Changes in version 0.2.7 - Remove automatic sorting introduced in newer versions of tidytable Changes in version 0.2.6 - Switch default workflow executor to targets - Update vignettes to demonstrate current workflow procedure (@hgao1, #59) - Remove dplyr dependency Changes in version 0.2.5 - Add capability to embed custom messages to check results, using check.errors attribute in returned logical value - Refactor side-effect messages from built-in checks to check.errors Changes in version 0.2.4 - Fix for issue where labelling large datasets would take a very long time Changes in version 0.2.3 - Add ability to filter variable lineage with blueprintr::filter_variable_lineage() - Add tooltop for variable node in variable lineage - Fix issue where grouped nodes in visNetwork would scale, harming legibility - Convert empty folder warning to option. Enable with options(blueprintr.warn_empty_blueprints_dirs = TRUE) Changes in version 0.2.2 - Adds .SOURCE("table_name") and mark_source(obj) to add variable (and table) lineage to non-blueprint-created tables. Particularly useful for including data fetched from the web into lineage - Fixes lineage legends to focus on object type, rather than grouping. In large projects, the table groups would make the legend explode with the number of labels, rendering it useless. Changes in version 0.2.1 - Implements variable lineage through options(blueprintr.use_variable_uuids = TRUE). Visualize variable lineage with blueprintr::vis_variable_lineage or inspect specific parts of the graph with an igraph generated with blueprintr::load_variable_lineage. - Visualize blueprint table lineage with blueprintr::vis_table_lineage and inspect specific parts of the graph with blueprintr::load_table_lineage. - Implements improved annotations through an option blueprintr.use_improved_annotations. Metadata now always overwrites annotations in the output dataset, but mutate_annotation() and mutate_annotation_across() will override the metadata safely. No longer a need to mess with annotate_mutate! Changes in version 0.2.0 - Adds ability to run macro statements in interactive mode. Very useful when debugging pipelines! Set the blueprintr.interactive_eval_macros option to TRUE to enable. - Adds ability to set metadata_file_path based on the directory and name of the affiliated blueprint script. For example, a blueprint defined in "blueprints/ex/test.R" will get its default metadata file location set to "blueprints/ex/test.csv". Set the blueprintr.use_local_metadata_path option to TRUE to enable. - Fixes a bug where targets patterns were being evaluated prematurely Changes in version 0.1.3 - Fixes an issue where qualified expressions like "package::thing" were not being translated correctly Changes in version 0.1.2 - Fixes double-sided formulae rendering in blueprints - Exports bpstep() and bp_add_step() for customization - Allows argument forwarding to kfa::kfa() using the kfa_args parameter in bp_export_kfa_report() Changes in version 0.1.1 - Clarifies which coding string cannot be evaluated during codebook generation or labelling - Fixes a bug where bp_extend() would clobber the results of bp_add_bpstep() Changes in version 0.1.0 - Adds capability to define custom pipeline steps, executed after completion of the _final stage - Moves codebook export step to new custom step - Adds kfa report generation step capability Changes in version 0.0.11 - Uses UTF-8 encoding when parsing blueprint files (#41) - Ignores private (dot-prefixed) metadata fields during variable annotation (#40) Changes in version 0.0.10 - Adds a .with_names parameter to mutate_annotation_across that, when TRUE, sends a column and its name as arguments to .fn. Changes in version 0.0.9 - Bugfix to ensure the "description" field is present in the event that parent, annotated datasets don't have the "description" field filled in at all - Bugfix that didn't specify a join column, causing nuisance noise Changes in version 0.0.8 - Adds "annotations", a variable attribute system that makes it easier to propagate metadata provenance, rather than depending on variable naming from .TARGET() calls. - Adds mutate_annotation() and mutate_annotation_across() which allows runtime transformation of metadata. Useful for batch transformations before metadata creation. Changes in version 0.0.7 - Adds recurse parameter to load_blueprints() and tar_blueprints() to recursively load blueprints from the provided directory Changes in version 0.0.6 - Integration with panelcleaner to import homogenized codings if target data.frame is a mapped_df - Adds a new bp_include_panelcleaner_meta() extension to signal blueprintr to add mapped_df meta Changes in version 0.0.5 - Fixed bug that would not translate anonymous functions and formulae correctly (#22 and #29) Changes in version 0.0.4 - Add targets support, with tar_blueprint() and tar_blueprints() - Change create_metadata_file() to return the metadata file path. To compensate, a new build step *_meta_path has been added that removes the metadata logic inconsistency. Target topology should be much cleaner and clearer. - Added capability to group variables together for shared descriptions in codebooks with group field - Add file parameter for bp_export_codebook() to access codebook_file parameter in blueprint Changes in version 0.0.3 - Added blueprint extension functions bp_export_codebook() and bp_label_variables() that create a pipe-able method of setting blueprint configuration. Changes in version 0.0.2 - Created codebook export feature with render_codebook() (#3) - Allow labelled dataset generation with blueprint(..., labelled = TRUE). Handy to have when making datasets intended to be used in a STATA or SPSS environment. - Load blueprints from file or folder using load_blueprint() or load_blueprints() (#10)