How ise-fizkarlsruhe/chemotion-kg works
Chemotion-KG is a chemical-research knowledge graph, with repository code focused on acquiring publication metadata and transforming RDF/JSON-LD into graph outputs. The visible workflow downloads publication JSON-LD into a year/month data store; separate processors merge records, create per-year graphs, or transform RDF into NFDI Core form, and another utility partitions an RDF graph by subject. These scripts are not wired into one end-to-end runner in the sampled evidence. The README describes FAIR graph data, but does not document a query service; the repository also includes a Turtle graph and query documentation.
Main entry points: worker_downloader.py Publication downloader · data Downloaded JSON-LD · main_parallel_downloader.py Parallel downloader · all.py Graph merger
Components 13
Metadata acquisition
- Parallel downloader
processing/main_parallel_downloader.py - Publication downloader
processing/worker_downloader.py - Downloaded JSON-LD
data
RDF processing
- Graph merger
processing/all.py - Yearly converter
processing/convert-all.py - NFDI Core transform
processing/all-nfdicore.py - Subject splitter
processing/split_by_subject.py
Graph outputs
- Merged RDF
- Yearly RDF
- RDF chunks
- Knowledge graph
data/chemotion-kg.ttl
Other components
- Research user
- Chemotion API
Connections 12
- Research user → Parallel downloader: starts download
- Parallel downloader → Publication downloader: launches subprocess
processing/main_parallel_downloader.py - Publication downloader → Chemotion API: fetches publications
processing/worker_downloader.py - Publication downloader → Downloaded JSON-LD: writes JSON-LD
processing/worker_downloader.py - Downloaded JSON-LD → Graph merger: reads JSON-LD
processing/all.py - Graph merger → Merged RDF: serializes merged graph
processing/all.py - Downloaded JSON-LD → Yearly converter: reads JSON-LD
processing/convert-all.py - Yearly converter → Yearly RDF: writes yearly graphs
processing/convert-all.py - Research user → NFDI Core transform: provides RDF input
- NFDI Core transform → Research user: returns transformed graph
- Research user → Subject splitter: provides RDF input
processing/split_by_subject.py - Subject splitter → RDF chunks: writes chunks
processing/split_by_subject.py