team_testing_code.m

function [binary_output, probability_output] = team_testing_code(model, record, data_folder, index, verbose)
%==========================================================================
% Edit this function to add your team's testing code. 
%==========================================================================

% Extract identification details from the record
    patient_id = record.BidsFolder;
    session_id = record.SessionID;
    
    if isnumeric(session_id)
        session_str = num2str(session_id);
    else
        session_str = char(session_id);
    end
    
    % Load physionlogical and algorithmic files as raw data matrices and header structures
    [row, phys_sig, phys_hdr, algo_sig, algo_hdr] = load_challenge_data(data_folder, record, index);
   
    % Pack physiological data symmetrically
    phys_data.signals = phys_sig;
    phys_data.labels  = phys_hdr.Label;
    phys_data.hdr     = phys_hdr; 
    
    % Pack algorithmic data identically to handle hospital variations!
    algo_data.signals = algo_sig;
    algo_data.labels  = algo_hdr.Label;
    % --- ADDED FOR BIOSIG TIMING FIX ---
    % Store total_seconds directly within algo_data so get_features doesn't rely on vector length
    if isfield(algo_hdr, 'NRec') && isfield(algo_hdr, 'Dur')
        algo_data.total_seconds = algo_hdr.NRec * algo_hdr.Dur;
    else
        algo_data.total_seconds = 0; % Fallback
    end
    % -----------------------------------
    
    % Extract features
    features = get_features(row, phys_data, algo_data);
    % if verbose
    %     fprintf('%f,',features);
    % end
    
    % =====================================================================
    % Run Classifier Inference
    % =====================================================================
    % if verbose
    %     fprintf('           Running model prediction\n');
    % end
    
    % Extract the object from the model structure
    actual_tree_model = model.model_outcome;
    
    % Run prediction on the features
    [predicted_labels, posterior_scores] = predict(actual_tree_model, features);
    
    % Format outputs to match expected numeric types
    binary_output      = int32(str2double(predicted_labels{1}));
    probability_output = double(posterior_scores(1, 2));
    
end