+
−% Test ao/removeVal for:
+
−% - functionality
+
−%
+
−% M Hueller 10-03-10
+
−%
+
−% $Id: test_ao_removeVal.m,v 1.1 2010/03/13 06:36:20 mauro Exp $
+
−%
+
−
+
−%% Start with an AO with no "bad" data in it
+
−a_orig = ao(plist(...
+
−  'waveform', 'sinewave', ...
+
−  'Nsecs', 100, ...
+
−  'fs', 1, ...
+
−  'f', 0.02, ...
+
−  'yunits', 'V', ...
+
−  't0', '2010-03-10 20:00:00.000' ...
+
−  )) + 2;
+
−a_orig.setName('original');
+
−
+
−%% 1) Test the case where we remove only Inf, removing data
+
−value = Inf;
+
−% Prepare a list of index where we add the "bad" points
+
−N   = randi(10, 1);
+
−bad_index   = randi(numel(a_orig.x), N, 1);
+
−
+
−% Mix the AO with the "bad" values
+
−y_orig = a_orig.y();
+
−y_bad = y_orig;
+
−y_bad(bad_index)   = value;
+
−a_bad = a_orig.setY(y_bad);
+
−a_bad.setName('corrupted');
+
−
+
−% Run the ao/removeVal method
+
−a_clean = a_bad.removeVal(plist('value', value));
+
−a_clean.setName('cleaned');
+
−
+
−% Compare against the original:
+
−% units
+
−if ~eq(a_clean.yunits, a_bad.yunits)
+
−  disp('Warning: wrong yunits!');
+
−end
+
−if ~eq(a_clean.xunits, a_bad.xunits)
+
−  disp('Warning: wrong xunits!');
+
−end
+
−
+
−% t0
+
−if ~eq(a_clean.t0, a_bad.t0)
+
−  disp('Warning: wrong t0!');
+
−end
+
−
+
−index   = true(size(a_bad.x));
+
−index(bad_index) = false;
+
−
+
−% "good" values: x
+
−x_clean = a_clean.x;
+
−x_bad   = a_bad.x;
+
−if ~eq(x_bad(index), x_clean)
+
−  disp('Warning: wrong x values!');
+
−end
+
−
+
−% "good" values: y
+
−y_clean = a_clean.y;
+
−y_bad   = a_bad.y;
+
−if ~eq(y_bad(index), y_clean)
+
−  disp('Warning: wrong y values!');
+
−end
+
−
+
−%% 2) Test the case where we remove only NaN, removing data
+
−value = NaN;
+
−% Prepare a list of index where we add the "bad" points
+
−N   = randi(10, 1);
+
−bad_index   = randi(numel(a_orig.x), N, 1);
+
−
+
−% Mix the AO with the "bad" values
+
−y_orig = a_orig.y();
+
−y_bad = y_orig;
+
−y_bad(bad_index)   = value;
+
−a_bad = a_orig.setY(y_bad);
+
−a_bad.setName('corrupted');
+
−
+
−% Run the ao/removeVal method
+
−a_clean = a_bad.removeVal(plist('value', value));
+
−a_clean.setName('cleaned');
+
−
+
−% Compare against the original:
+
−% units
+
−if ~eq(a_clean.yunits, a_bad.yunits)
+
−  disp('Warning: wrong yunits!');
+
−end
+
−if ~eq(a_clean.xunits, a_bad.xunits)
+
−  disp('Warning: wrong xunits!');
+
−end
+
−
+
−% t0
+
−if ~eq(a_clean.t0, a_bad.t0)
+
−  disp('Warning: wrong t0!');
+
−end
+
−
+
−index   = true(size(a_bad.x));
+
−index(bad_index) = false;
+
−
+
−% "good" values: x
+
−x_clean = a_clean.x;
+
−x_bad   = a_bad.x;
+
−if ~eq(x_bad(index), x_clean)
+
−  disp('Warning: wrong x values!');
+
−end
+
−
+
−% "good" values: y
+
−y_clean = a_clean.y;
+
−y_bad   = a_bad.y;
+
−if ~eq(y_bad(index), y_clean)
+
−  disp('Warning: wrong y values!');
+
−end
+
−
+
−
+
−%% 3) Test the case where we remove only 0s, removing data
+
−value = 0;
+
−% Prepare a list of index where we add the "bad" points
+
−N   = randi(10, 1);
+
−bad_index   = randi(numel(a_orig.x), N, 1);
+
−
+
−% Mix the AO with the "bad" values
+
−y_orig = a_orig.y();
+
−y_bad = y_orig;
+
−y_bad(bad_index)   = value;
+
−a_bad = a_orig.setY(y_bad);
+
−a_bad.setName('corrupted');
+
−
+
−% Run the ao/removeVal method
+
−a_clean = a_bad.removeVal(plist('value', value));
+
−a_clean.setName('cleaned');
+
−
+
−% Compare against the original:
+
−% units
+
−if ~eq(a_clean.yunits, a_bad.yunits)
+
−  disp('Warning: wrong yunits!');
+
−end
+
−if ~eq(a_clean.xunits, a_bad.xunits)
+
−  disp('Warning: wrong xunits!');
+
−end
+
−
+
−% t0
+
−if ~eq(a_clean.t0, a_bad.t0)
+
−  disp('Warning: wrong t0!');
+
−end
+
−
+
−index   = true(size(a_bad.x));
+
−index(bad_index) = false;
+
−
+
−% "good" values: x
+
−x_clean = a_clean.x;
+
−x_bad   = a_bad.x;
+
−if ~eq(x_bad(index), x_clean)
+
−  disp('Warning: wrong x values!');
+
−end
+
−
+
−% "good" values: y
+
−y_clean = a_clean.y;
+
−y_bad   = a_bad.y;
+
−if ~eq(y_bad(index), y_clean)
+
−  disp('Warning: wrong y values!');
+
−end
+
−
+
−
+
−%% 4) Test the case where we remove Inf AND NaN AND 0s, removing data
+
−% Prepare a list of index where we add the "bad" points
+
−N_nan   = randi(10, 1);
+
−N_inf   = randi(10, 1);
+
−N_zeros = randi(10, 1);
+
−nan_index   = randi(numel(a_orig.x), N_nan, 1);
+
−inf_index   = randi(numel(a_orig.x), N_inf, 1);
+
−zeros_index = randi(numel(a_orig.x), N_zeros, 1);
+
−
+
−% Mix the AO with NaN, Inf, and 0s inside
+
−y_orig = a_orig.y();
+
−y_bad = y_orig;
+
−y_bad(nan_index)   = NaN;
+
−y_bad(inf_index)   = Inf;
+
−y_bad(zeros_index) = 0;
+
−a_bad = a_orig.setY(y_bad);
+
−a_bad.setName('corrupted');
+
−
+
−% Run the ao/removeVal method
+
−value = [Inf NaN 0];
+
−a_clean = a_bad.removeVal(plist('value', value));
+
−a_clean.setName('cleaned');
+
−
+
−% Compare against the original
+
−% units
+
−if ~eq(a_clean.yunits, a_bad.yunits)
+
−  disp('Warning: wrong yunits!');
+
−end
+
−if ~eq(a_clean.xunits, a_bad.xunits)
+
−  disp('Warning: wrong xunits!');
+
−end
+
−
+
−% t0
+
−if ~eq(a_clean.t0, a_bad.t0)
+
−  disp('Warning: wrong t0!');
+
−end
+
−
+
−index   = true(size(a_bad.x));
+
−index(nan_index) = false;
+
−index(inf_index) = false;
+
−index(zeros_index) = false;
+
−
+
−% "good" values: x
+
−x_clean = a_clean.x;
+
−x_bad   = a_bad.x;
+
−if ~eq(x_bad(index), x_clean)
+
−  disp('Warning: wrong x values!');
+
−end
+
−
+
−% "good" values: y
+
−y_clean = a_clean.y;
+
−y_bad   = a_bad.y;
+
−if ~eq(y_bad(index), y_clean)
+
−  disp('Warning: wrong y values!');
+
−end
+
−
+
−% plots the original, "bad", and "clean" data
+
−iplot(a_orig, a_bad, a_clean, plist('LineStyles', {'-', '-', 'None'}, 'Markers', {'.', '.', 'o'}));
+
−
+
−%% 5) Test the case where we remove Inf AND NaN AND 0s, interpolating data
+
−% Prepare a list of index where we add the "bad" points
+
−N_nan   = randi(10, 1);
+
−N_inf   = randi(10, 1);
+
−N_zeros = randi(10, 1);
+
−nan_index   = randi(numel(a_orig.x), N_nan, 1);
+
−inf_index   = randi(numel(a_orig.x), N_inf, 1);
+
−zeros_index = randi(numel(a_orig.x), N_zeros, 1);
+
−
+
−% Mix the AO with NaN, Inf, and 0s inside
+
−y_orig = a_orig.y();
+
−y_bad = y_orig;
+
−y_bad(nan_index)   = NaN;
+
−y_bad(inf_index)   = Inf;
+
−y_bad(zeros_index) = 0;
+
−a_bad = a_orig.setY(y_bad);
+
−a_bad.setName('corrupted');
+
−
+
−% Run the ao/removeVal method
+
−value = [Inf NaN 0];
+
−a_clean = a_bad.removeVal(plist('value', value, 'method', 'interp'));
+
−a_clean.setName('cleaned');
+
−
+
−% Compare against the original
+
−% units
+
−if ~eq(a_clean.yunits, a_bad.yunits)
+
−  disp('Warning: wrong yunits!');
+
−end
+
−if ~eq(a_clean.xunits, a_bad.xunits)
+
−  disp('Warning: wrong xunits!');
+
−end
+
−
+
−% t0
+
−if ~eq(a_clean.t0, a_bad.t0)
+
−  disp('Warning: wrong t0!');
+
−end
+
−
+
−% plots the original, "bad", and "clean" data
+
−iplot(a_orig, a_bad, a_clean, plist('LineStyles', {'-', '-', 'None'}, 'Markers', {'.', '.', 'o'}));
+
−