avianz_of_raven,

    avianz_of_raven, #not working right, 1 file per label I think

    one_hot_labels,

# Raven selections.txt to AviaNZ .data

# ASSUMES 60 second files, also wav and txt file in same folder
# Raven selections.txt to AviaNZ .data 

function avianz_of_raven(f::String)

function avianz_of_raven(f::String) #not working right, check

        duration = Float64[],

    dur = data[1]["Duration"]

                    duration = dur,

# if specified folder must include trailing /
function label_summary(folder::String = "", avianz = true)

# if specified folder must include trailing /, can be "" for current folder
function label_summary(folder::String, avianz::Bool = true)

            duration = Float64[],

        select!(df, [:File, :start_time, :end_time, :low_f, :high_f, :Species])

        select!(df, [:File, :duration, :start_time, :end_time, :low_f, :high_f, :Species])

end
#allow 0.2 overlap each side
#round end_time down to nearest 5
function et(end_time::Float64)
    end_time % 5 > 0.2 ? (c0 = ceil(end_time / 5) * 5) : c0 = floor(end_time / 5) * 5
    return c0
end
#allow 0.2 overlap each side
#round start_time down to nearest 5
function st(start_time::Float64)
    start_time % 5 < 4.8 ? (f = floor(start_time / 5) * 5) : f = ceil(start_time / 5) * 5
    return f

#labels must be a df loaded from label_summary run over avianz data (not raven)
function one_hot_labels(labels::DataFrame)
    gdf = groupby(labels, :File)
    vdf = []
    for group in gdf
        dur = first(group.duration)
        nrows = dur ÷ 5
        seil = nrows * 5
        df = DataFrame(
            file = [first(group.File) for x = 1:nrows],
            start_time = collect(0:5:seil-1),
            end_time = collect(5:5:seil),
        )
        for row in eachrow(group)
            fst = st(row.start_time)
            @assert fst >= 0
            lst0 = et(row.end_time)
            #end time must not be greater than duration
            lst0 > dur ? lst = lst0 - 5 : lst = lst0
            @assert lst <= dur
            f_idx = fst ÷ 5 + 1 |> Int
            l_idx = lst ÷ 5 |> Int
            vect = [false for x = 1:nrows]
            for idx = f_idx:l_idx
                vect[idx] = true
            end
            col_name = row.Species
            df[!, Symbol(col_name)] = vect
        end
        #(names(df) |> length) > 4 && println(df)
        push!(vdf, df)
    end
    cdf1 = reduce(
        (x, y) ->
            outerjoin(x, y, matchmissing = :equal, on = intersect(names(x), names(y))),
        vdf,
    )
    cdf2 = coalesce.(cdf1, false)
end