Skip to contents

Changelog

Source: NEWS.md

yspec 0.8.0

  • ys_document() no longer prints unit in define.pdf when it is set to an empty string (#173).

  • ys_add_factors() and ys_factors() now retain attributes on data columns that are converted to factors (#177).

  • ys_extend() and ys_join() now properly copy namespace information from the extension spec to the parent, so namespaces defined only in an extension spec are recognized without error (#179).

  • Added ys_document_namespace as a SETUP__ meta field to control which namespaces are invoked when rendering define.pdf; falls back to the tex namespace by default (#180).

  • Fixed a bug in the regulatory define.pdf where the file extension was incorrect in section headers (#181).

yspec 0.7.0

  • Added utilities to work with flags in the spec (#166), including
    • ys_flags() to extract a named list of flag data
    • ys_flags_chr() to extract a character vector of flagged data names
    • ys_select_fl() to select from a yspec object based on flagged data names
    • ys_factors_fl() to create factors in a data frame based on flagged data names
  • ys_add_labels() gains a strict argument, allowing the user to label any data frame if there is at least one name in common between the spec and the data frame (#167).

yspec 0.6.3

  • Links in “regulatory” define pdf file generated by ys_document() now appear in default color (black) rather than previous magenta (#164).

  • Updates for compatibility with glue 1.8.0 (#162).

yspec 0.6.2

Breaking changes

  • nm_input() was significantly refactored to produce output more typically seen in practice as well as adding some convenience features (#150).
    • the default output is now wide format
    • data set labels can get renamed with <new name> = <old name> syntax, passed under ...
    • columns with type <character> are dropped by default
    • the cat argument was renamed .cat
    • the .drop argument was added to indicate a group of data items to drop from the problem
    • the previous long output can be recreated using .long = TRUE
    • the .decodes argument was added to control appearance of categorical data decode information in the long output

yspec 0.6.1

yspec 0.6.0

  • Adds ys_factors() to replace data set columns with their factor version, optionally retaining the original values in new columns (#141).

yspec 0.5.3

  • Fix bug where character type was inferred whenever values were character, overriding what the user specified; now, only infer when user doesn’t specify type (#129).

  • Add ys_col_note() to generate a string of column names and definitions based on short or label; use this when column names appear in a table and require explanation in a a footnote (#119, #120).

  • Put stories in yaml format; add script to build validation docs from the yaml file (#123, #124).

yspec 0.5.2

  • Adds ys_extend() to extend a spec object with additional columns with any external yaml file; adds extend_file field in SETUP__: to identify a single file which can be automatically called upon for extension #112, #113.

  • ys_add_factors() refactored to only add factors for available columns #88, #91.

  • ys_add_factors() restores tidyselect semantics when specifying a subset of columns for adding factors #109, #110.

  • Add ys_table() to render the spec in TeX table code to be included in a report appendix #98, #104.

  • ys_select() returns an empty spec object when nothing is selected #100, #101.

  • Add ys_prune() to select any data columns that appear in the spec as well as other optional columns #87, #97. #99.

  • Add ys_recode() to convert a vector of data set column names to short version, optionally adding unit #86, #114.

  • ys_load() will now detect if values / decode are mis-coded as a list of lists; all items in values and decode must be atomic #106, #108.

  • ys_load() will issue more informative error message when short contains too many characters #102, 105.

  • ys_document() now accepts sponsor and projectnumber arguments which get passed through to the rendered define document #90, #92.

yspec 0.5.1

  • Fix bug where dots in lookup file were not properly inherited by spec columns that didn’t already have a dots list (#69)

  • Fix bug in namespace creation when multiple fields include namespace info (#81)

  • Fix bug where specification document didn’t properly render when using R 4.1 (#73)

  • ys_add_factors() now substitutes missing values (NA) with a character string supplied by the user prior to making the factor; the decode list is also updated so that the “missing” data is the last level; the string is supplied by the new argument .missing argument (#79)

yspec 0.5.0

  • ys_select() will now rename columns if new names are provided (#62)
  • ys_filter() will subset a yspec object based on logical criteria (#65)
  • ys_rename() will rename columns in a yspec object (#63)
  • ys_join() will combine two or more yspec objects together (#64)
  • New SETUP__ field called flags that can be used to flag sets of columns (#62)
  • ys_fill_dots() will supplement the dots list with default values (#62)
  • Add vignettes back to the package

yspec 0.4.2

  • Add ys_select() to select a subset of columns from a yspec object #24
  • Refactor axis_col_labs to add vars argument; if missing, all col / labs will be created; otherwise, the user can select a subset to be created
  • Deprecate mrgtemplate(); added definetemplate() that is used as default for a more styled define document #37
  • Add ability to include alternate representations of field information through namespacing #38
  • Fix bug in ys_check where the return value was not correct when error_on_fail was set to FALSE #31
  • Add nm_input function to print nonmem input data text to the console #8
  • Fix bug where some meta data (data_stem, name, spec_file) was not updated when importing from another spec #5
  • Fix bug where ys_get_short_unit was not returning a list #2
  • Fix bug where axis_col_labs doesn’t do title case or short max #26
  • Better reporting of bad values in discrete data column #27

yspec 0.4.1

  • Add minimum version numbers for dependency packages
  • Add long argument to pander_table to create longer tables
  • Fix bug where data_stem meta field was not getting propagated into data set name in the define document #35
  • Add ability to update project number and sponsor information in a yspec object or a yproj object #36
  • Fix bug where unit input was not properly validated #45
  • Fix bug where ys_check didn’t issue error when data set columns were not in the right order #54
  • New field for data set columns: label; this will be used when labeling data sets for xpt export #60
  • The label field can be made optional with ys.require.label option #99
  • Add function (ys_add_labels) to label data set columns #63
  • Add function (update_short) to change the value in the short field #82
  • Add meta field (import) to enable import of contents of another spec #84
  • Add meta field option (character_last) to push all character columns to the back of the spec object #86
  • Add meta field option (comment_col) to identify the column that is for comments; character comment columns will not be pushed to back #86
  • Add argument to ys_check so that an error is not generated when checking the data, but rather FALSE is returned #88
  • Add functionality to add all possible factors to data set with yspec_add_factors #91
  • Add new data column field make_factor that will tell yspec_add_factors to make that column into a factor #91
  • The following fields will be collapsed into a single value when they are entered as array: source, comment, long #93
  • Add alias ys_add_factors to yspec_add_factors #97
  • Add functions to form label, unit, and short; methods for yspec and ycol #100

yspec 0.3.0

  • Released into production

yspec 0.2.0.9000

  • Fixed bug where column check error were assigned to the wrong column name #7
  • Added as_proj_spec function that allows creation of a project spec file from individual data set spec objects
  • Added more flexibility for specifying project and data set spec files in different locations #5
  • Function ys_load preferred alias for load_spec
  • Function ys_load_file preferred alias for load_spec_file
  • Function ys_project preferred alias for as_proj_spec
  • Function ys_document is a wrapper around both render_define and render_fda_define. Use the type argument to request a “working” document or a “regulatory” document
  • Added internal lookup source (ysdb_internal.yml)
  • It is no longer an error to put no data in a column specification; the lookup attribute will be added and yspec will attempt to look up the column in formation in one of the lookup tables
  • ys_get_lookup will recreate the lookup list for a yspec object
  • ys_lookup_source will report on the source of each column in an yspec object
  • In general, we no longer write out a project file; use ys_project with a vector of files or yspec objects instead
  • The table of contents for the regulatory document (fda_content_table) now links to the corresponding .xpt file; to facilitate testing, arguments loc and ext have been added so that the linking can be better tested (e.g. create a document that links to a .csv file, which should be easier to open on computers that don’t have SAS installed). #20
  • Arguments data_path and data_stem have been added to ys_load
  • Handler !look has been added to indicate that column data should be searched for in the lookup resources
  • Handler !value:decode has been added so that maps can be created with the values on the left hand side (- 0 : male)
  • Handler !decode:value has been added that recreates the default map handling, but also allows several values to be mapped to the same decode
  • An error is generated if the number of characters in a column name is greater than 8. Set the ys.col.len option to control the max allowable characters.
  • Added verbose argument to print information to the console while processing a yaml file
  • Added ys_help object that includes example / demo materials, including a function that will export a folder of assets that will work out of the box
  • Revised package vignettes
  • Expanded the ?yspec help topic to outline the suggested workflow and point to other locations for help
  • Added custom sanitizer ys_sanitize that can be changed through the ys.sanitize option
  • Added a glue field for SETUP__, where a yaml map can be used to connect value with a name; when the name is inserted in to yaml code as <<name>>, the value will be substituted in with glue; this is the proper way to evade the sanitizer when including TeX code in the document
  • Only “numeric” or “character” will be accepted in the type field; and error is generated if anything else appears there
  • Added c method to combine two yspec objects together

yspec 0.1.0.9001

  • render_define changed to generic function, with methods for yproj and `character
  • Removed projectnumber and sponsor arguments to render_define; that information will be pulled from the SETUP__ block as in render_fda_define
  • The column widths for x_table were made smaller so that they fit on our standard page (8.5 x 11 with 7 cm margins on the left and the right)

yspec 0.1.0

  • Initial validated version