postProcess: Generic function to post process objects

In PredictiveEcology/reproducible: Enhance Reproducibility of R Code

Generic function to post process objects

Description

The method for GIS objects (terra ⁠Spat*⁠ & sf classes) will crop, reproject, and mask, in that order. This is a wrapper for cropTo(), fixErrorsIn(), projectTo(), maskTo() and writeTo(), with a required amount of data manipulation between these calls so that the crs match.

Usage

postProcess(x, ...)

## S3 method for class 'list'
postProcess(x, ...)

## Default S3 method:
postProcess(x, ...)

Arguments

x

A GIS object of postProcessing, e.g., Spat* or sf*. This can be provided as a rlang::quosure or a normal R object.

...

Additional arguments passed to methods. For spatialClasses, these are: cropTo(), fixErrorsIn(), projectTo(), maskTo(), determineFilename(), and writeTo(). Each of these may also pass ... into other functions, like writeTo(). This might include potentially important arguments like datatype, format. Also passed to terra::project, with likely important arguments such as method = "bilinear". See details.

Value

A GIS file (e.g., RasterLayer, SpatRaster etc.) that has been appropriately cropped, reprojected, masked, depending on the inputs.

Post processing sequence

If the rasterToMatch or studyArea are passed, then the following sequence will occur:

1. Fix errors fixErrorsIn(). Currently only errors fixed are for SpatialPolygons using buffer(..., width = 0).

2. Crop using cropTo()

3. Project using projectTo()

4. Mask using maskTo()

5. Determine file name determineFilename()

6. Write that file name to disk, optionally writeTo()

NOTE: checksumming does not occur during the post-processing stage, as there are no file downloads. To achieve fast results, wrap prepInputs with Cache

Backwards compatibility with rasterToMatch and/or studyArea arguments

For backwards compatibility, postProcess will continue to allow passing rasterToMatch and/or studyArea arguments. Depending on which of these are passed, different things will happen to the targetFile located at filename1.

See Use cases section in postProcessTo() for post processing behaviour with the new from and to arguments.

If targetFile is a raster (⁠Raster*⁠, or SpatRaster) object:

	rasterToMatch	studyArea	Both
extent	Yes	Yes	rasterToMatch
resolution	Yes	No	rasterToMatch
projection	Yes	No*	rasterToMatch*
alignment	Yes	No	rasterToMatch
mask	No**	Yes	studyArea**

*Can be overridden with useSAcrs.

**Will mask with NAs from rasterToMatch if maskWithRTM.

If targetFile is a vector (⁠Spatial*⁠, sf or SpatVector) object:

	rasterToMatch	studyArea	Both
extent	Yes	Yes	rasterToMatch
resolution	NA	NA	NA
projection	Yes	No*	rasterToMatch*
alignment	NA	NA	NA
mask	No	Yes	studyArea

*Can be overridden with useSAcrs

See Also

prepInputs

Examples

if (requireNamespace("terra", quietly = TRUE) && requireNamespace("sf", quietly = TRUE)) {
  library(reproducible)
  od <- setwd(tempdir2())
  # download a (spatial) file from remote url (which often is an archive) load into R
  # need 3 files for this example; 1 from remote, 2 local
  dPath <- file.path(tempdir2())
  remoteTifUrl <- "https://github.com/rspatial/terra/raw/master/inst/ex/elev.tif"

  localFileLuxSm <- system.file("ex/luxSmall.shp", package = "reproducible")
  localFileLux <- system.file("ex/lux.shp", package = "terra")

  # 1 step for each layer
  # 1st step -- get study area
  studyArea <- prepInputs(localFileLuxSm, fun = "terra::vect") # default is sf::st_read

  # 2nd step: make the input data layer like the studyArea map
  # Test only relevant if connected to internet -- so using try just in case
  elevForStudy <- try(prepInputs(url = remoteTifUrl, to = studyArea, res = 250,
                                 destinationPath = dPath, useCache = FALSE))

  # Alternate way, one step at a time. Must know each of these steps, and perform for each layer
  
    dir.create(dPath, recursive = TRUE, showWarnings = FALSE)
    file.copy(localFileLuxSm, file.path(dPath, basename(localFileLuxSm)))
    studyArea2 <- terra::vect(localFileLuxSm)
    if (!all(terra::is.valid(studyArea2))) studyArea2 <- terra::makeValid(studyArea2)
    tf <- tempfile(fileext = ".tif")
    download.file(url = remoteTifUrl, destfile = tf, mode = "wb", quiet = TRUE)
    Checksums(dPath, write = TRUE, files = tf)
    elevOrig <- terra::rast(tf)
    studyAreaCrs <- terra::crs(studyArea)
    elevForStudy2 <- terra::project(elevOrig, studyAreaCrs, res = 250) |>
      terra::mask(studyArea2) |>
      terra::crop(studyArea2)

    isTRUE(all.equal(elevForStudy, elevForStudy2)) # TRUE!
  

  # sf class
  studyAreaSmall <- prepInputs(localFileLuxSm)
  studyAreas <- list()
  studyAreas[["orig"]] <- prepInputs(localFileLux)
  studyAreas[["reprojected"]] <- projectTo(studyAreas[["orig"]], studyAreaSmall)
  studyAreas[["cropped"]] <- suppressWarnings(cropTo(studyAreas[["orig"]], studyAreaSmall))
  studyAreas[["masked"]] <- suppressWarnings(maskTo(studyAreas[["orig"]], studyAreaSmall))

  # SpatVector-- note: doesn't matter what class the "to" object is, only the "from"
  studyAreas <- list()
  studyAreas[["orig"]] <- prepInputs(localFileLux, fun = "terra::vect")
  studyAreas[["reprojected"]] <- projectTo(studyAreas[["orig"]], studyAreaSmall)
  studyAreas[["cropped"]] <- suppressWarnings(cropTo(studyAreas[["orig"]], studyAreaSmall))
  studyAreas[["masked"]] <- suppressWarnings(maskTo(studyAreas[["orig"]], studyAreaSmall))
  if (interactive()) {
    par(mfrow = c(2,2));
    out <- lapply(studyAreas, function(x) terra::plot(x))
  }

  setwd(od)
}

PredictiveEcology/reproducible documentation built on Dec. 12, 2024, 7:26 p.m.