rectify returns a cropped result

[sigmafelix]

Thank you for maintaining the package. I found that rectify for curvilinear grids (i.e., MODIS swath data) returns a cropped rectilinear SpatRaster. Per the statement in the issue template, I did not include plots of rasters.

library(terra)
library(luna)
tdir <- tempdir()
# luna::getProducts("MOD06")

date_start <- "2023-06-01"
date_end <- "2023-06-02"
aoi_in <- c(-79.5, -77, 34, 36)
mod06f <-
    luna::getNASA(
        product = "MOD06_L2",
        start_date = date_start,
        end_date = date_end,
        aoi = aoi_in,
        version = "6.1",
        server = "LAADS",
        download = TRUE,
        username = "****",
        password = "****",
        path = tdir)

pathsds <- sprintf("HDF4_EOS:EOS_SWATH:%s:mod06:Cloud_Fraction", mod06f[1])

cloudfrac <- rast(pathsds)
# Warning message:
# [rast] the data in this file are rotated. Use 'rectify' to fix that
cloudfrac
# class       : SpatRaster 
# dimensions  : 406, 270, 1  (nrow, ncol, nlyr)
# resolution  : 0.01513012, 0.008826409  (x, y)
# extent      : -77.26448, -73.17935, 27.19393, 30.77745  (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84) 
# source      : MOD06_L2.A2023152.0325.061.2023152155452.hdf:mod06:Cloud_Fraction 
# varname     : MOD06_L2.A2023152.0325.061.2023152155452 
# name        : Cloud Fraction in Retrieval Re~m Pixels) from 1-km Cloud Mask 

cloudfrac_r <- rectify(cloudfrac)
cloudfrac_r
# class       : SpatRaster 
# dimensions  : 470, 336, 1  (nrow, ncol, nlyr)
# resolution  : 0.01513012, 0.008826409  (x, y)
# extent      : -78.26307, -73.17935, 26.62904, 30.77745  (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84) 
# source(s)   : memory
# name        : Cloud Fraction in Retrieval Re~m Pixels) from 1-km Cloud Mask 
# min value   :                                                             0 
# max value   :                                                             1 

It appears that rectify conserves the extent of the original curvilinear raster.

However, stars::st_warp seems to work as expected:

library(stars)
cloudfrac_st <- read_stars(pathsds)
# threshold set to 0.231967 : set a larger value if you see missing values where they shouldn't be
# Warning message:
# In transform_grid_grid(st_as_stars(src), st_dimensions(dest), threshold) :
#   using Euclidean distance measures on geodetic coordinates
cloudfrac_st_r <- st_warp(cloudfrac_st, crs = "OGC:CRS84")
# stars object with 2 dimensions and 1 attribute
# attribute(s):
#                 Min. 1st Qu. Median      Mean 3rd Qu. Max.  NA's
# Cloud_Fraction     0    0.08   0.88 0.6097394       1    1 30183
# dimension(s):
#   from  to offset    delta refsys x/y
# x    1 385 -101.9  0.07729 WGS 84 [x]
# y    1 286  45.98 -0.07729 WGS 84 [y]

I guess this issue related to the source in gdal_algs.cpp.