Bases: BaseProcessor

Processor for CellO automated cell type annotations.

CellO provides automated cell type predictions using the Cell Ontology (CL) for bulk RNA-seq samples. The raw source is a JSON object keyed by SRX accession with a list of CL term IDs as the value for each sample.

The processor explodes the per-sample term lists into one row per (sample, CL term), resolves each CL ID to a human-readable name via the CL OBO file, and drops any terms that cannot be resolved.

process(output_dir, **kwargs)

Process the CellO JSON file into standardized sample annotations.

Reads the JSON source, builds one row per (sample, CL term), maps each CL ID to its human-readable label via the CL OBO, drops unmapped terms, and writes the result to a parquet file.

Parameters:
  • output_dir (Path) –

    Directory where the processed parquet file will be written.

  • **kwargs (Any, default: {} ) –

    input_path (Path | str) — override the default CellO JSON input path (defaults to CELLO_JSON from config).
Returns:
  • DataFrame –

    Standardized annotations with columns COL_ACCESSION, annotation_type, term_id, term_label, and ecode.

validate(data)

Validate that processed CellO data meets minimum requirements.

Parameters:
  • data (DataFrame) –

    Processed annotations DataFrame to validate.
Returns:
  • bool –

    True if validation passes.
Raises: