odmlns + str(uuid.uuid4()))
+                # self.g.add((bag, RDF.type, RDF.Bag))
+                # self.g.add((curr_node, fmt.rdf_map(k), bag))
+                # for v in values:
+                #     self.g.add((bag, RDF.li, Literal(v)))
+
+                counter = 1
                 for v in values:
-                    self.g.add((bag, RDF.li, Literal(v)))
+                    pred = "%s_%s" % (str(RDF), counter)
+                    self.g.add((seq, URIRef(pred), Literal(v)))
+                    counter = counter + 1
+
             # adding entities' properties
             else:
                 val = getattr(e, k)
@@ -258,9 +275,20 @@ def parse_property(self, prop_uri):
             elems = list(self.g.objects(subject=prop_uri, predicate=attr[1]))
             if attr[0] == "value" and len(elems) > 0:
                 prop_attrs[attr[0]] = []
+
+                # rdflib does not respect order with RDF.li items yet, see comment above
+                # support both RDF.li and rdf:_nnn for now.
+                # Remove rdf:_nnn once rdflib respects RDF.li order in an RDF.Seq obj.
                 values = list(self.g.objects(subject=elems[0], predicate=RDF.li))
-                for v in values:
-                    prop_attrs[attr[0]].append(v.toPython())
+                if len(values) > 0:
+                    for v in values:
+                        prop_attrs[attr[0]].append(v.toPython())
+                else:
+                    # rdf:__nnn part
+                    valseq = Seq(graph=self.g, subject=elems[0])
+                    for seqitem in valseq:
+                        prop_attrs[attr[0]].append(seqitem.toPython())
+
             elif attr[0] == "id":
                 prop_attrs[attr[0]] = prop_uri.split("#", 1)[1]
             else:

test/resources/version_conversion.yaml

@@ -11,10 +11,10 @@ Document:
     properties:
     - name: prop_name
       values:
-      - dtype: !!python/object/apply:odml.dtypes.DType
+      - dtype:
         - string
         value: '[''one'', ''two'']'
-      - dtype: !!python/object/apply:odml.dtypes.DType
+      - dtype:
         - int
         value: '1'
 odml-version: '1'

test/resources/version_conversion_int.yaml

@@ -14,23 +14,23 @@ Document:
       values:
       - dtype: string
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - string
         filename: filename val 1
         reference: ref val 1
         uncertainty: 11
         unit: arbitrary
         value: '1'
       - definition: def val 2
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 2
         reference: ref val 2
         uncertainty: 12
         unit: arbitrary 2
         value: '2'
       - definition: def val 3
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - string
         filename: filename val 3
         reference: ref val 3
@@ -44,63 +44,63 @@ Document:
       values:
       - dtype: string
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '1'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '2'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '3'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '4'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '4'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '4'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
         uncertainty: 12
         unit: arbitrary
         value: '4'
       - definition: def val 1
-        dtype: !!python/object/apply:odml.dtypes.DType
+        dtype:
         - int
         filename: filename val 1
         reference: ref val 1
@@ -132,31 +132,31 @@ Document:
         values:
         - dtype: string
         - definition: def val 1
-          dtype: !!python/object/apply:odml.dtypes.DType
+          dtype:
           - int
           filename: filename val 1
           reference: ref val 1
           uncertainty: 12
           unit: arbitrary
           value: '4'
         - definition: def val 1
-          dtype: !!python/object/apply:odml.dtypes.DType
+          dtype:
           - int
           filename: filename val 1
           reference: ref val 1
           uncertainty: 12
           unit: arbitrary
           value: '4'
         - definition: def val 1
-          dtype: !!python/object/apply:odml.dtypes.DType
+          dtype:
           - int
           filename: filename val 1
           reference: ref val 1
           uncertainty: 12
           unit: arbitrary
           value: '4'
         - definition: def val 1
-          dtype: !!python/object/apply:odml.dtypes.DType
+          dtype:
           - int
           filename: filename val 1
           reference: ref val 1

test/test_parser_odml.py

@@ -85,8 +85,11 @@ def test_rdf_file(self):
         self.rdf_writer.write_file(self.odml_doc, self.rdf_file)
         rdf_doc = self.rdf_reader.from_file(self.rdf_file, "turtle")
 
-        # RDF does not preserve the order of sections,
-        # need to check the attributes by hand.
+        self.assertEqual(self.odml_doc, rdf_doc[0])
+
+        # RDF does not preserve the order of sections or properties,
+        # check the attributes by hand to make sure everything
+        # was correctly imported.
         self.assertEqual(len(rdf_doc), 1)
         self.assertEqual(rdf_doc[0].author, self.odml_doc.author)
         self.assertEqual(rdf_doc[0].version, self.odml_doc.version)

test/test_rdf_writer.py

@@ -90,6 +90,8 @@ def test_adding_values(self):
         w = RDFWriter([doc])
         w.convert_to_rdf()
         self.assertEqual(len(list(w.g.subject_objects(predicate=RDF.li))), 0)
+        self.assertEqual(len(list(
+            w.g.subject_objects(predicate=URIRef("%s_1" % str(RDF))))), 0)
 
         doc = parse("""
             s1[t1]
@@ -98,14 +100,22 @@ def test_adding_values(self):
 
         w = RDFWriter([doc])
         w.convert_to_rdf()
-        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val")))), 1)
+        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li,
+                                               object=Literal("val")))), 0)
+        self.assertEqual(len(list(w.g.subjects(predicate=URIRef("%s_1" % str(RDF)),
+                                               object=Literal("val")))), 1)
 
         doc.sections[0].properties[0].append("val2")
         w = RDFWriter([doc])
         w.convert_to_rdf()
-        self.assertEqual(len(list(w.g.subject_objects(predicate=RDF.li))), 2)
-        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val")))), 1)
-        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val2")))), 1)
+        self.assertEqual(len(list(w.g.subject_objects(predicate=RDF.li))), 0)
+        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val")))), 0)
+        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val2")))), 0)
+
+        self.assertEqual(len(list(w.g.subjects(predicate=URIRef("%s_1" % str(RDF)),
+                                               object=Literal("val")))), 1)
+        self.assertEqual(len(list(w.g.subjects(predicate=URIRef("%s_2" % str(RDF)),
+                                               object=Literal("val2")))), 1)
 
         doc = parse("""
              s1[t1]
@@ -117,7 +127,9 @@ def test_adding_values(self):
 
         w = RDFWriter([doc])
         w.convert_to_rdf()
-        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val")))), 3)
+        self.assertEqual(len(list(w.g.subjects(predicate=RDF.li, object=Literal("val")))), 0)
+        self.assertEqual(len(list(w.g.subjects(predicate=URIRef("%s_1" % str(RDF)),
+                                               object=Literal("val")))), 3)
 
     def test_section_subclass(self):
         p = os.path.join(dirname(dirname(abspath(__file__))), 'doc', 'section_subclasses.yaml')