Track file description information

[shirey]

For primary datasets file type information can be derived from the ingest validation directory schemas

For processed datasets file information by assay/directory schema is included in, for example for salmon rnaseq processed: bulk_pipeline_manifest.json file or for SPRM processed data pipeline_manifest.json

How do we resolve the manifest.json info files for various pipelines?

API/Use case/mock-up info

[AlanSimmons]

The proposal is that assay file type metadata be incorporated into the HuBMAP ontology. Each assay (both primary and derived) would be associated with file paths to the relevant manifest files (YAML for primary; JSON for derived).

The File API feature would thus depend on the HuBMAP application ontology.

To obtain file information on an assay type, a File API would:

• query the HuBMAP ontology knowledge graph (via a Cypher query against the ontology neo4j database) to find the file nodes that relate to the assay type--e.g., the DirectorySchemaURL node.
• Open the directory schema file (YAML/JSON).
• Obtain file type information (e.g., description of files) from the schema file.

[AlanSimmons]

@kburke @shirey This is done, no?

[shirey]

@AlanSimmons Not fully tested or released yet.