This organization contains various repositories for the ArtifactDB project, a storage system for analysis-ready artifacts. It aims to provide easy access to datasets and analysis results across multiple programming frameworks such as R and Python. The ArtifactDB system was originally developed at @Genentech to store the outputs of various genomics analysis pipelines (plus the associated metadata); scientists can then pull these artifacts into their analysis environments for further processing.
The alabaster suite implements functions to save and read Bioconductor objects via language-agnostic file formats. This is the workhorse of our R-based data serialization pipelines, managing the conversion of various objects into files for long-term storage.
The gypsum R package implements an interface to the gypsum REST API. This handles the upload of files and associated metadata to cloud storage for large-scale distribution.
The dolomite suite implements functions to save and read Bioconductor objects via language-agnostic file formats. This is equivalent to alabaster for Python and is based heavily on the classes from the BiocPy project.
The gypsum Python package implements an interface to the gypsum REST API. This is equivalent to the R package of the same name.
The gypsum worker implements a REST API for storing and serving artifacts via the Cloudflare stack. This uses R2 for storage and Workers to handle authenticated uploads via flexible permission schemes.
The Gobbler manages artifacts across users on a shared filesystem such as those used in HPC clusters. This is effectively an on-premise version of gypsum that is simpler and more efficient for local applications.
The SewerRat provides a search index for metadata on a shared filesystem. This enables HPC users to find and share objects within an organization.
The takane library contains language-agnostic specifications for all Bioconductor object types. These are enforced by validator functions written in C++, which are used by both alabaster and dolomite to verify compliance.
The scRNAseq R package and its Python counterpart use gypsum to store takane representations of various single-cell datasets.
The celldex R package and its Python counterpart use gypsum to store takane representations of cell type reference datasets.
