Merge pull request #76 from pyscal/improved_workflow_annotation

Improved workflow annotation

environment-workflows.yml

@@ -1,4 +1,4 @@
-name: workflow-rdf-v0.1
+name: workflow-rdf-v0.2
 channels:
   - conda-forge
 dependencies:

environment.yml

@@ -16,3 +16,6 @@ dependencies:
   - owlready2
   - plotly
   - ipywidgets
+  - sqlalchemy
+  - pip:
+    - "git+https://github.com/RDFLib/rdflib-sqlalchemy.git@develop"

pyscal_rdf/__init__.py

@@ -1,2 +1,3 @@
 from pyscal_rdf.graph import KnowledgeGraph
-from pyscal_rdf.structure import System
+from pyscal_rdf.structure import System
+from pyscal_rdf.workflow.workflow import Workflow

pyscal_rdf/graph.py

@@ -5,8 +5,6 @@
 """
 
 from rdflib import Graph, Literal, Namespace, XSD, RDF, RDFS, BNode, URIRef, FOAF, SKOS, DCTERMS
-from rdflib.store import NO_STORE, VALID_STORE
-from rdflib import plugin
 
 import os
 import numpy as np
@@ -26,9 +24,12 @@
 from pyscal_rdf.network.ontology import read_ontology
 from pyscal_rdf.structure import System
 import pyscal_rdf.properties as prp
+from pyscal_rdf.stores import create_store
+
 #from pyscal3.core import System
 from pyscal3.atoms import Atoms
 
+
 CMSO = Namespace("http://purls.helmholtz-metadaten.de/cmso/")
 PLDO = Namespace("http://purls.helmholtz-metadaten.de/pldo/")
 PODO = Namespace("http://purls.helmholtz-metadaten.de/podo/")
@@ -70,12 +71,6 @@ def _replace_keys(refdict, indict):
                 refdict[key] = val
     return refdict
 
-def _setup_structure_store(structure_store):
-    if structure_store is None:
-        structure_store = os.path.join(os.getcwd(), 'rdf_structure_store')
-    if not os.path.exists(structure_store):
-        os.mkdir(structure_store)
-    return structure_store
 
 class KnowledgeGraph:
     def __init__(self, graph_file=None, 
@@ -85,36 +80,12 @@ def __init__(self, graph_file=None,
         ontology=None,
         structure_store=None):
 
-        self.store_file = store_file
-        self.structure_store = structure_store
-
-
-        if store == "Memory":
-            self.graph = Graph(store="Memory", identifier=identifier)
-
-
-        elif store=="SQLAlchemy":
-            #check for modules
-            try:
-                import sqlalchemy as sa
-            except ImportError:
-                raise RuntimeError('Please install the sqlalchemy package')
-            try:
-                import rdflib_sqlalchemy as rsa
-            except ImportError:
-                raise RuntimeError('Please install the rdllib-sqlalchemy package. The development version is needed, please do pip install git+https://github.com/RDFLib/rdflib-sqlalchemy.git@develop')
-
-            if store_file is None:
-                raise ValueError("store file is needed if store is not memory")
-
-            self.graph = Graph(store="SQLAlchemy", identifier=identifier)            
-            uri = Literal(f"sqlite:///{store_file}")
-            self.graph.open(uri, create=True)        
-        else:
-            raise ValueError("Memory or SQLAlchemy")
+
+        create_store(self, store, identifier, 
+            store_file=store_file,
+            structure_store=structure_store)
 
         #start the storage
-        self.structure_store = _setup_structure_store(self.structure_store)
 
         #start binding
         self.graph.bind("cmso", CMSO)
@@ -135,7 +106,11 @@ def __init__(self, graph_file=None,
         self._atom_ids = None
         self.store = store
 
-
+
+    def add_structure(self, structure):
+        structure.graph = self
+        structure.to_graph()
+
     def add(self, triple):
         if str(triple[2].toPython()) != 'None':
             self.graph.add(triple)

pyscal_rdf/stores.py

@@ -0,0 +1,64 @@
+from rdflib.store import NO_STORE, VALID_STORE
+from rdflib import plugin
+from rdflib import Graph, Literal
+
+import os
+#special methods; for supporting workflow envs
+from pyscal_rdf.workflow import inform_graph
+
+
+def create_store(kg, store, identifier, 
+    store_file=None,
+    structure_store=None):
+
+    kg.store_file = store_file
+    if store == 'Memory':
+        store_memory(kg, store, identifier, store_file=store_file, structure_store=structure_store)
+    elif store == 'SQLAlchemy':
+        store_alchemy(kg, store, identifier, store_file=store_file, structure_store=structure_store)
+    elif type(store).__name__ == 'Project':
+        store_pyiron(kg, store, identifier, store_file=store_file, structure_store=structure_store)
+    else:
+        raise ValueError('Unknown store found!')
+
+
+def store_memory(kg, store, identifier, store_file=None, structure_store=None):
+    graph = Graph(store="Memory", identifier=identifier)
+    kg.graph = graph
+    kg.structure_store = _setup_structure_store(structure_store=structure_store)
+
+def store_alchemy(kg, store, identifier, store_file=None, structure_store=None):
+    _check_if_sqlalchemy_is_available()
+    if store_file is None:
+        raise ValueError("store file is needed if store is not memory")
+
+    kg.graph = Graph(store="SQLAlchemy", identifier=identifier)            
+    uri = Literal(f"sqlite:///{store_file}")
+    kg.graph.open(uri, create=True)
+    kg.structure_store = _setup_structure_store(structure_store=structure_store)
+
+
+def store_pyiron(kg, store, identifier, store_file=None, structure_store=None):        
+    structure_store = os.path.join(store.path, 'rdf_structure_store')
+    kg.structure_store = _setup_structure_store(structure_store=structure_store)
+    store_file = os.path.join(store.path, f'{store.name}.db')
+    store_alchemy(kg, store, identifier, store_file, structure_store=structure_store)      
+    #finally update project object
+    inform_graph(store, kg)
+
+def _check_if_sqlalchemy_is_available():
+    try:
+        import sqlalchemy as sa
+    except ImportError:
+        raise RuntimeError('Please install the sqlalchemy package')
+    try:
+        import rdflib_sqlalchemy as rsa
+    except ImportError:
+        raise RuntimeError('Please install the rdllib-sqlalchemy package. The development version is needed, please do pip install git+https://github.com/RDFLib/rdflib-sqlalchemy.git@develop')
+
+def _setup_structure_store(structure_store=None):
+    if structure_store is None:
+        structure_store = os.path.join(os.getcwd(), 'rdf_structure_store')
+    if not os.path.exists(structure_store):
+        os.mkdir(structure_store)
+    return structure_store

pyscal_rdf/structure.py

@@ -42,7 +42,7 @@ def _make_crystal(structure,
     element=None,
     primitive=False,
     graph=None,
-    names=True):
+    names=False):
 
     atoms, box, sdict = pcs.make_crystal(structure, 
         lattice_constant=lattice_constant,
@@ -70,7 +70,7 @@ def _make_general_lattice(positions,
     noise = 0,
     element=None,
     graph=None,
-    names=True):
+    names=False):
 
     atoms, box, sdict = pcs.general_lattice(positions,
         types,
@@ -98,7 +98,7 @@ def _make_grain_boundary(axis,
     repetitions = (1,1,1),
     overlap=0.0,
     graph=None,
-    names=True):
+    names=False):
 
     gb = GrainBoundary()
     gb.create_grain_boundary(axis=axis, sigma=sigma, 
@@ -194,7 +194,7 @@ def __init__(self, filename = None,
             species = None,
             source=None,
             graph=None,
-            names=True):
+            names=False):
 
         super().__init__(filename = filename, 
             format = format, 
@@ -543,7 +543,7 @@ def _add_lattice_properties(self):
         """
         data = self.schema.material.crystal_structure.unit_cell.lattice_parameter()
         lattice_parameter = URIRef(f'{self._name}_LatticeParameter')
-        self.graph.add((self.unit_cell, CMSO.hasLatticeParamter, lattice_parameter))
+        self.graph.add((self.unit_cell, CMSO.hasLatticeParameter, lattice_parameter))
         self.graph.add((lattice_parameter, RDF.type, CMSO.LatticeParameter))
         self.graph.add((lattice_parameter, CMSO.hasLength_x, Literal(data[0], datatype=XSD.float)))
         self.graph.add((lattice_parameter, CMSO.hasLength_y, Literal(data[1], datatype=XSD.float)))

pyscal_rdf/visualize.py

@@ -99,9 +99,10 @@ def visualize_graph(g,
             #we collapse sample information
             #if cmso.connector is found, only use it is it is cmso.hasCalculated
             #all sub sample props, indicated by sample_x_jsjsj will be ignored.
+            green_list = ["hasCalculatedProperty", "wasCalculatedBy", "hasValue"]
             ssplit = string3.split('.')
             if (len(ssplit) == 2):
-                if (ssplit[0] == 'cmso') and (ssplit[1] != "hasCalculatedProperty"):
+                if (ssplit[0] == 'cmso') and (ssplit[1] not in green_list):
                     plot = False
             if string3 == 'subClassOf':
                 plot = False