function [MP, MP_old] = get_mp(fld,monkey, sessions)
% this function concatenates the mp table that was created for individual
% sessions. it has 1 row per channel, so the resulting MP table will have 1
% row per channel per session. this function also removes rows that never
% have a stimulus in their RF. the original (with all rows intact) is
% returned as MP_old. 

datadir = fld.procdatadir;

MP = [];
for sid = 1:length(sessions)
    % session info
    session = sessions{sid};
    q = strsplit(session,'_');
    cdate = q{2};
    block = str2num(q{3}(end-4));

    % load the data 
    session_id = [monkey '_' cdate '_' num2str(block)];
    sessiondir = fullfile(datadir, monkey, session_id);
    try 
        load(fullfile(sessiondir, [session_id '_channels.mat']));
    catch
        continue
    end
    %load([sessiondir session_id '_LUT.mat']);
    
    % add some session data 
    session_id = repmat(sid,size(mp,1),1);
    date = repmat({cdate},size(mp,1),1);
    block = repmat(block,size(mp,1),1);
    
    mp = addvars(mp,session_id,date,block,'After','monkey_id');
    
    % concatenate
    MP = [MP; mp];
end

%% count rows that have empty values for Tar
% in constructing MP, we added all channels that were available. each
% channel is given a value for RFPos, which is available in MP.rfpos. this
% value is technically between 1 and 6 and denotes the stimulus positions
% that falls in the RF, but in practice, it's only ever 1 or empty. 
% this means that channels with RFPos=1 have a stimulus in their RF. we can
% exclude channels that have an empty RFPos. 
emptyrows = cellfun('isempty',MP.Tar);
disp(['There are ' num2str(sum(emptyrows)) ' empty rows']);

MP_old = MP; 

MP(emptyrows,:) = [];
disp(['There are ' num2str(length(unique(MP.chan_id))) ' usable channels for this monkey']);