fix issue with gdal-warp

[mlampros]

First of all thanks for making this repository publicly available. This pull request fixes issues #93 and #92
The following is a reproducible example on Linux Mint 18.2

sp_path = ebirdst::ebirdst_download("example_data")
abd = ebirdst::load_raster(product = "abundance", sp_path)

mod_date = MODIS::transDate(begin = as.Date('2018-02-01'),
                            end = as.Date('2018-02-02'))

mod_tile = MODIS::getTile(x = abd)

tifs = MODIS::runGdal(product = "MODOCGA",
                      begin = mod_date$beginDOY,
                      end = mod_date$endDOY,
                      tileH = mod_tile@tileH,
                      tileV = mod_tile@tileV,
                      extent = mod_tile, 
                      job = "modis")

By running the previous code I received,

Error in sf::gdal_utils(util = "warp", source = gdalSDS, destination = ofile,  : 
  gdal_utils warp: an error occured

The error also revealed,

Output driver `79' not recognised or does not support

The output files in the temporary directory were all of type ".tif" (GTIFF driver).

The runGdal() function calls internally the sf::gdal_utils() function which takes params as input which normally should be a vector of parameters such as,

'-of'  'GTIFF'   etc.

however it actually is

'-of'  '79'   etc.

In the 'runGdal()' function the 'GTIFF' driver is assigned to the dataFormat variable which equals to

dataFormat = checkGdalWriteDriver(dataFormat)

The checkGdalWriteDriver() function calls internally the getGdalWriteDrivers() function which in my case returns,

'data.frame':	45 obs. of  7 variables:
 $ name     : Factor w/ 210 levels "AAIGrid","ACE2",..: 197 79 129 87 46 117 15 147 148 92 ...
 $ long_name: Factor w/ 209 levels "ACE2","Aeronav FAA",..: 201 68 126 43 36 95 121 152 153 93 ...
 $ write    : logi  TRUE TRUE TRUE TRUE TRUE TRUE ...
 $ copy     : logi  TRUE TRUE TRUE TRUE FALSE FALSE ...
 $ is_raster: logi  TRUE TRUE TRUE TRUE TRUE TRUE ...
 $ is_vector: logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
 $ vsi      : logi  TRUE TRUE TRUE TRUE TRUE FALSE ...

Therefore the nms variable,

nms = getGdalWriteDrivers()$name

is of type factor and rather than returning the driver as a character string (i.e. 'GTIFF') it returns the level position,

 Factor w/ 210 levels "AAIGrid","ACE2",..: 197 79 129 87 46 117 15 147 148 92 ...

In this PR I converted the nms variable to a character string as the 'checkGdalWriteDriver()' function is called only once in the MODIS R package.

[fdetsch]

Marvellous, thanks for the PR. I'll merge into 'develop' and check back with those guys that encountered problems.

Out of interest, are you running R >= 4.0.0? I just want to rule out that this is related to recent changes in the default behavior of stringsAsFactors.

[mlampros]

Hi, I run the code of this PR in R version 3.6.3.
Yes, you are right, a user having 4.0.0 installed will highly probable won't encounter this error as the default behaviour has changed to

stringsAsFactors = FALSE

[fdetsch]

Right. Still, a highly valuable PR - I didn't realize this was related to format changes in sf::st_drivers() for different versions of R. Otherwise, I would have had to add R (>= 4.0.0) to Depends, which I figure isn't particularly desirable.