dynamo-preprocess

What problem does it solve?

Preprocessing notebooks into reusable, source-grounded agent workflows by wrapping the current dynamo Preprocessor API, enabling deterministic pipelines and standardizable outputs across recipes.

Core Features & Use Cases

• Wrapper around dynamo.preprocessing.Preprocessor.preprocess_adata(...) to select a recipe (monocle, seurat, sctransform, pearson_residuals, or monocle_pearson_residuals) and to configure per-recipe parameters.
• Supports both a wrapper workflow for common cases and a stepwise path for debugging or custom pipelines.
• Validates essential outputs like adata.obs/var keys, obsm X_pca, and normalized layers for velocity-related analyses, enabling downstream dynamo analyses and UMAP visualization.
• Repository references and source-grounding notes are included for traceability, evaluation, and compatibility guidance.

Quick Start

Preprocess an AnnData object by choosing a recipe (monocle, seurat, sctransform, pearson_residuals, or monocle_pearson_residuals) and validating key outputs.