This is an in-depth guide on contributing to the package. In order to get started with contributing, please read our contributing guide
Adding new Tables, Listing and Graphs
In order to add new TLGs, two things are needed: - a TLG definition specified in the tlg.yaml file - an implementation function that is responsible for generating an entry (table, list, graph).
TLG definition
Specified in tlg.yaml file. Entry in that file is responsible for providing some metadata regarding the TLG (like name, descriptions, links), a function reference for creating the resulting TLG and options, taken as arguments by the generating function, that allow for quick and easy definition of widgets to be rendered in the application interface. This then allows the user to customize plots in accordance to their needs.
General format
The entry should have the following format. Identifying keys (wrapped
in <>
) should be provided by the creator and be
unique within their scope (indentation level).
# unique identifier for given entry
<entry id>:
# true / false whether TLG should be included as default
is_default:
# type of the TLG, in this case must be Graph
type:
# name of the dataset
dataset:
# standarized id of the TLG, e.g. pkcg01
pkid:
# short label to be displayed as tab name
label:
# longer descriptions, to be displayed in the order table
description:
# link to the documentation of the TLG
link:
# name of the function exported by the package, responsible for generating TLG, must return a list of plots to be displayed
fun:
# options that can be passed as arguments to the function, will generate an editing widget on the TLG page
options:
# see below mor more information on options
template:
# You can also specify template definitions. This is a character string with id of another TLG definition. All options will be copied over from template id. Any provided parameters will be a) overwritten if existing in the template or b) added as new.
Options
Options should be passable as arguments into the implementation
function. Each provided option will generate a widget for the user to
provide customized values. This is the main interface to allow the users
to customize their reports. Options should be identified by a unique
option_id
, that is the same as the named argument provided
to the implementation function. In example, if the implementation
function allows to specify xaxis_label
as an argument, the
option_id
should be the same.
There are several types of input widgets that you can specify.
text
Standard shiny textInput
.
# option id, the same as the argument that is passed to the rendering function, must be unique in the scope of the TLG entry
<option id>:
# type of the option/widget
type: text
# label to be displayed in the editing widget
label:
# default value to be provided in the field;
# if provided, will overwrite function argument defaults;
# if left empty, function defaults will be applied;
default:
numeric
Standard shiny numericInput
.
# option id, the same as the argument that is passed to the rendering function, must be unique in the scope of the TLG entry
<option id>:
# type of the option/widget
type: numeric
# label to be displayed in the editing widget
label:
# default value to be provided in the field;
# if provided, will overwrite function argument defaults;
# if left empty, function defaults will be applied;
default:
select
Standard shiny selectInput
.
# option id, the same as the argument that is passed to the rendering function, must be unique in the scope of the TLG entry
<option id>:
# type of the option/widget
type: select
# label to be displayed in the editing widget
label:
# choices to pick from the dropdown, either specified outright or using a special keyword:
# - '.colnames' keyword will pull the choices from the data column names
# - '$COLUMN_NAME' keyword will pull choices from values of a specific column
choices:
# whether to allow for multiple values to be selected
multiple:
# default value to be provided in the field;
# if provided, will overwrite function argument defaults;
# if left empty, function defaults will be applied;
# '.all' keyword can be applied to select all choices;
default:
table
# option id, the same as the argument that is passed to the rendering function, must be unique in the scope of the TLG entry
<option id>:
# type of the option/widget
type: table
# label to be displayed in the editing widget
label: "Variables formatting"
# column definition for the input table, specifying each column type
cols:
# colum name, as accepted in the implementation function
<column_name>:
# type of the input widget, either `text` or `select`
type:
# label for the column header
label:
# applicable to `select` type, choices to pick from the dropdown, either specified outright or using a special keyword:
# - '.colnames' keyword will pull the choices from the data column names
# - '$COLUMN_NAME' keyword will pull choices from values of a specific column
choices:
# default rows definition, should contain a list of rows that should be applied as defaults;
# each row should be an array containing value for each defined column;
# if row value for specific column should be empty, use special `$NA` keyword.
default_rows:
Examples
For examples of TLG definition implementations, please see tlg.yaml file in the repository.
Implementation function
Function should be defined within aNCA
package and meet
the following criteria:
- name should be prefaced with letter appropriate for tlg type:
-
t_*
for tables -
l_*
for listings -
g_*
for graphs
-
- should accept a
data
argument for providing results to be rendered as a given TLG - should return a list of entries or a single entry.