Version 0.10 • pipette

pipette 0.10.0 (2022-10-25)

Major changes:

This release attempts to harden the primary arguments supported for export and import. Note that file is intentionally not supported in favor of con, which is the preferred convention used in BiocIO.
Reduced the number of exported methods, tightening on missing instead of using missingOrNULL, which can cause unwanted inheritance issues that collide with BiocIO package.
export now supports any type of atomic vector.

Major changes:

import: Added support for import of gene cluster text (GCT) file format. Currently can set return parameter to import as a matrix (default) or data.frame with Name and Description columns retained.

Minor changes:

factorize: Bug fix to ensure that logical columns are not coerced to factor, as these are expected to have repeated values.
Now requiring Bioconductor 3.16 release.

Minor changes:

Minor changes:

as.DataFrame: Reworked the list and SimpleList methods:
- We removed the option to set row.names for list and SimpleList methods. Instead, row names will get defined automatically when the first element in the list is named.
- The list method now checks for length mismatched input at the start of the function. This should also be handled in the downstream DataFrame generator step, but we can set a custom error message instead.
- We’re now calling new internally with listData and nrows defined, which is faster and less problematic then attempting to construct a matrix (via cbind) and then coerce to DataFrame instead.
- Our list method now attempts to reorder named elements when possible.
decode: The internal code for DataFrame method has been simplified, due to rework of our internal as.DataFrame method code. Note that list elements are now unname during the decoding step, but has no effect on row name handling.
droplevels2: Hardened DataFrame method to properly coerce DataFrame to SimpleList before proceeding. Using as method with "List" class does not currently unclass DataFrame as expected.

Minor changes:

Minor changes:

Added line break to naStrings handling, which improves automatic sanitization of some malformed gene CSV files on the Ensembl FTP server.

Minor changes:

import: Switched default quoting for delim file using base engine. Also removed empty string assert checks, which can be problematic for some data types returned by readr and data.table engines.

Minor changes:

import and export now support overriding default quote handling with the quote argument. Changing this is not recommended by default, but is useful for some edge cases with annoying gene metadata files from Ensembl. Who decided that B" is an acceptable gene name?

Minor changes:

Renamed S4 class definitions: DataFrame to DFrame; GenomicRanges to GRanges; and IntegerRanges to IRanges.

Major changes:

import: Improved support for FASTA files. The function will now intentionally error on any warnings, which can occur if the moleculeType argument is not set correctly. For example, we’re defaulting to DNA input here, but that is not always the case. miRBase FASTA files are in RNA format, so use moleculeType argument here to set "RNA" instead of "DNA". We’ve also added amino acid support via "AA" argument, which passes to Biostrings package in a similar method for DNA and RNA. Note that for miRBase FASTA files, we now attempt to get metadata from the FASTA identifiers, which are defined as a DataFrame slotted into metadata as attributes.

Minor changes:

import: Updated data.table engine settings to use header = "auto" when colnames argument is declared as TRUE. This will attempt to handle malformed columns, similar to readr/vroom approach.
Improved internal usage consistency of goalie boolean grepl wrapper functions, such as isMatchingFixed and isMatchingRegex, which improves code legibility.
Removed legacy Docker usage instructions in README.