Version 0.9 • pipette

pipette 0.9.0 (2022-05-04)

Major changes:

Switched primary import/export engine from readr back to base, to avoid strong dependency on readr package. Note that both readr and/or data.table packages can optionally be used for import/export by specifying "engine".
Removed coercion method support for as_tibble and as.data.table, to remove strong dependencies on data.table and tibble packages.
Removed as coercion support in favor of simply using as.DataFrame S4 generic approach, to avoid potential conflicts with Bioconductor.
as.DataFrame: Reduced the number of supported classes, removing support of non-Bioconductor classes, specifically data.table and tbl_df.
Removed re-exported functions: column_to_rownames, getURL, rbindlist, rownames_to_columns, and tibble.

Minor changes:

Package code is now formatting using styler package.
Now exporting method support in droplevels2 instead of droplevels, to avoid method conflict with new Bioconductor 3.15 update.
All requireNamespaces calls are now wrapped by assert.
Removed previously deprecated functions that are no longer in use. These were previously defined in deprecated.R: sanitizeColData, sanitizeRowData, sanitizeRowRanges, and writeCounts.

Minor changes:

import: Added support for Open Biomedical Ontologies (OBO) format. This uses ontologyIndex internally. See also BiocSet for alterative import method that uses es_set and other accessor functions.

Minor changes:

Minor changes:

atomize: Hardened edge cases of empty DataFrame and GRanges input. Also improved code coverage to test handling of these edge case events, which can occur when exporting metadata from SummarizedExperiment in upstream AcidExperiment and bcbioRNASeq packages for objects with minimal metadata.

Minor changes:

Updated and hardened unit tests to avoid file system lock issues on Windows related to readr engine.

Minor changes:

export: Bug fix to provide compatibility in Matrix method for export of sparse matrices without dimnames (e.g. row and/or column names). Hit this edge case in the update of example data in our Chromium package.

Major changes:

factorize: Reworked internal method to only coerce columns with repeated values to factor. Now uses anyDuplicated internally to check for which columns to factorize.

Minor changes:

import: Hardened the internal readr engine to use base make.names for column name repair for delimited file import (e.g. CSV, TSV). This also eliminates unwanted CLI messages about name repair when using readr.

Minor changes:

factorize: Improved handling of DataFrame input that contains complex S4 columns, such as CharacterList.