The mapper implements the stOTTR language with extensions for mapping asset structures inspired by the Epsilon Transformation Language. Implemented with Apache Arrow in Rust using Pola.rs. We provide a Python wrapper for the library, which allows us to create mappings using DataFrames.

The API is simple, and contains only one class and a few methods.

from maplib import Mapping, ResolveIRI, MintingOptions, to_graph
import polars as pl

Mapping-objects are initialized using a list of stOttr documents (each a string).

doc = """
    @prefix ex:<http://example.net/ns#>.
    ex:ExampleTemplate [?MyValue] :: {
    ottr:Triple(ex:myObject, ex:hasValue, ?MyValue)
    } .
    """
mapping = Mapping([doc])

In order to extend this template, we provide a DataFrame with a particular signature. We use the extend-method of the mapping-object.

from typing import Optional, Dict
def extend(self, 
           template: str,
           df: pl.DataFrame,
           language_tags: Optional[Dict[String, String]]=None
           ) 

The template-argument specifies the template we want to expand. We are allowed to use the prefixes in the stOttr documents when referring to these templates unless there are conflicting prefix-definitions. The parameters of the templates must be provided as identically-named columns. To provide a null-argument, just make a column of nulls.

Multiple alternatives exist to export the mapped triples. The fastest way to serialize is the write_ntriples-method.

mapping.write_ntriples(file:str)

Alternatively, we can export the mapping to an rdflib-graph.

gr = to_graph(mapping)