Skip to contents

Version 0.16

pipette 0.16.0 (2026-05-31)

Major changes:

  • import: Added support for Apache Parquet files (.parquet). Uses nanoparquet by default; falls back to arrow when available.
  • import: Added support for DuckDB database files (.duckdb). Uses DBI and duckdb. Single-table databases are returned as a data.frame; multi-table databases are returned as a named list. A table or query parameter can be passed to select or filter data.
  • import: Added support for AnnData HDF5 files (.h5ad), the standard format for single-cell and spatial analyses from the scverse ecosystem (scanpy, squidpy, AnnData). Returns a SingleCellExperiment. Uses anndataR (native R, no Python required) by default; falls back to zellkonverter when specified via the engine argument.
  • import: Added support for Web Ontology Language files (.owl). Parses OBO Foundry-style RDF/XML using xml2. Returns a DataFrame with id, name, isA, def, and obsolete columns.
  • import: Enhanced OBO file parser to extract additional fields: namespace, def, is_a, synonym, and xref. A fields parameter controls which columns are returned.
  • import: Added transparent AWS S3 support. Pass s3://bucket/key URIs directly to import(). Requires the paws.storage package.
  • export: Added Parquet export for data.frame, DFrame, matrix, and GRanges objects. Use a .parquet file extension with export().
  • export: Added transparent AWS S3 support. Pass s3://bucket/key URIs as the con argument to export(). Requires the paws.storage package.

Minor changes:

  • Updated linter configuration (.lintr) to include unnecessary_lambda_linter() from lintr 3.1.
  • Added air.toml for the Air R formatter with matching 80-character line width and 4-space indent settings.

License changes:

  • Switched from AGPL-3.0 to Apache-2.0.

pipette 0.16.1 (2026-06-19)

Minor changes:

  • Modernized tooling: removed per-package .lintr in favour of global ~/.lintr; added air.toml; migrated to roxygen2 8.0.