NEWS.md

In acidgenomics/transformer: Coercion methods for biological data

transformer 0.2.12 (2010-01-08)

New functions

• as.DataFrame: New generic defined to allow for easy list to DataFrame coercion. Avoid defining the list method for as/coerce, as this will result in unexpected behavior for data.frame to DataFrame coercion. The standard DataFrame generator currently doesn't handle lists very well, unlike as_tibble, so we defined an alternative approach via as.DataFrame here in the package, to allow for easy list column generation.

Minor changes

• mutateAll: Further improvements for list column handling and DataFrame return consistency.
• Renamed coerceS4ToList in favor of coerceToList. The old function spelling is soft deprecated and will be deprecated/made defunct in a future release.
• Requiring 100% code coverage in CI checks, since this is an essential infrastructure package for Acid Genomics workflows.

transformer 0.2.11 (2020-01-07)

Minor changes

• mutateAll: Fix for improved handling of nested list columns. Requires that as_tibble is called internally before handing off to DataFrame.

transformer 0.2.10 (2019-11-06)

Minor changes

• Bumped Bioconductor dependencies to require 3.10 release.
• Updated unit test code to work with newly saved objects in acidtest update.

transformer 0.2.9 (2019-10-24)

New functions

• metadata2 and metadata2<-: Assign and extract metadata from either S3 or S4 objects in a consistent manner. Requires the which argument, similar to the approach employed in base attr. Working example shows clearly how this differs internally between S3 (attr) and S4 objects extending Annotated class (metadata).

Minor changes

• Made deprecated relevel methods (DataFrame, Ranges, SummarizedExperiment) defunct in favor of droplevels usage.

transformer 0.2.8 (2019-10-22)

Updated NAMESPACE to work with migration of some low-level functions into new acidbase package.

Minor changes

• matchRowNameColumn: Converted function to S4 method that works on data.frame and DataFrame class objects. Added choices argument, which lets the user change the column matching values, if desired.
• as.SummarizedExperiment: Redefined generic to include ... in case we need to add arguments in a future update.
• coerceS4ToList: Converted to S4 method approach. Now works specifically on objects that extend Annotated class, such as SummarizedExperiment.

transformer 0.2.7 (2019-09-09)

Minor changes

• Made flatFiles defunct, in favor of coerceS4ToList usage.

transformer 0.2.6 (2019-09-03)

Minor changes

• mutateAll: Bug fix that improves column creation with DataFrame.

transformer 0.2.5 (2019-08-26)

Major changes

• Renamed dplyr-like generics into camel case: innerJoin, leftJoin, rightJoin, fullJoin, semiJoin, antiJoin; mutateAll, mutateAt, mutateIf; selectIf.
• Reworked mutate* and select* verbs to use simple base R code internally, and now methods are only defined for DataFrame class. This way we don't run into any masking issues with dplyr, and we don't need to list it as a suggested package.

Minor changes

• Dropped support for select_all and select_all. These verbs don't really make sense in the context of Bioconductor-related internal code.

transformer 0.2.4 (2019-08-15)

New functions

Improved S4 method support for dplyr verbs working on DataFrame class.

• Joins: inner_join, left_join (already defined), right_join, full_join, semi_join, and anti_join. These methods use merge internally but ensure that rows are not unexpectedly reordered.
• Scoped variants of mutate and select: mutate_all, mutate_at, mutate_if, transmute_at, transmute_if; select_all, select_at, select_if. These variants keep track of row names. Internally, these functions coerce to tbl_df and then call the dplyr function. Return is coerced back to DataFrame, with row names preserved.

Major changes

• relevel: Made S3 methods for relevel defunct of favor of recommend S4 method usage of droplevels instead. S4 method support for droplevels is already provided in S4Vectors (i.e. List, DataFrame). Here we are improving droplevels functionality by adding additional method support for Ranges (to support IRanges, GRanges) as well as SummarizedExperiment, which enables automatic leveling of factors in colData or rowData.

transformer 0.2.3 (2019-08-13)

Minor changes

• Fix for automatic decoding of Rle columns, which will error otherwise when attempting to coerce from DataFrame to data.frame.

transformer 0.2.2 (2019-08-08)

Minor changes

• Reexporting dplyr left_join generic. Improved DataFrame method for left_join to keep row names, which correspond to the x object.

transformer 0.2.1 (2019-08-05)

Minor changes

• Improved documentation consistency, using new acidroxygen package.

transformer 0.2.0 (2019-07-19)

New methods

• relevel: Added S3 method support for DataFrame and GRanges objects containing factor columns.

Major changes

• Migraged atomize, decode, and encode from brio here.
• Removed as.DataFrame S3 method in favor of simply using S4 as coercion.
• Tightened up as.SummarizedExperiment S3 methods to support SummarizedExperiment and RangedSummarizedExperiment specifically.

Minor changes

• Improved S3 method documentation. Use rdname tag instead of method (or legacy s3method).

transformer 0.1.12 (2019-07-12)

Major changes

• Improved automatic row name handling for data.table and tbl_df when coercing to DataFrame.
• Added coercion support from IRanges to data.frame, data.table, and tbl_df, which works like the default GRanges coercion methods.
• Added new matchRowNameColumn function, which can detect row name columns inside data frame classes that don't support row names assignment, currently data.table and tbl_df.

transformer 0.1.11 (2019-05-04)

Minor changes

• Importing GRanges class from GenomicRanges package.
• Improved Travis CI and AppVeyor configuration.

transformer 0.1.10 (2019-04-25)

Minor changes

• S4 generic reexport documentation fixes.

transformer 0.1.9 (2019-04-22)

Minor changes

• Initial DataFrame S4 method support for left_join.

transformer 0.1.8 (2019-03-28)

• Updated working examples and unit tests to use acidtest package.

transformer 0.1.7 (2019-03-22)

Minor changes

• Migrated code to [Acid Genomics].

transformer 0.1.6 (2019-02-25)

Minor changes

• Improve documentation regarding default row name handling for as_tibble and as.data.table.

transformer 0.1.5 (2019-02-18)

Major changes

• Added support for coercion to data.table class. DataFrame and GRanges class objects are initially supported.

transformer 0.1.4 (2019-01-23)

Minor changes

• Fix for DataFrame to tbl_df coercion that doesn't make names syntactically valid during internal as.data.frame() call.

transformer 0.1.3 (2019-01-22)

Major changes

• Migrated these functions to brio package: atomize(), factorize(), encode(), decode().

transformer 0.1.2 (2019-01-20)

Major changes

• Migrated decode() and encode() S4 methods from basejump.
• Improved atomize() compatibility with Bioconductor 3.7 release. Now DataFrame objects containing Rle columns will get decoded consistently without applying stringsAsFactors in an unexpected fashion. This works correctly in Bioconductor 3.8 simply using as.data.frame() internally, but has unwanted coercion of character to factor columns for Rle-encoded data when using the Bioconductor 3.7 release.

Minor changes

• Improved code coverage to 100%.

transformer 0.1.1 (2019-01-14)

Major changes

• Renamed package from S4Transformer to simply transformer.
• Migrated atomize() and factorize() from basejump.

Minor changes

• Ensuring tibble::as_tibble() gets reexported.
• Now exporting as.DataFrame() S3 coercion methods, for consistency.
• Improved documentation for as() coercion methods.
• Improved Travis CI and [AppVeyor][] CI build settings.

transformer 0.1.0 (2019-01-06)

Initial release. Migrated useful S4 coercion methods from basejump for easier unit testing in a separate package.

acidgenomics/transformer documentation built on Jan. 9, 2020, 11:34 a.m.