wobbegong

wobbegong converts Bioconductor objects (like BiocFrame, SummarizedExperiment, and SingleCellExperiment) into a set of static files optimized for HTTP range requests.

It includes a native Python client that allows you to query these datasets remotely, fetching only the specific genes or metadata columns you need without downloading the entire file.

[!NOTE]

Check out the R version of this package here.

Install

To get started, install the package from PyPI

pip install wobbegong

Quick Start

1. Convert Data (wobbegongify)

Use wobbegongify to convert your objects into static files.

from wobbegong import wobbegongify
from biocframe import BiocFrame
from singlecellexperiment import SingleCellExperiment
from scipy import sparse
import numpy as np

# 1. Create a SingleCellExperiment
counts = sparse.random(100, 20, density=0.1, format="csr")
pca = np.random.randn(20, 5)
sce = SingleCellExperiment(
    assays={"counts": counts},
    reduced_dims={"PCA": pca}
)

# 2. Convert to wobbegong format (defaults to zlib compression)
wobbegongify(sce, "output/my_study")

# OR use lz4 compression
wobbegongify(sce, "output/my_study", compression = "lz4")

2. Read Data (load)

Use wobbegong.load() to read data from a local path or a remote URL. The client automatically handles HTTP range requests for you.

import wobbegong

# Load from a local directory (or URL)
sce = wobbegong.load("output/my_study")

# Access Assays (Row-wise access is optimized)
counts = sce.get_assay("counts")
gene_expression = counts.get_row(0)

# Access Reduced Dimensions
pca = sce.get_reduced_dim("PCA")
pc1 = pca.get_column(0)

Note

This project has been set up using BiocSetup and PyScaffold.