scxpand.pretrained.inference_api

scxpand.pretrained.inference_api#

Internal inference API for pre-trained models.

This module provides internal functions for running inference with pre-trained models, handling automatic model downloading and loading. For external use, use scxpand.run_inference().

Functions

fetch_model_and_run_inference([model_name, ...])

Internal function for running inference with pre-trained models.
scxpand.pretrained.inference_api.fetch_model_and_run_inference(model_name=None, model_url=None, data_path=None, adata=None, save_path=None, batch_size=None, num_workers=4, eval_row_inds=None)#

Internal function for running inference with pre-trained models.

This is an internal function that handles automatic model downloading, loading, and inference in a single call. Works with both file-based and in-memory data. For external use, use scxpand.run_inference() instead.

Parameters:

  • model_name (str | None (default: None)) – Name of pre-trained model from registry (alternative to model_url)

  • model_url (str | None (default: None)) – Direct URL to model ZIP file (alternative to model_name)

  • data_path (str | Path | None (default: None)) – Path to input data file (h5ad format) - alternative to adata

  • adata (AnnData | None (default: None)) – In-memory AnnData object - alternative to data_path

  • save_path (str | Path | None (default: None)) – Directory to save prediction results (optional)

  • batch_size (int | None (default: None)) – Batch size for inference (uses model default if None)

  • num_workers (int (default: 4)) – Number of workers for data loading

  • eval_row_inds (default: None) – Specific cell indices to evaluate (None for all cells)

Returns:

Structured results containing predictions, metrics (if available), and model info

Return type:

InferenceResults

Raises:

  • ValueError – If neither model_name nor model_doi provided, or neither data_path nor adata provided

  • FileNotFoundError – If data_path does not exist

Note

This is an internal function. For external use, use scxpand.run_inference() instead.

Contents