glam/scripts/owl_to_plantuml.py

#!/usr/bin/env python3
"""
Convert OWL/Turtle ontology to PlantUML class diagram.

Usage:
    python owl_to_plantuml.py input.owl.ttl output.puml
"""

import sys
from pathlib import Path
from rdflib import Graph, Namespace, RDF, RDFS, OWL
from collections import defaultdict

# Namespaces
LINKML = Namespace("https://w3id.org/linkml/")
HC = Namespace("https://nde.nl/ontology/hc/")
SCHEMA = Namespace("https://schema.org/")
SKOS = Namespace("http://www.w3.org/2004/02/skos/core#")

def simplify_uri(uri, graph):
    """Convert URI to short form using prefixes."""
    try:
        # Try to get qname (prefix:localname)
        qname = graph.qname(uri)
        if ':' in qname:
            prefix, local = qname.split(':', 1)
            # For LinkML generated URIs, just use the local name
            if prefix in ['hc_class', 'hc_slot', 'hc_enum', 'hc']:
                return local
            return qname
        return qname
    except Exception:
        # Fallback to last part of URI
        return str(uri).split('/')[-1].split('#')[-1]

def get_label(graph, uri):
    """Get human-readable label for a URI."""
    # Try rdfs:label first
    label = graph.value(uri, RDFS.label)
    if label:
        return str(label)
    
    # Try skos:prefLabel
    label = graph.value(uri, SKOS.prefLabel)
    if label:
        return str(label)
    
    # Fallback to simplified URI
    return simplify_uri(uri, graph)

def parse_owl_to_plantuml(owl_file, output_file):
    """Parse OWL/Turtle and generate PlantUML diagram."""
    
    print(f"Loading OWL file: {owl_file}")
    graph = Graph()
    graph.parse(owl_file, format='turtle')
    
    print(f"Loaded {len(graph)} triples")
    
    # Find all classes
    classes = set()
    for s in graph.subjects(RDF.type, OWL.Class):
        classes.add(s)
    for s in graph.subjects(RDF.type, RDFS.Class):
        classes.add(s)
    
    # Filter to keep only HC namespace classes (our ontology)
    hc_classes = [c for c in classes if str(c).startswith(str(HC))]
    
    print(f"Found {len(hc_classes)} HC classes")
    
    # Find properties
    properties = defaultdict(list)  # class -> [(property, range, cardinality)]
    
    for cls in hc_classes:
        # Find properties via rdfs:domain
        for prop in graph.subjects(RDFS.domain, cls):
            prop_range = graph.value(prop, RDFS.range)
            properties[cls].append({
                'uri': prop,
                'label': get_label(graph, prop),
                'range': get_label(graph, prop_range) if prop_range else 'string',
                'type': 'property'
            })
    
    # Find inheritance relationships
    subclass_of = {}  # child -> parent
    for cls in hc_classes:
        parent = graph.value(cls, RDFS.subClassOf)
        if parent and parent in hc_classes:
            subclass_of[cls] = parent
    
    # Generate PlantUML
    lines = [
        "@startuml",
        "!theme plain",
        "skinparam classAttributeIconSize 0",
        "skinparam linetype ortho",
        "skinparam class {",
        "  BackgroundColor white",
        "  BorderColor black",
        "  ArrowColor black",
        "}",
        "",
        "' Heritage Custodian Ontology - Class Diagram",
        f"' Generated from: {Path(owl_file).name}",
        "",
    ]
    
    # Sort classes by name for consistent output
    sorted_classes = sorted(hc_classes, key=lambda c: get_label(graph, c))
    
    # Output classes
    for cls in sorted_classes:
        cls_label = get_label(graph, cls)
        
        # Get class description
        description = graph.value(cls, RDFS.comment)
        
        lines.append(f"class {cls_label} {{")
        
        # Add properties
        if cls in properties:
            for prop in sorted(properties[cls], key=lambda p: p['label']):
                prop_name = prop['label']
                prop_range = prop['range']
                lines.append(f"  {prop_name}: {prop_range}")
        
        lines.append("}")
        
        # Add note if there's a description
        if description:
            desc_short = str(description).replace('\n', '\\n')[:150]
            if len(str(description)) > 150:
                desc_short += "..."
            lines.append(f"note right of {cls_label}")
            lines.append(f"  {desc_short}")
            lines.append("end note")
        
        lines.append("")
    
    # Output inheritance relationships
    lines.append("' Inheritance relationships")
    for child, parent in sorted(subclass_of.items(), key=lambda x: get_label(graph, x[0])):
        child_label = get_label(graph, child)
        parent_label = get_label(graph, parent)
        lines.append(f"{parent_label} <|-- {child_label}")
    
    lines.append("")
    lines.append("@enduml")
    
    # Write output
    print(f"Writing PlantUML to: {output_file}")
    with open(output_file, 'w') as f:
        f.write('\n'.join(lines))
    
    print(f"Generated PlantUML diagram with {len(sorted_classes)} classes")
    print(f"To render: plantuml {output_file}")

if __name__ == '__main__':
    if len(sys.argv) != 3:
        print("Usage: python owl_to_plantuml.py input.owl.ttl output.puml")
        sys.exit(1)
    
    owl_file = Path(sys.argv[1])
    output_file = Path(sys.argv[2])
    
    if not owl_file.exists():
        print(f"Error: Input file not found: {owl_file}")
        sys.exit(1)
    
    parse_owl_to_plantuml(owl_file, output_file)