function [acc,confusm,lhood,models,sy,gt] = trntest_1fold(trainset, testset, ...
                                                    datapath, dataext, ...
                                                    ngmm, nsamp, dims, ...
                                                    beatlevel, verb)
% [acc,confusm,lhood,models,sy,gt] = trntest_1fold(trainset, testset, 
%                                            datapath, dataext, ngmm,
%                                            nsamp, dims,
%                                            beatlevel, verb)
%
% Example baseline artist ID task.  <trainset> is the name of a file listing 
% all the training examples (with corresponding label file assumed).
% <testset> is the name of a file that lists all the test files.
% <datapath> is a path to prefix to all data file names.
% <dataext> is an extension to append to all data file names.
% <ngmm> (1) is the number of gaussians to use in each mixture model
% (<ngmm>==1 is a special case for full-covariance single gaussians)
% <nsamp> (1000) is the number of randomly-selected samples to train with.
% <dims> (all) is vector of columns in the feature vectors to use.
% <beatlevel> (0) says to load beat-level features instead of HTK files.
% <verb> (0) selects verbose progress reporting.
% <acc> returns the overall accuracy (0..1)
% <confusm> is a confusion count matrix
% <lhood> returns the raw scores for all tracks across all models
% <models> is an array of per-label model structures
% <sy> is a vector of the raw system classes (integers), and 
% <gt> is the corresponding ground truth values.  acc = mean(sy==gt);
%
% 2007-04-04 Dan Ellis dpwe@ee.columbia.edu
% $Header: /homes/drspeech/data/uspop2002/baseline/RCS/do_expt.m,v 1.1 2007/04/06 15:44:05 dpwe Exp dpwe $

%   Copyright (c) 2007 Columbia University.
% 
%   This file is part of LabROSA-artist20-baseline
% 
%   artist20-baseline is free software; you can redistribute it and/or modify
%   it under the terms of the GNU General Public License version 2 as
%   published by the Free Software Foundation.
% 
%   artist20-baseline is distributed in the hope that it will be useful, but
%   WITHOUT ANY WARRANTY; without even the implied warranty of
%   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
%   General Public License for more details.
% 
%   You should have received a copy of the GNU General Public License
%   along with artist20-baseline; if not, write to the Free Software
%   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
%   02110-1301 USA
% 
%   See the file "COPYING" for the text of the license.

%%%% Input parameters
if nargin < 1; trainset = 'a20-trn-tracks.list'; end
if nargin < 2; testset = 'a20-val-tracks.list'; end
% How many mixture components?
if nargin < 5; ngmm = 1; end
% Train on how many samples per file?
if nargin < 6; nsamp = 1000; end
% Using which cepstral dimensions?
if nargin < 7; dims = []; end
% load beat-level feature files?
if nargin < 8; beatlevel = 0; end
% progress messages?
if nargin < 9; verb = 0; end

if beatlevel
  % different defaults
  if nargin < 3; datapath = '../beatmfcc'; end
  if nargin < 4; dataext = '.mat'; end
else
  if nargin < 3; datapath = '../mfccs'; end
  if nargin < 4; dataext = '.htk'; end
end

% plot confusion matrix?
doplot = 0;
if verb
  doplot = 1;
  % seed the RNG when doing a verbose run for consistent results
  rand('state', 0);
end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Train models
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

train_files = textread(trainset, '%s\n');
train_labs = labelsfor(trainset);

% Make the train files into full relative paths
for i = 1:length(train_files)
  train_files{i} = fullfile(datapath, [train_files{i}, dataext]);
end

% List of all unique labels
ulabs = unique(train_labs);
nlabs = length(ulabs);

% Train models one by one

for model = 1:nlabs
  if verb; disp(['training for ',ulabs{model},' ...']); end

  % Select filenames that have this label as ground truth
  files = train_files(strcmp(train_labs, ulabs{model}));
    
  models(model) = train_model(files, ngmm, nsamp, dims, beatlevel);
  
end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Test models
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

test_files = textread(testset,'%s\n');
test_labs = labelsfor(testset);

ntest = length(test_files);

for file = 1:ntest

  testfile = test_files{file};
  
  if verb; disp(['testing ',testfile,'...']); end
  
  lhood(:,file) = test_track(fullfile(datapath,[testfile,dataext]), ...
                             models, nsamp, dims, beatlevel);
  
  % Which model should it have been according to ground truth label?
  gt(file) = find(strcmp(ulabs, test_labs{file}));
  
end

% Which is the most likely model for each test item?
[maxv,sy] = max(lhood);

% Overall error rate
acc = mean(sy==gt);
if verb; disp(['Classification accuracy = ', num2str(100*acc),'%']); end

% Matrix of which track was classified to which class, and ground truth;
cm = 0*lhood;
gtm = 0*lhood;
for i = 1:ntest
  cm(sy(i),i) = 1;
  gtm(gt(i),i) = 1;
end
% So confusion matrix
confusm = gtm*cm';
% rows are true class, columns are reported (model) class

if doplot
  trntest_plot_results(lhood, test_labs, ulabs, confusm);
end