write listw to ArcGIS SWM dbf format

[JosiahParry]

This PR introduces a new function that enables a user to write a listw object to the .dbf format that can be converted into a binary .swm file within ArcGIS Pro.

[JosiahParry]

Similarly here, the region.id must be coercible to integer to be compatible with the swm file format.

Should tryCatch() be used to test if the data is coercible?

[JosiahParry]

Reprex:

library(spdep)
#> Loading required package: spData
#> To access larger datasets in this package, install the spDataLarge
#> package with: `install.packages('spDataLarge',
#> repos='https://nowosad.github.io/drat/', type='source')`
#> Loading required package: sf
#> Linking to GEOS 3.11.0, GDAL 3.5.3, PROJ 9.1.0; sf_use_s2() is TRUE

# read the NC SIDS dataset 
nc_sf <- sf::st_read(system.file("gpkg/nc.gpkg", package="sf")[1])
#> Reading layer `nc.gpkg' from data source 
#>   `/Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/library/sf/gpkg/nc.gpkg' 
#>   using driver `GPKG'
#> Simple feature collection with 100 features and 14 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS:  NAD27

# create a UniqueID field
nc_sf$UniqueID <- as.character(1:nrow(nc_sf))

# create a listw object
listw <- nb2listw(
  poly2nb(nc_sf$geom)
)

# create temporary file
tmp <- tempfile(fileext = ".dbf")

# write it out
write.swmdbf(listw, tmp, "UniqueID", region.id = nc_sf$UniqueID)
#> Loading required namespace: foreign

# read it back
read.swmdbf2listw(tmp, style = "B")
#> Warning in read.swmdbf2listw(tmp, style = "B"): region.id not given, c(MYID,
#> NID) range is 1:100
#> Characteristics of weights list object:
#> Neighbour list object:
#> Number of regions: 100 
#> Number of nonzero links: 490 
#> Percentage nonzero weights: 4.9 
#> Average number of links: 4.9 
#> 
#> Weights style: B 
#> Weights constants summary:
#>     n    nn  S0       S1       S2
#> B 100 10000 100 44.65023 410.4746

[rsbivand]

Thanks, this looks sensible, do you @JosiahParry or @scw have any further changes?

[JosiahParry]

@rsbivand I've gone ahead and fixed the spaces v tab issue. That should be resolved. I'm good with this one!

@mjanikas shared some code with me a while ago I rediscovered for reading the binary version of the swm. I'll try and prepare a PR for that in the coming days. Many of us use {spdep} for validation of statistics and ideas and having a way to convert to and from is proving to make our work much nicer! Thank you for being receptive. 🙏🏼

[rsbivand]

CI stuck on rgeoda missing from CRAN because BH geometry headers are broken in clang 19 GeoDaCenter/rgeoda#49 eddelbuettel/bh#101. I guess I'll drop the remainder of rgeoda until BH is updated next month or thereafter.

[JosiahParry]

Oof! The issue is looks to be for Fedora 36 which reached end of life 18 months ago. FWIW I have a CRAN package that does not compile on Fedora 36. As long as the error is informative and occurs during the configure step with a note in cran-comments.md, it seems to be acceptable.

[rsbivand]

No, it applies for forthcoming clang, Boost headers are not defensively written, and have fixed this upstream, but a new BH package release is needed.

[rsbivand]

Problems also because of https://forum.posit.co/t/r-cmd-check-fails-on-macos-latest-release-r-lib-actions-setup-r-v2/195084. Local R CMD check is more reliable than CI, at least in this case. This PR seems to be associated with a stale version of main, as only macOS fails for my commit to remove rgeoda.

[rsbivand]

Apart from the macOS CI, looks good!

[rsbivand]

1.3-7 submitted 13:41 CET, online 14:58 CET