Read BirdNET selection tables

Reads the selection tables that BirdNET writes in the Raven format (one tab-separated .txt file per recording or per site), with the column names standardized as in the 'sheepdog' package: lower case, without units, and with spaces replaced by underscores (Begin Time (s) becomes begin_time).

Usage

fp_read_birdnet(path, pattern = "\\.txt$", recursive = TRUE, tz = "UTC")

Arguments

path

Paths of selection tables, or of directories that hold them.

pattern

Regular expression for the table files inside directories.

recursive

Search directories recursively?

tz

Time zone of the recording times in the file names.

Value

A data frame with one row per detection, with the standardized columns of the tables plus table (the selection table it came from), file (the name of the audio file), start (start of the detection in the audio file, in seconds: file_offset when the table has it, otherwise begin_time), and time (recording start parsed from a YYYYMMDD_HHMMSS stamp in the file name, plus start; NA when the name has none).

See also

fp_aggregate(), fp_read_validation()

Examples

tab <- tempfile(fileext = ".txt")
writeLines(c(
  paste("Selection", "View", "Channel", "Begin Time (s)", "End Time (s)",
        "Begin Path", "File Offset (s)", "Species Code", "Common Name",
        "Confidence", sep = "\t"),
  paste(1, "Spectrogram 1", 1, 0, 3, "/rec/SITE1_20240101_200000.wav", 0,
        "compau", "Common Pauraque", 0.81, sep = "\t")
), tab)
fp_read_birdnet(tab)
#>   selection          view channel begin_time end_time
#> 1         1 Spectrogram 1       1          0        3
#>                       begin_path file_offset species_code     common_name
#> 1 /rec/SITE1_20240101_200000.wav           0       compau Common Pauraque
#>   confidence                                table                      file
#> 1       0.81 /tmp/Rtmp1DsqfQ/file1b92529aa44e.txt SITE1_20240101_200000.wav
#>   start                time
#> 1     0 2024-01-01 20:00:00