Error in sits_accuracy()

[cs-caroline]

I got this error when applying the sits_accuracy function:
Error: sits_accuracy_class_cube: NULL value not allowed for pred_ref - invalid validation data

I exported the validation samples that SITS generated and checked them in QGIS, after that, I exported the file and uploaded it to R again. I've tried the CSV, GPKG and SHP formats, but they all gave an error. The file has the columns label, start_date, end_date, longitude and latitude, and is in the EPSG:32722 projection. I checked if the labels are compatible, if the formats are correct, if the dates match, if there are any null values, but I didn't find anything that could lead to the error. Any idea what might be happening?

a2024_samples_sf <- sits_stratified_sampling(
  cube = a2024_map,
  sampling_design = a2024_sampling_design,
  alloc = "alloc_120",
  multicores = 4,
  shp_file = "D:/carol/Área de Trabalho/SER-411-Avançado/BDC/Cubo 2024/Classificação/validacao_2024.shp"
)

# Get ground truth points
valid_csv <- "D:/carol/Área de Trabalho/SER-411-Avançado/BDC/Cubo 2024/Classificação/r_2024.csv"
validation_data <- read.csv(valid_csv)

validation_data <- validation_data %>%
mutate(
start_date = as.Date("2024-01-01"),
end_date = as.Date("2024-10-01")
)

# Calculate accuracy according to Olofsson's method
area_acc <-
  sits_accuracy(
    a2024_map,
    validation = validation_data
    multicores = 4
)

[OldLipe]

Hi @cs-caroline, thank you for reporting the issue.

I could not reproduce your error. Could you please share your validation sample file with us? It can be just a few samples so we can run with them and try to reproduce the error. Also, it would be good if you shared the full script you are using so we can explore the previous steps as well.