function tex_string = rankingReportTeX(a_bundle, crit_bundle, crit_db, props)
% rankingReportTeX - Generates a report by comparing a_bundle with the given match criteria, crit_db from crit_bundle.
%
% Usage:
% tex_string = rankingReportTeX(a_bundle, crit_bundle, crit_db, props)
%
% Description:
% Generates a LaTeX document with:
% - (optional) Raw traces compared with some best matches at different distances
% - Values of some top matching a_db rows and match errors in a floating table.
% - colored-plot of measure errors for some top matches.
% - Parameter distributions of 50 best matches as a bar graph.
%
% Parameters:
% a_bundle: A dataset_db_bundle object that contains the DB to compare rows from.
% crit_bundle: A dataset_db_bundle object that contains the criterion dataset.
% crit_db: A tests_db object holding the match criterion tests and STDs
% which can be created with matchingRow.
% props: A structure with any optional properties.
% caption: Identification of the criterion db (not needed/used?).
% num_matches: Number of best matches to display (default=10).
% rotate: Rotation angle for best matches table (default=90).
%
% Returns:
% tex_string: LaTeX document string.
%
% See also: displayRowsTeX
%
% $Id$
%
% Author: Cengiz Gunay <cgunay@emory.edu>, 2005/12/13
% Copyright (c) 2007 Cengiz Gunay <cengique@users.sf.net>.
% This work is licensed under the Academic Free License ("AFL")
% v. 3.0. To view a copy of this license, please look at the COPYING
% file distributed with this software or visit
% http://opensource.org/licenses/afl-3.0.php.
if ~ exist('props', 'var')
props = struct([]);
end
tex_string = '';
a_ranked_db = rankMatching(a_bundle.joined_db, crit_db);
joined_db = joinOriginal(a_ranked_db);
ranked_num_rows = dbsize(a_ranked_db, 1);
% LaTeX likes '_' to be '\_'
a_db_id = strrep(lower(get(a_bundle.joined_db, 'id')), '_', '\_');
crit_db_id = strrep(lower(get(crit_db, 'id')), '_', '\_');
if ranked_num_rows > 0
% Display raw data traces from dataset
% prepare a landscape figure with two rows
% one row for original and 5 matching data traces
% second row for original and 5 matching spike shapes
% TODO: This should be in a *_profile class
crit_trace_d100 = ctFromRows(crit_bundle, crit_db, 100);
crit_trace_h100 = ctFromRows(crit_bundle, crit_db, -100);
if isempty(crit_trace_h100) || isempty(crit_trace_d100)
error(['Cannot find one of 100 or -100 pA cip traces in ' get(crit_bundle, 'id') '.']);
end
crit_trace_id = strrep(get(crit_trace_d100(1), 'id'), '_', '\_');
num_plots = 5;
trace_d100_plots = cell(1, num_plots);
trace_h100_plots = cell(1, num_plots);
for plot_num=1:num_plots
rank_num = (plot_num - 1) * 10 + 1;
trial_num = get(onlyRowsTests(joined_db, rank_num , 'trial'), 'data');
if plot_num > 1
sup_props = struct('noLegends', 1);
crit_traces = 1;
else
sup_props = struct;
crit_traces = ':';
end
trace_d100_plots{plot_num} = ...
superposePlots([plotData(crit_trace_d100(crit_traces)), ...
plotData(ctFromRows(a_bundle, trial_num, 100))], {}, ...
['Rank ' num2str(rank_num) ', t' num2str(trial_num)], ...
'plot', sup_props);
trace_h100_plots{plot_num} = ...
superposePlots([plotData(crit_trace_h100(crit_traces)), ...
plotData(ctFromRows(a_bundle, trial_num, -100))], {}, '', ...
'plot', sup_props);
end
% Make a full figure with the best matching guy
plotFigure(plot_stack([trace_d100_plots{1}, trace_h100_plots{1}], ...
[0 3000 -150 80], 'y', ...
['The best match to ' crit_trace_id ], ...
struct('xLabelsPos', 'bottom', 'titlesPos', 'none')));
best_filename = properTeXFilename([crit_db_id ' - best matching model from ' a_db_id ]);
orient tall; print('-depsc2', [ best_filename '.eps' ] );
horiz_props = struct('titlesPos', 'all', 'yLabelsPos', 'left', 'xLabelsPos', 'none');
d100_row_plot = plot_stack([trace_d100_plots{2:num_plots}], [0 3000 -80 80], 'x', '', ... %+100 pA CIP
mergeStructs(struct('xLabelsPos', 'none'), horiz_props));
h100_row_plot = plot_stack([trace_h100_plots{2:num_plots}], [0 3000 -150 80], 'x', '-100 pA CIP', ...
mergeStructs(struct('titlesPos', 'none'), horiz_props));
plotFigure(plot_stack({d100_row_plot, h100_row_plot}, [], 'y', ...
['Best matches to ' crit_trace_id ], ...
struct('titlesPos', 'none')));
filename = properTeXFilename([ crit_db_id ' - top matching models from ' a_db_id ]);
print('-depsc2', [ filename '.eps' ] );
% Put the best match in a figure float
short_caption = [ 'Best matching model to ' crit_db_id ...
' (' crit_trace_id ').' ];
caption = [ short_caption ...
' All available raw traces from the criterion cell are shown.' ];
tex_string = [ tex_string ...
TeXfloat([ '\includegraphics{' best_filename '}' ], ...
caption, struct('floatType', 'figure', ...
'center', 1, ...
'height', '.9\textheight', ...
'shortCaption', short_caption)) ];
% Put other matches in a figure float
% TODO: indicate distances of best and furthest matches
short_caption = [ 'Raw traces of the ranked to the ' crit_db_id ...
' (' crit_trace_id ').' ];
caption = [ short_caption ...
' Traces are taken from 5 equidistant matches from the best' ...
' 50 ranks from ' a_db_id '.' ...
' Criterion cell trace is superposed with each model trace.'];
tex_string = [ tex_string ...
TeXfloat([ '\includegraphics{' filename '}' ], ...
caption, struct('floatType', 'figure', ...
'center', 1, ...
'rotate', 90, ...
'height', '.9\textheight', ...
'shortCaption', short_caption)) ...
'\clearpage%' sprintf('\n') ];
% Display values of 10 best matches
if isfield(props, 'num_matches')
num_best = props.num_matches;
else
num_best = 13;
end
top_ranks = onlyRowsTests(a_ranked_db, 1:min(num_best, ranked_num_rows), ':', ':');
short_caption = [ a_db_id ' ranked to the ' crit_db_id '.' ];
caption = [ short_caption ...
' Only ' num2str(num_best) ' best matches are shown.' ];
tex_string = [ tex_string displayRowsTeX(top_ranks, caption, ...
struct('shortCaption', short_caption, ...
'rotate', 0)) ];
% Display colored-plot of top 50 matches
num_best = min(50, ranked_num_rows);
plotFigure(plotRowErrors(a_ranked_db, 1:num_best));
% Save it as a picture
filename = properTeXFilename([ crit_db_id ' - colorgraph of ' num2str(num_best) ...
' top matching models from ' a_db_id ]);
orient tall; print('-depsc2', [ filename '.eps' ]);
% Put it in a figure float
% TODO: indicate distances of best and furthest matches
short_caption = [ 'Individual measure distances color-coded for top matches of ' ...
a_db_id ' ranked to the ' crit_db_id ' (' crit_trace_id ').'];
caption = [ short_caption ...
' Increasing distance represented with colors starting from ' ...
'blue to red. Dark blue=0 STD; light blue=1xSTD; yellow=2xSTD; ' ...
'and red=3xSTD difference.' ];
tex_string = [ tex_string ...
TeXfloat([ '\includegraphics{' filename '}' ], ...
caption, struct('floatType', 'figure', ...
'center', 1, ...
'width', '\textwidth', ...
'shortCaption', short_caption)) ];
% Display sorted colored-plot of top 50 matches
plotFigure(plotRowErrors(a_ranked_db, 1:num_best, struct('sortMeasures', 1)));
% Save it as a picture
filename = properTeXFilename([ crit_db_id ' - sorted colorgraph of ' num2str(num_best) ...
' top matching models from ' a_db_id ]);
orient tall; print('-depsc2', [ filename '.eps' ]);
% Put it in a figure float
% TODO: indicate distances of best and furthest matches
short_caption = [ 'Sorted individual measure distances color-coded for top matches of ' ...
a_db_id ' ranked to the ' crit_db_id ' (' crit_trace_id ').'];
caption = [ short_caption ...
' Increasing distance represented with colors starting from ' ...
'blue to red. Dark blue=0 STD; light blue=1xSTD; yellow=2xSTD; ' ...
'and red=3xSTD difference. Measures sorted with overall match quality. ' ];
tex_string = [ tex_string ...
TeXfloat([ '\includegraphics{' filename '}' ], ...
caption, struct('floatType', 'figure', ...
'center', 1, ...
'width', '\textwidth', ...
'shortCaption', short_caption)) ];
% Display parameter distributions of 50 best matches
num_best = 50;
top_ranks = onlyRowsTests(joined_db, 1:min(num_best, ranked_num_rows), ':', ':');
all_param_cols = true(1, get(top_ranks, 'num_params'));
all_param_cols(tests2cols(top_ranks, 'trial')) = false;
p_hists = paramsHists(onlyRowsTests(top_ranks, ':', all_param_cols));
plotFigure(plot_stack(num2cell(plotEqSpaced(p_hists)), [], 'x', ...
['Parameter distribution histograms of ' num2str(num_best) ...
' best matches' ], ...
struct('yLabelsPos', 'left', 'titlesPos', 'none')));
% Save it as a picture
filename = properTeXFilename([ crit_db_id ' - params distribution of top 50 matches to ' ...
a_db_id ]);
print('-depsc2', [ filename '.eps' ]);
% Put it in a figure float
% TODO: indicate distances of best and furthest matches
short_caption = [ 'Parameter distributions of the best ranked to the ' crit_db_id '.' ];
caption = [ short_caption ...
' Only ' num2str(num_best) ' best matches from ' a_db_id ...
' are taken.' ];
tex_string = [ tex_string ...
TeXfloat([ '\includegraphics{' filename '}' ], ...
caption, struct('floatType', 'figure', ...
'center', 1, ...
'width', '0.9\textwidth', ...
'shortCaption', short_caption)) ...
'\clearpage%' sprintf('\n')];
end