This release attempts to solve problems hosting the shiny_cast app, while also implementing functions to preserve as much meta data as possible from the REDCap database when exporting data.
The hosting on shinyapps.io has given a lot of trouble recently.
Modified package structure a little around the
shiny_cast(), to accommodate an alternative hosting
approach with all package functions included in a script instead of
requiring the package.
NEW: A new option to raw_or_label in
read_readcap_tables() has been added: “both”. Get raw
values with REDCap labels applied as labels. Use
as_factor() to format factors with original labels and use
the gtsummary package to easily get beautiful tables with
original labels from REDCap. Use fct_drop() to drop empty
levels.
NEW: fct_drop() has been added with an extension to
forcats::fct_drop(), that works across data.frames. Use as
fct_drop().
CHANGE: the default data export method of
easy_redcap() has been changed to use the new labelled data
export with read_readcap_tables().
24.11.1 was rejected on CRAN based on wrong title capitalisation. This was an opportunity to extend the package overhaul. And this actually turned out to be a major step towards a very usable shiny app which have received most of the focus.
I have implemented option to specify categorical variables to
factorize, but doing this with a modified version of {forcats} and
{haven}’s as_factor(), that will preserve any attributes
applied to the data to be able to upload and cast REDCap meta data from
richly formatted data (use .rds). No matter the input type, all input is
parsed using the default options from the {readr} package. Also to avoid
mis-labelling, logicals are converted to factors as REDCap truefalse
class follows different naming conversion compared to R. Also correct
support for variable labels as field labels (use .rds formatted data and
label with labelled::var_label())
Vignettes and documentation have been restructured.
This package has been detached from the REDCapRITS, which it was originally forked from. The data split function will be kept, while testing will be rewritten. This projects has evolved away from the original fork.
Revised tests.
Documentation has been slightly updated to highlight the shiny app for casting REDCap metadata. I am working on hosting my own Shiny Server.
Bug: ‘form.name’ specified to ‘ds2dd_detailed()’ was ignored.
Corrected to only be ignored if ‘form.sep’ is specified. Added handling
of re-occurring form.sep pattern.
New: export_redcap_instrument() is a new version of
create_instrument_meta(), that will only export a single
instrument. Multiple instrument export can be done with
lapply() or purrr::map(). This allows for
inclusion of this functionality in the Shiny implementation and is
easier to handle. create_instrument_meta() is
deprecated.
Improved: shiny_cast() app has been updated to
actually work if you install the package and not clones the whole
repository.
New: Major overhaul of the app interface with the introduction of
bslib for building the page. Also Detailed documentation
added for the app workflow.
New: Export a REDCap instrument ready to add to your database
based on an uploaded spreadsheet. This is thanks to the
export_redcap_instrument() function. This functionality is
intended for projects in production and adding instruments should be
handled manually and not by API upload.
Bug: Export datadictionary with “” instead of “NA” for NAs. Upload to REDCap failed. Not anymore.
The shiny implementation is included with this package. Implementing in shinylive may be looked into again later.
Updated links and spelling.
Minor changes to pass tests and renv is out. rhub is
really not running as smooth as previously.
Fix: read_redcap_tables(): field names testing
allows to include “[form_name]_complete” fields.
Fix: ds2dd_detailed(): default record ID name is now
“record_id”, the REDCap default. Default is still to use the first
column name. Support was added to interpret column name prefix or suffix
as instrument names. See the examples.
New: create_instrument_meta(): creates zip with
instrument files to allow adding new instruments to project in
production. Takes data dictionary as input and creates a zip for each
instrument specified by the form_name column.
New: doc2dd(): function to convert document table to
data dictionary. This allows to specify instrument or whole data
dictionary in text document, which for most is easier to work with and
easily modifiable. The generic case is a data frame with variable names
as values in a column. This is a format like the REDCap data dictionary,
but gives a few options for formatting. Has a few related functions for
data handling and formatting. One interesting function is
case_match_regex_list(), which allows for a dynamic
dplyr::case_when()-like approach for regex-matching. I
think it is neat at least.
shiny_cast() with
shinylive, I need to remove curl as a
dependency. To accomplish this, the shiny_deploy() helper
functions has been moved to the package project.aid.
This was before realising that REDCapR has
curl as dependency, which is the culprit.
REDCapCAST is not going to be a shinylive
web-app without removing REDCapR dependency or any other
REDCap database interaction, which would defy the purpose. I’ll stick to
hosted Shiny app instead.Fix: ds2dd(): uses correct default dd column names.
Will be deprecated.
Fix: easy_redcap(): fixed to actually allow project
naming. also specifically asks for uri. widening updated to
work.
Fix: redcap_wider(): updated to accept more formats
and allow handling of simple projects without repeating instruments and
not longitudinal.
Fix: read_redcap_tables(): now handles
non-longitudinal project without repeatable instruments.
NEW: ds2dd_detailed(): extension of the
ds2dd(), which serves to preserve as much metadata as
possible automatically. Depends on a group of helper functions also
introduced. Of special note is the
guess_time_only_filter(), which will try to guess which
columns/variables should be formatted as time only formats. Supports hms
time format. DETAILED INSTRUCTION AND VIGNETTE IS PENDING.
NEW: read_redcap_instrument(): convenience function
to retrieve complete instrument. Goes a little against the focused
approach. With REDCapR::redcap_read() you can specify a
form to download. You have to also specify the record id variable
though. This is done for you with read_redcap_instrument().
Nothing fancy.
NEW: shiny_cast(): Shiny application to ease the process
of converting a spreadsheet/data set to a REDCap database. The app runs
locally and data is transferred securely. You can just create and upload
the data dictionary, but you can also transfer the given data in the
same process. I plan to host the app with shinyapps.io, but for now you
can run it locally.
I believe renv has now been added and runs correctly.
After clone, do renv::restore() to install all necessary
package to modify the package. This seems to always be back and forth.
renv may be on its way out again.
Added a Code of Conduct.
read_redcap_tables(): checking form names based on
data dictionary to allow handling of non-longitudinal projects. Prints
invalid form names and invalid event names. If invalid form names are
supplied to REDCapR::redcap_read() (which is the backbone),
all forms are exported, which is not what we want with a focused
approach. Invalid event names will give an output with a rather peculiar
formatting. Checking of field names validity is also added.One new function to ease secure dataset retrieval and a few bug fixes.
easy_redcap() function to ease the retrieval of a
dataset with read_redcap_tables() with
keyring-package based key storage, which handles secure API
set, storage and retrieval. Relies on a small helper function,
get_api_key(), which wraps relevant
keyring-functions. Includes option to cast the data in a
wide format with flag widen.data.REDCap_split(): when using this function on its
own, supplying a data set with check boxes would fail if metadata is
supplied as a tibble. Metadata is now converted to data.frame.
Fixed.read_redcap_tables(): fixed bug when supplying
events.This version marks the introduction of a few helper functions to handle database creation.
New: ds2dd() function migrating from the
stRoke-package. Assists in building a data dictionary for
REDCap from a dataset.
New: strsplitx() function to ease the string
splitting as an extension of base::strsplit(). Inspiration
from https://stackoverflow.com/a/11014253/21019325 and
https://www.r-bloggers.com/2018/04/strsplit-but-keeping-the-delimiter/.
New: d2n() function converts single digits to
written numbers. Used to sanitize variable and form names in REDCap
database creation. For more universal number to word I would suggest
english::word() or xfun::numbers_to_words(),
though I have not been testing these.
Page added. Vignettes to follow.
GithubActions tests added and code coverage assessed. Badge galore..
To reflect new functions and the limitation to only working in R, I have changed the naming of the fork, while still, of course, maintaining the status as a fork.
The versioning has moved to a monthly naming convention.
The main goal this package is to keep the option to only export a
defined subset of the whole dataset from the REDCap server as is made
possible through the REDCapR::redcap_read() function, and
combine it with the work put into the REDCapRITS package and the
handling of longitudinal projects and/or projects with repeated
instruments.
read_redcap_tables() NEW: this
function is mainly an implementation of the combined use of
REDCapR::readcap_read() and REDCap_split() to
maintain the focused nature of REDCapR::readcap_read(), to
only download the specified data. Also implements tests of valid form
names and event names. The usual fall-back solution was to get all
data.
redcap_wider() NEW: this function
pivots the long data frames from read_redcap_tables() using
tidyr::pivot_wider().
focused_metadata() NEW: a hidden
helper function to enable a focused data acquisition approach to handle
only a subset of metadata corresponding to the focused dataset.