function [trainX, valX, testX, trainY, valY, testY] = split_data(X, Y, ratios)
% SPLIT_DATA  Split dataset into train/validation/test sets (stratified)
%
%   [trainX, valX, testX, trainY, valY, testY] = split(X, Y, ratios)
%
%   ratios : [trainRatio, valRatio, testRatio] (e.g. [0.6 0.2 0.2])
%
%   Stratified split ensures class proportions remain consistent.

    if nargin < 3
        ratios = [0.6 0.2 0.2];
    end
    assert(abs(sum(ratios) - 1) < 1e-6, 'Ratios must sum to 1.');

    n = size(X,1);
    classes = unique(Y);
    idxTrain = []; idxVal = []; idxTest = [];

    for c = classes'
        idx = find(Y == c);
        idx = idx(randperm(length(idx)));  % randomize within class

        nTrain = round(ratios(1)*length(idx));
        nVal   = round(ratios(2)*length(idx));

        idxTrain = [idxTrain; idx(1:nTrain)];
        idxVal   = [idxVal; idx(nTrain+1:nTrain+nVal)];
        idxTest  = [idxTest; idx(nTrain+nVal+1:end)];
    end

    % Shuffle within each subset to mix classes
    idxTrain = idxTrain(randperm(length(idxTrain)));
    idxVal   = idxVal(randperm(length(idxVal)));
    idxTest  = idxTest(randperm(length(idxTest)));

    trainX = X(idxTrain,:);  trainY = Y(idxTrain);
    valX   = X(idxVal,:);    valY   = Y(idxVal);
    testX  = X(idxTest,:);   testY  = Y(idxTest);
end