Adding a usage example.

[saeedizade74]

On the examples, you need a reference alignment file for evaluation. However this is just for using OntoAligner:

!pip install ontoaligner
and since it does not support Turtle, make it xml:

from rdflib import Graph
def rdf_to_xml(input_file, output_file):
    g = Graph()
    g.parse(input_file)
    g.serialize(destination=output_file, format="xml")

rdf_to_xml('O1.ttl', 'O1.xml')
rdf_to_xml('O2.ttl', 'O2.xml')

here is the alignment:

from ontoaligner.ontology import GenericOMDataset
from ontoaligner.encoder import ConceptParentLightweightEncoder
from ontoaligner.aligner import SimpleFuzzySMLightweight
from ontoaligner import AlignerPipeline
from ontoaligner.utils import xmlify  # for exporting results, optional

task = GenericOMDataset()
# task.track = "MyDomain"          # optional, just a label
# task.ontology_name = "OntoA-OntoB"  # optional, just a label

# 2. Collect — note: NO reference_matching_path passed at all
dataset = task.collect(
    source_ontology_path="/content/Logs4batch_id 1.xml",
    target_ontology_path="/content/Logs4batch_id 2.xml",
)


# 3. Encode + align (lightweight fuzzy matching, no LLM/GPU needed)
aligner_pipeline = AlignerPipeline(
    encoder=ConceptParentLightweightEncoder(),
    aligner=SimpleFuzzySMLightweight(fuzzy_sm_threshold=0.9),
    om_dataset=dataset,
)

matchings = aligner_pipeline.generate()

# 4. No evaluation call since there's no reference — just inspect/save matchings
rows = [
    (
        row["source"].split("/")[-1],
        row["target"].split("/")[-1],
        f"{row['score']:.3f}",
    )
    for row in sorted(matchings, key=lambda r: r["score"], reverse=True)
]

src_width = max(len(r[0]) for r in rows)
tgt_width = max(len(r[1]) for r in rows)

print(f'{"Source":<{src_width}}  {"Target":<{tgt_width}}  Score')
print(f'{"-" * src_width}  {"-" * tgt_width}  -----')

for source, target, score in rows:
    print(f"{source:<{src_width}}  {target:<{tgt_width}}  {score:>5}")

[saeedizade74]

Here is an example that the sample code does not work
https://colab.research.google.com/drive/1BYeU05ST8vR93EshtoroGA6s-dQJsBDv?usp=sharing