When deleting atoms, and inserting substitutions and interstitials, atomRDF currently sets the predicate as hasSymbol. It should be hasChemicalSymbol instead.

Hi,

I am trying to delete an atom in the lattice and writing it to a turtle file. The final hasNumberOfAtoms recorded in the turtle file is incorrect.

Steps to reproduce

from pyscal_rdf import KnowledgeGraph, System
kg = KnowledgeGraph()

# Create Al lattice
struct_Al = System.create.element.Al(graph=kg)

# Write initial data to ttl file
kg.write("initial.ttl", format="ttl")

# Delete one atom
struct_Al.delete(indices=[0])

# Write final data to file
kg.write("final.ttl", format="ttl")

Version

pyscal3: 3.1.7
pyscal-rdf: 0.2.3

Issue

The hasNumberOfAtoms field in initial.ttl is 4, which is correct. However, the hasNumberOfAtoms field in final.ttl is 2, which seems incorrect.

Can you kindly look into this?

Thanks

I am retrieving system object using get_system_from_sample, I expected system.graph to return a ref to the Knowledge graph. However it returns None.

Code to reproduce

from atomrdf import System, KnowledgeGraph
# Create data file
kg = KnowledgeGraph()
struct_Mg =  System.create.element.Mg(graph=kg, repetitions=(10,10,10))

kg.write("Mg.ttl", format="ttl")


# Read file
kg_2 = KnowledgeGraph(graph_file="Mg.ttl")
system = kg_2.get_system_from_sample(kg_2.samples[0])

print(system.graph)

Versions
atomrdf: 0.4.4
pyscal3: 3.1.12

when calling the query_sample() on the getting_started and grain_boundaries notebooks, one gets the following error:

---

UnboundLocalError Traceback (most recent call last)
Cell In[11], line 1
----> 1 res = g.query_sample(g.ontology.terms.cmso.hasAltName,
2 condition=(g.ontology.terms.cmso.hasAltName=='bcc'))
3 res

File ~/miniconda3/envs/rdf/lib/python3.11/site-packages/pyscal_rdf/graph.py:930, in RDFGraph.query_sample(self, destination, condition, return_query, enforce_types)
929 def query_sample(self, destination, condition=None, return_query=False, enforce_types=True):
--> 930 return self.auto_query(self.ontology.terms.cmso.AtomicScaleSample, destination,
931 condition=condition, return_query=return_query, enforce_types=enforce_types)

File ~/miniconda3/envs/rdf/lib/python3.11/site-packages/pyscal_rdf/graph.py:919, in RDFGraph.auto_query(self, source, destination, condition, return_query, enforce_types)
916 return res
917 else:
918 query = self.ontology.create_query(source, destination,
--> 919 condition=condition, enforce_types=val)
920 if return_query:
921 return query

UnboundLocalError: cannot access local variable 'val' where it is not associated with a value

• Ontology network module sometimes has owl#Thing as domain. This is currently ignored. Instead, this should be assigned to all classes.
• superproperties do not inherit domains from their subproperties. This is probably what is expected; nevertheless we should discuss if it makes sense for implementation purposes.

process_structure in RDFGraph allows giving a filename instead of a system object for the calculations. However, there is possibility to specify the format. These method should be ideally removed.

Features

• Add substitutional defects (#78 )
• Add interstitial defects (#79 )
• Wrapper for dislocations using atomman
• Reach test coverage
• Add ASE bulk as a direct option for structure creation
• Add API docs
• Add type hinting
• lint with black
• Add example notebooks

Creating a HCP Mg lattice with primitive=True and no repetitions shows a volume of "2993.63" (hasVolume field in turtle file). This is higher than expected as the cell lengths are (3.21,3.21,5.25).

Code to reproduce error:

kg = KnowledgeGraph()

struct_Mg = System.create.element.Mg(graph=kg, primitive=True)
kg.write("test_mg.ttl", format="ttl")

Versions
pyscal_rdf: 0.3.2

• DC3 variety of structures at different temperatures
• Carbon dataset ML potential fitting
• Copper dataset ML potential fitting
• Magnesium dataset ML potential fitting
• GB dataset GB, surface data for a variety of elements
• Si Va dataset
• Fe-H
• Al-Li
• GB dataset from MD
• https://doi.org/10.1038/sdata.2015.9
• https://edmond.mpg.de/dataset.xhtml?persistentId=doi:10.17617/3.DYLLSS

species should be known for StructureGraph methods. In case of create_structure, this is currently optional. For read_structure, when the format itself does not specify species information, for example lammps-dump, this will throw an error.

Query results can be returned in a pandas dataframe for further processing

Currently creating HCP lattice using primitive=False, leads to lattice angle 90,90,90. On discussion with @aazocar we think that lattice angle should still be 90,90,120. The cell angle however will still be 90,90,90 which is currently done correctly.

Code to reproduce:

kg = KnowledgeGraph()

struct_Mg = System.create.element.Mg(graph=kg,  repetitions=(10,10,10))
kg.write("test_mg.ttl", format="ttl")

Versions
pyscal_rdf : 0.3.2

When creating an l12 structure, it gives the following error:

TypeError: System.init() got an unexpected keyword argument 'species'

The versions of pyscal and pyscal_rdf in conda are:

pyscal-rdf 0.1.5
pyscal3 3.1.1