required.jl

using CSV
using DataFrames
using Plots
using EcologicalNetworks
using EcologicalNetworksPlots
using Distributions

function load_data()
    df = CSV.read("WoodEtal_Append1_v2.csv")

    # Separate based on WebScale
    scale = [df[df.WebScale .== u,:] for u in unique(df.WebScale)]

    # select for edge interaction columns only
    edgelist_T = scale[1][:, [:PredTSN, :PreyTSN]]

    # view all entries that contain "san" in PredTSN
    pred_san = edgelist_T[startswith.(edgelist_T.PredTSN, "san"),:]
    # remove "san"
    pred_san.PredTSN = replace.(pred_san.PredTSN, "san" .=> "")
    pred_san
    #view unique predID without san
    unique(pred_san,1)
    edgelist_T.PredTSN == "102" #manually check if in original df

    # remove the "san" from the dataframe itself
     edgelist_T.PredTSN = replace.(edgelist_T.PredTSN, "san" .=> "")

    # Do the same for PreyTSN
    backup2 = copy(edgelist_T)
    old_ids2 = unique(backup2.PreyTSN)
    # keep the indexes before the change
    prey_indexes = startswith.(edgelist_T.PreyTSN, "san")
    # remove the "san" from the dataframe itself
    edgelist_T.PreyTSN = replace.(edgelist_T.PreyTSN, "san" .=> "")
    edgelist_T
    #check unique preyIDs without san
    check_uniques2 = [in(new_id2, old_ids2) for new_id2 in edgelist_T.PreyTSN[prey_indexes]]
    filter(isone, check_uniques2)

    # convert "string" to "integer"
    edgelist_T.PredTSN = parse.(Int64, edgelist_T.PredTSN)
    edgelist_T.PreyTSN = parse.(Int64, edgelist_T.PreyTSN)

    return edgelist_T
end