Introduction
If you are reading this vignette you are most probably to contribute to the mapme.biodiversity package. This is great news and we are very happy to receive Pull-Requests extending the package’s functionality! Below you will find important in-depth information about how to add resources and indicators to make the process as seamless as possible for both you and the package’s maintainers. Please make sure to read and understand this guide before opening a PR. If in doubt, especially if you feel that the framework does not support your use case, always feel free to raise an issue and we will happily discuss how we can support your ideas. If you have not already done so, make sure to read the Terminology vignette to get familiar with the most important concepts of this package.
Note that we use the tidyverse
style guide for the package. That specifically means that function
and variable names should follow the snake case pattern. We also use the
arrow assignment operator (<-). When submitting a PR
that does not consistently follow the tidyverse style guide, the
maintainers of the package might change the code to adhere to this code
style without further notice before accepting the PR.
Getting started
Ideally, you clone the GitHub repository via the git command in a command line on Linux and MacOS systems or via the GitHub Desktop application on Windows. On Linux, the command would look like this:
We do not accept pushes to main, thus the first step would be to
create a specific branch for your extension. In this tutorial, we will
pretend to re-implement the nasa_srtm resource and the
associated elevation indicator, so we will create a branch
reflecting this. Don’t forget to check out to the newly created
branch!
Below, we will assume that you develop your extension to the package in R Studio. The general guidelines to follow also apply if you choose different tooling for your development process, however, it will not be covered in this vignette. We assume that all R development dependencies are installed. The easiest way to ensure this is using devtools:
devtools::install_dev_deps()Adding a resource
Overview of adding a resource
A resource is a supported dataset that can be made available from a
user’s perspective by specifying one or more functions to
get_resources(). Currently, the package supports only
raster and vector resources. If you wish to submit the support of a new
resource, please be aware that we will only accept new resources if they
are associated with at least one indicator calculation. The very first
step to adding a resource is to create a new file that will be holding
the required code.
Once you checked out to the new branch and having the project opened in
R Studio, adapt the following command to open the a new resource
file:
file.edit("R/get_<your-new-resource>.R")
# e.g. file.edit("R/get_soildgrids.R")Documenting the new resource
In the first part of a resource function, make sure to include detailed documentation. This documentation should explain what this resource represents, where it comes from (including a citation), and the user-facging arguments that should be specified during runtime.
Importantly, this documentation MUST receive the
roxygen tag @keywords resource, so that the documentation
will be identified as a resource. Also, add the bare name of the
resource as the @name tag (e.g. in the case of our example
that translates to @name nasa_srtm).
#' Short title
#'
#' One or more description paragraphs might follow here. Please describe
#' the spatio-temporal structure of your resource here briefly.
#'
#' @name <the short name of your resource>
#' @param <any user-facing arguments>
#' @keywords resource <identifies the documentation as a resource>
#' @references <ideally a citable scientific publication>
#' @source <a link in the \url{} tag linking to an online documentation>
#' @returns A function that makes a resource available for a portfolio
#' @include register.R
#' @exportThe last two tags are important to add as well. The include statement is mandatory for the register functionality (more on that below) to be loaded before your resource function. The export tag is important so that the resource is actually exposed to the users of the package.
Constructing a resource function - Outer level
Resource functions are constructed as closures, i.e. functions that return a function. The outer level exposes arguments to be set by users of the function to fine-control the flow of the function. Note, it is important to check the user input in this outer level for correctness so that warning/error messages in case of any miss specifications are thrown immediately.
For nasa_srtm, this outer level does not look really
exciting becuase there are no user-facing arguments to be checked (we
will see how to check user-facing arguments when constructing the
indicator below):
get_nasa_srtm <- function() {
# .... inner function level
}Note, that there are some exported helper functions for re-occurring
argument checks that you are free to use
(e.g. check_available_years() in case you query the user
for a temporal time frame). The arguments defined in the outer level of
the resource function are then ready to be used in the inner level, that
we will have a look at next.
Constructing a resource function - Inner level
The inner level of a resource function has a mandatory function
signature that will be checked during run-time. Your function is
required to exactly specify the below signature. For the
nasa_srtm resource, this looks like this:
function(x,
name = "nasa_srtm",
type = "raster",
outdir = mapme_options()[["outdir"]],
verbose = mapme_options()[["verbose"]]) {
# ... function body
}The x argument here represents the portfolio object
handed over by the user when calling get_resources() which
is an sf-object and can thus be used to derive the spatial
extent of the portfolio. Next, comes the name and the type of the
resource which is required for the backend to correctly handle the
output and log the resource once it has been made available.
The other arguments should default to the their respective output
values of mapme_options() and represent a character vector
for the output directory, a logical to control the verbosity. Note, that
the output directory might be NULL in case a user wishes to
access data directly from a remote source. We will look into how these
things come together now as we peak into constructing the actual body of
a resource function.
Constructing a resource function - Body
The expected output of a resource function is an
sf-object with geometries representing the bounding box for
single resource elements (e.g. tiles in case of a raster resource). We
provide you with functionality to seamlessly produce such a footprint
object via make_footprints(). Note, this function either
excepts a character vector to GDAL readable data sources or an
sf object. In case you provide a charachter vector, the
bounding box information will be retrieved automatically. This comes
with a performance penalty for remote sources, because each file has to
be opened once, so always opt for constructing an sf object
in the resource function, if feasible.
The footprints sf object is expected to contain a column
called source that points to a GDAL readable data source
and the geometries to correspond to the bounding box of each single
resource element. You might opt to supply a filename
argument to make_footprints(). This defaults to
basename(srcs[["source]]), so you might supply a better
suited filename, i.e. in the case the source location ends with an API
key value, or similar.
Next to specifying if the the resource you wish to turn into a
footprint object represents a raster or vector resource, you can specify
opening and creation options. Opening options are relevant if opening a
remote source requires location or driver dependent options
(e.g. specifying non-standard columns names for longitude/latitude with
the CSV driver). Creation options are relevant if a user specified a
outdir argument and and refer to the arguments used by gdal_translate.
These should specifically specify data type and compression algorithms
for raster resources, but you are otherwise free to optimize the data
layout for efficient access.
Both oo and co can be specified as a single
character vector, in which case the same options are applied to all
elements of a resource, or as a list in order to accommodate file
specific options.
Use the verbose argument to decide if informative
messages should be printed, e.g. to inform users about download
progress. Errors or warnings should be emitted in either case.
If there is no intersection between the x object and
your resource, make sure to return NULL as early as
possible.
Adding sample resource for package internal testing
We ask you to provide a small subset of your resource to
inst/res/resource_name so that indicators that depend on
the resource can be tested without the need to actual download the
resource.
Because there are some restrictions to the final size of the package,
we ask you to put substantial effort in reducing the size of the files
to a minimum. This includes cropping all resource samples to the spatial
extent of the polygon provided in
inst/extdata/sierra_de_neibe_478140.gpkg or a polygon of
similar size supplied by you in case it the spatial extent does not
intersect with your resource.
For raster resources, if the original raster is encoded as float, consider changing the data type to integer by introducing a scale factor. Also, please use a compression algorithm to further reduce the file size. For vector resources, consider reducing the number of vertices in case the geometries are very complex.
Finally, put your processing script of the resource into
data-raw to ensure reproducibility. Then, you are required
to write a unit-test for your resource function, which should execute as
much of your code as possible without actually conducting a
download.
A note on dependencies for resources
Note, that a resource SHALL NOT add additional dependencies to the package. If you add dependencies we require you to add a supporting statement to your PR explaining why these dependencies are needed and why other approaches would fail. Before accepting your PR, we might request you to change your code to remove these dependencies, if it is feasible to achieve the same functionality without.
Adding an indicator
The process of adding an indicator is very similar to the one for resources. However, some input-output requirements are different. Note, that in case that you added a new resource we also expect a new indicator taking advantage of that resource in your PR.
As you will see, there are two new important concepts to have in mind when adding an indicator. These are the processing mode and computational engines. We will briefly explain these concepts below, however, you can also head over to the Terminology vignette if you are interested in a more comprehensive definition of these two terms.
Overview of adding a new indicator
An indicator is a logical routine depending on one or more resources
that extracts numeric outputs for all assets in a portfolio. From a
user’s perspective, indicators are processed via the
calc_indicators() function. You as a developer will have to
construct an indicator function as a closure, e.g. a function that
returns another function. The outer level exposes user-facing arguments
and checks that they are correctly specified, while the inner level is
required to follow a specified signature and returns a tibble.
Once you checked out to the new branch and having the project opened in R Studio, adapt the following command to open the a new indicator file:
file.edit("R/calc_<your-new-indicator>.R")
# e.g. file.edit("R/calc_precipitation")Documenting the new indicator
In the first part of an indicator function, make sure to include
detailed documentation. This documentation should explain which
resources are required to calculate the indicator, the user-facing
arguments that should be specified during runtime and the structure of
the output tibble. Importantly, this documentation MUST
receive the roxygen tag @keywords indicator, so that the
documentation will be identified as an indicator. Also, add the bare
name of the indicator as the @name tag
(e.g. @name elevation).
#' Short title
#'
#' One or more description paragraphs might follow here. Please describe
#' required resource and user arguments here.
#' Please document which processing engines are available for your indicator
#' and briefly describe how the indicator is derived from its inputs.
#'
#' @name <the short name of your indicator, same as in the backlog>
#' @param <any user-facing arguments>
#' @keywords indicator <identifies the documentation as an indicator>
#' @returns A function that calculates an indicator for a portfolio
#' @include register.R
#' @exportThe last two tags are important to add as well. The include statement is mandatory for the register functionality (more on that below) to be loaded before your indicator function. The export tag is important so that the resource is actually exposed to the users of the package.
Constructing an indicator function - Outer level
Indicator functions are constructed as closures, i.e. functions that return a function. The outer level exposes arguments to be set by users of the function to fine-control the flow of the function. Note, it is important to check the user input in this outer level for correctness so that warning/error messages in case of any miss specifications are thrown immediately.
For elevation, this outer level could look something
like this:
calc_elevation <- function(engine = "extract",
stats = "mean") {
engine <- check_engine(engine)
stats <- check_stats(stats)
# ... inner function level
}There are some exported helper functions for re-occurring argument
checks that you are free to use (e.g. check_engine()).
Note, that the arguments defined this way in the outer level of the
indicator function are then ready to be used in the inner level that we
will have a look at next.
Constructing an indicator function - Inner level
The inner level of an indicator function has a mandatory function
signature that will be checked during run-time. Your function is
required to exactly specify the below signature. For the
elevation indicator, this looks like this:
function(x,
nasa_srtm = NULL,
name = "elevation",
mode = "asset",
aggregation = "stat",
verbose = mapme_options()[["verbose"]]) {
# ... function body
}The x argument here represents the portfolio object
handed over by the user when calling get_resources() which
is an sf-object with 'POLYGON' as features.
Next, comes the name(s) of the required resource(s) and the name of the
indicator. What follows is the computation mode, that must be one of
"asset" or "portfolio".
We realized, that for large (potentially global) portfolios, depending on the spatial resolution of a resource, different processing modes substantially impact the time needed for a computation. For high to medium resolution raster resources, processing on the asset level benefits computation time. However, spatially cropping coarse resolution datasets for a high number of assets introduces significant overhead, thus processing these resources on a portfolio level is more efficient.
If neither of the two processing modes lead to satisfactory processing times for your indicator, please leave an issue/comment to discuss the addition of another processing mode with the maintainers of the package.
The argument aggregation governs how chunked results for
large polygons are to be combined into a single indicator. In case uses
supply polygons larger than the specified code path or assets of type
MULTIPOLYGON, a code path is triggered that splits assets
into sub-components. The aggregation method specifies which statistic is
used to combine values that share the same values in the remaining
indicator columns (i.e. datetime, variable,
and unit). The stat keyword is a special
keyword used for indicators where statistics are specified by the user
and will trigger to select the respective statistic as the aggregation
statistic (e.g. take the sum of sums).
The available statistics are:
mapme.biodiversity:::available_stats
#> Error in get(paste0(generic, ".", class), envir = get_method_env()) :
#> object 'type_sum.accel' not found
#> [1] "mean" "median" "sd" "min" "max" "sum" "var"The argument verbose defaults to the corresponding
package-wide option and should control the verbosity of you indicator
function.
Constructing an indicator function - Body
The expected output of an indicator function is a tibble. Depending
on the mode specified for processing, it is a single tibble
for mode = "asset", or a list of tibbles equal to the rows
of x in case mode = "portfolio".
You may use helper functions provided by the package for a common
interface e.g. for vector-raster zonal statistics (e.g. by using
select_engine()).
You are encouraged to write your own helper function that are needed for your indicator processor. These should be located in the same file as the main processor, start with a dot and should not be exported.
If you wish to include roxygen documentation for your helpers, make
sure to add the @keywords internal and @noRd
tags to your functions. If you feel that one or more of your helper
functions would be of benefit to more that just one indicator, please
comment in and issue/pull-request to discuss with the package
maintainers if your helper function could be moved to
R/utils.R.
Use the verbose argument to decide if informative
messages should be printed, e.g. to inform users about processing
progress. Errors or warnings should be emitted in either case.
If there is no intersection between the x object and the
required resources, or for any other reason why your indicator might not
be calculated with the given configuration, make sure to return
NA as early as possible.
Adding units tests for an indicator
You are required to add unit tests for your indicator using the package internal example data sets for resources. Make sure to properly test for miss-specification of user-facing arguments and also check for the correctness of numerical results of your indicator.
You might not need to construct a portfolio from scratch to test you
indicator function. Instead, you can directly call the returned function
on an appropriate polygon with the respective required resource. For the
elevation indicator, this looks like this:
x <- read_sf(system.file(
"extdata", "sierra_de_neiba_478140.gpkg",
package = "mapme.biodiversity"
))
nasa_srtm <- list.files(
system.file(
"res", "nasa_srtm",
package = "mapme.biodiversity"
),
pattern = ".tif$", full.names = TRUE
)
nasa_srtm <- rast(nasa_srtm)
ce <- calc_elevation(stats = c("mean", "median", "sd"))
result_multi_stat <- ce(shp, nasa_srtm)
expect_equal(
names(result_multi_stat),
c("elevation_mean", "elevation_median", "elevation_sd")
)