gsoc26: Mapping Conversion Layer (Layer 3)

[DhanashreePetare]

Feature Request

Describe the feature you'd like
Add Layer 3 (cross-class mapping conversion) to the Databus Python Client download pipeline. Users should be able to convert downloaded RDF files to CSV and vice versa, and convert between RDF triple and quad formats, using the existing --convert-format flag extended with --graph-name and --base-uri flags.

Why is this feature important?
Many applications consume tabular data (CSV/TSV) but datasets on the Databus are published in RDF formats. Cross-class conversion eliminates the manual mapping step and makes Databus datasets directly consumable by tabular applications. It also enables named graph workflows by supporting triples-to-quads promotion and quads-to-triples splitting.

Describe alternatives you've considered
The Java client uses TARQL for CSV to RDF mapping. For the Python client, a direct rdflib-based approach is used instead to avoid adding a new dependency. The companion metadata JSON file approach is used to address data loss of RDF datatypes and language tags during RDF to CSV conversion, enabling lossless round trips when both files are used together.

Additional context
Supported mapping directions:

From	To	Notes
RDF Triples	RDF Quads	Requires --graph-name. Lossless.
RDF Quads	RDF Triples	Splits into one file per named graph. Lossless.
RDF Triples	CSV	Quasi-equal. Companion .meta.json generated.
CSV	RDF Triples	Lossless if companion file present. Requires --base-uri.

Out of scope for this issue:

• RDF Quads to CSV (stretch goal, significant added complexity)
• TARQL integration