using Glob, Skraak, CSV
folders=Glob.glob("*/2024-10-18/")

using Glob, Skraak, CSV, DataFrames
folders=Glob.glob("*/2025-05-18/")

        CSV.write("/media/david/Pomona-4/Pomona/pomona_files_20241126_new.csv", df; append=true)

        CSV.write("/media/david/Pomona-4/Pomona/pomona_files_20250522_new.csv", df; append=true)
        #CSV.write("/media/david/Pomona-4/Pomona/pomona_files_20241126_new.csv", df; append=true)

        xxh64 = String[],

            xxh64 = readchomp(`xxh-hash $file`) #run(`xxh-hash $file`)

                xxh64,

end
#=
using Opus, PNGFiles, Skraak
function image_from_opus(file)
    signal, freq = Opus.load(file)
    if size(signal, 2) == 2
        signal = (signal[:, 1] + signal[:, 2]) / 2
    end
    if freq > 8000
        signal, freq = Skraak.resample_to_8000hz(signal, freq)
    end
    freq = freq |> Float32
    sample = signal[:]
    image = Skraak.get_image_from_sample(sample, freq)
    PNGFiles.save("$(file).png", image)
    print(".")

=#

# assumes kiwi

##################################################################

    filter!(row -> row.label == label, df)

    filter!(row -> row.label == 1, df)

label = "Rowi"

function move_clips_to_folders(df::DataFrame)

function move_clips_to_folders(df::DataFrame, label::String)

    call_map = Dict(1 => "Duet", 2 => "Female - Solo", 3 => "Male - Solo")

        dst = "$(row.label)/$(row.file)"
        mkpath("$(row.label)/")

        if row.label == 4
            dst = "Don't Know/$(row.file)"
            mkpath("Don't Know/")
        else
            dst = "$label/$(call_map[row.label])/$(row.file)"
            mkpath("$label/$(call_map[row.label])")
        end
        #dst = "$(row.label)/$(row.file)"
        #mkpath("$(row.label)/")

            if isdir(video)

            #=if isdir(video)

            end

            end=#

predictions=[a;b;c]
Flora
a=glob("../Misc-1/Friends of Flora (1)/Flora */*/preds3_Kahurangi_1-5_2025-02-14.csv")
b=glob("../Misc-1/Friends of Flora (1)/Flora */*/*/preds3_Kahurangi_1-5_2025-02-1*.csv")
c=glob("../Misc-1/Friends of Flora (1)/Flora */*/*/*/preds3_Kahurangi_1-5_2025-02-1*.csv")

for file in list

for file in predictions

make_clips_generic(file, 1, "Kahurangi3_ST", true)

make_clips_generic(file, 1, "MOK_202505-_K1-5_T3", true) #####false

#note of false above, be close to the wavs to minimise length of fname, wont save anything if .. in filename
#=

            mkpath(f)
            outfile = "$f/$name"
            sample = signal[Int(st):Int(en)]
            WAV.wavwrite(sample, "$outfile.wav", Fs = Int(freq))
            image = get_image_from_sample(sample, freq)
            PNGFiles.save("$outfile.png", image)
        end
        print(".")
    end
    print(".")
end
=#
#make_clips_clusters(glob("*/"), "preds3_opensoundscape-kiwi-1.2_2025-07-09.csv")
#make_clips_clusters(glob("*/"), "predsST_opensoundscape-kiwi-1.5_2025-07-09.csv")
# Valid file names follow this pattern: preds3_opensoundscape-kiwi-1.2_2025-07-09.csv
function make_clips_clusters(
    clusters::Vector{String},
    csv_name::String
)::nothing
    x=split(csv_name, "_")
    @assert length(x) == 3
    @assert "preds" occursin x[1]
    @assert "opensoundscape" occursin x[2]
    @assert length(x[3]) == 14
    @assert ".csv" occursin x[3]
    model = split(x[2])
    sensitivity = replace(first(x), "preds"=>"")
    for cluster in clusters
        cd(cluster)
        preds = glob("*/$csv_name")
        for pred in preds
            l=CSV.read(pred, DataFrame) |> x -> names(x)
            for ebird in l
                try
                    make_clips_generic(pred, ebird, model, sensitivity, true)
                catch e
                    @info e
                end
            end
        end
        cd("..")
    end
    return nothing
end
function make_clips_generic(
    preds_path::String,
    label::String, ##column header, ie ebird or "Kiwi"
    model_name::String,
    sensitivity::String
    unique_file_names = true,
)
    # Assumes function run from Kahurangi Data
    #pth = replace(preds_path, "preds-2024-10-21.csv" => "")
    pth0 = split(preds_path, "/")
    length(pth0) > 1 ? (pth = joinpath(pth0[1:end-1]) * "/") : pth = ""
    function assert_detections_present_(df::DataFrame, label::String, preds_path)::DataFrame
        1 in levels(df.label) ? (return df) :
        @error "No detections for label = $label at $preds_path"
    end
    # Load and group data frame by file
    gdf =
        #! format: off
        DataFrames.DataFrame(CSV.File(preds_path)) |>
        x -> assert_not_empty(x, preds_path) |>
        x -> assert_detections_present_(x, label, preds_path) |>
        x -> filter_positives!(x, label) |>
        group_by_file!
        #! format: on
    # Make clip and spectrogram
    for (k, v) in pairs(gdf)
        #file_name = chop(v.file[1], head = 2, tail = 4)
        file_name, extension = path_to_file_string(v.file[1])
        #@info (file_name, extension)
        start_times = v.start_time |> x ->
            convert(Vector{Float64}, x) |>
            #dropmissing(x, disallowmissing = true) |> ######CHECK used to make cobb work. not working anymore, but convert works fine. This happens because the col type of dataframe is Float64? even though no missings, seems to ony happen with doc recorders
            sort
        detections = cluster_detections(start_times)
        isempty(detections) && continue
        signal, freq = WAV.wavread("$pth$(file_name).$(extension)")
        if size(signal, 2) == 2
            signal = (signal[:, 1] + signal[:, 2]) / 2
        end
        if freq > 8000
            signal, freq = resample_to_8000hz(signal, freq)
        end
        freq = freq |> Float32
        length_signal = length(signal)
        for detection in detections
            st, en = calculate_clip_start_end(detection, freq, length_signal)
            if unique_file_names == true
                name = "$file_name-$(Int(floor(st/freq)))-$(Int(ceil(en/freq)))" #leave off path, not necesaray if unique file names
            else
                p = replace(pth, "/" => "--") #replace / with -- including trailing /
                name = "$p$file_name-$(Int(floor(st/freq)))-$(Int(ceil(en/freq)))"
            end
            f = "Clips_$(model_name)_$(today())"