local file content substituted for dimension name

[Edzer Pebesma]

This is part of my client script:

temp_period = p$aggregate_temporal(data = datacube,
                                   intervals = intervals,
                                   reducer = function(data, context){p$median(data)},
                                   labels = labels,
                                   dimension = "t")

when I have a file called "t" in my working directory, its content is substituted for the value of dimension. If I remove or rename that file, it works fine.

[Edzer Pebesma]

Complete script:

library(openeo)
con = openeo::connect("https://openeo.cloud")
collection = "TERRASCOPE_S5P_L3_NO2_TD_V1"
coll_meta = describe_collection(collection)
w = -20; s = 0; e = 50; n = 70
# smaller:
w = 7; s = 51.5; e = 8; n = 52.5 
date1 = "2020-01-01"
date2 = "2021-01-01"
p = processes(con)
datacube = p$load_collection(
     id = "TERRASCOPE_S5P_L3_NO2_TD_V1",
      spatial_extent = list(west = w, south = s, east = e, north = n),
      temporal_extent=c(date1, date2),
      bands=c("NO2")
)
# interpolate = function(data,context) {
#     return(p$array_interpolate_linear(data = data))
# }
# datacube = p$apply_dimension(process = interpolate,
#                              data = datacube, dimension = "t")
intervals = list(c('2020-01-01', '2020-02-01'),
                 c('2020-02-01', '2020-03-01'),
                 c('2020-03-01', '2020-04-01'),
                 c('2020-04-01', '2020-05-01'),
                 c('2020-05-01', '2020-06-01'), 
                 c('2020-06-01', '2020-07-01'), 
                 c('2020-07-01', '2020-08-01'),
                 c('2020-08-01', '2020-09-01'),
                 c('2020-09-01', '2020-10-01'), 
                 c('2020-10-01', '2020-11-01'),
                 c('2020-11-01', '2020-12-01'), 
                 c('2020-12-01', '2020-12-30'))
# and labels
labels = sapply(intervals, head, 1) # create labels from list
temp_period = p$aggregate_temporal(data = datacube,
                                   intervals = intervals,
                                   reducer = function(data, context){p$median(data)},
                                   labels = labels,
                                   dimension = "t")
result = p$save_result(data = temp_period, format="NetCDF")
login()
compute_result(result, format = "NetCDF", output_file = "ndvi.nc", con = con)

same happens for apply_dimension when that part gets uncommented.

[Matthias Mohr]

I can confirm this with my R client that I compiled mid-October from dev on Windows! This is a huge privacy issue!

This script sends my local file "t" to the server as value for the dimension parameter.
I also tried t as collection name (id in load_colleciton), but there it doesn't load the file.

[Florian Lahn]

Sorry about that. The feature to check and read existing files was requested for UDFCode parameter. But somehow the code got duplicated and ended up in argument String.