HARMONYRules.m

function [data_table_out,OutDict] = HARMONYRules(data_table,Visits,Dictionnary)
%% DESCRIPTION
% Function to translate PRONIA variables to HARMONY ones. See
% /opt/PRONIASoftware/Developpment/AnneTesting/NDL_code/DataAllCron/DataQuery/DataDictionnary/HARMONY/HARMONY Clinical measures data dictionary Vipar 091720.xlsx and
% and
% /opt/PRONIASoftware/Developpment/AnneTesting/NDL_code/DataAllCron/DataQuery/DataDictionnary/HARMONY/Translator_RS_Oct2020.xlsx
% for more details and instruction on how the items are translated
%% INPUTS
% - data_table = table. Table containing all items necessary for the
% translation and recoding of the pronia variables to harmony variables
% - Visits = cell array of string. Currently not necessary, but if
% longitudinal data are required, this will become handy
% - Dictionnary = string. Full path to the dictionnary that contains the
% translation from pronia to harmony. Ideally the header of the table
% should be like the following: 
% 'ItemName'            % new name of the item (in that case harmony)
% 'OriginalName'        % original name from the PortalName (in that case pronia)
%% OUTPUT
% - data_table_out = table. Items recoded according to the rules in this
% function
% -Dict            = table. Dictionnary will be removed in the future. It
% is for the ROIs MRI
%% CREDENTIALS
% Created by Rachele Sanfelici, 10/2020, @PRONIA
% Modified by AnneRuef 28-Oct-2020 @PRONIA
% Doc improved by Anne Ruef 22-Apr-2021
%%
% inputs checker
if nargin == 0
    error('HARMONYRules: not enough inputs arguments')
end
if exist('Visits','var')
   Visits = {''}; % this is hard coded
   % not used yet but it will be if longitudinal data is required
end
% read the dictionnary
if exist('Dictionnary','var')
    if contains(Dictionnary,'xls')
        [~,~,Dict] = xlsread(Dictionnary);
        TempHead = Dict(1,:);
        Dict = cell2table(Dict(2:end,:));
        Dict.Properties.VariableNames = TempHead;
    elseif contains(Dictionnary,'csv')
        Dict = readtable(Dictionnary);
    else
        error('HARMONYRules: unknown dictionnary file type')
    end
end
%%
% initialise the table
data_table_out = table;
%%
% extract the consortium == 3 this is PRONIA (for other sites change that)
data_table_out.consortium = repmat(3,size(data_table,1),1);
%%
% extract the site and psn
% NOTE there is certainly a more elegant way to code this, please feel free
% to change that
% 20=“PRONIA/Bar�?,
% 21=“PRONIA/Basel�?,
% 22=“PRONIA/Birmingham�?,
% 23=“PRONIA/Cologne�?,
% 24=“PRONIA/Dusseldorf�?,
% 25=“PRONIA/Milan�?,
% 26=“PRONIA/Munich�?,
% 27=“PRONIA/Munster�?,
% 28=“PRONIA/Turku�?,
% 29=“PRONIA/Udine�?,
sites = {'UBARI','UBS','Uni BHAM','UKK','UDUS','MilanNig','LMU','UMUENS','Uni Turku','Uni Udine'};
s = 20:29;
data_table_out.site = nan(size(data_table,1),1);
data_table_out.subject_id = nan(size(data_table,1),1);
for i=1:size(sites,2)
    data_table_out.site(strcmp(data_table.INSTITUTE_SHORTNAME,sites{1,i})) = s(i);
    % create a random number of the participant
    TempRand = 1:999;
    [~,Idx] = sort(rand(numel(TempRand),1));
    RandNb = TempRand(Idx(1:sum(strcmp(data_table.INSTITUTE_SHORTNAME,sites{1,i}))));
    data_table_out.subject_id(strcmp(data_table.INSTITUTE_SHORTNAME,sites{1,i})) = RandNb;
end
% create the subject id 
for i = 1:size(data_table,1)
    % add 0 in case the randomly generated number is different from 4
    if numel(num2str(data_table_out.subject_id(i))) == 4
        data_table_out.subject_id(i) =  str2double(['3',num2str(data_table_out.site(i)),num2str(data_table_out.subject_id(i))]);
    elseif numel(num2str(data_table_out.subject_id(i))) == 3
        data_table_out.subject_id(i) =  str2double(['3',num2str(data_table_out.site(i)),num2str(data_table_out.subject_id(i)),'0']);
    elseif numel(num2str(data_table_out.subject_id(i))) == 2
        data_table_out.subject_id(i) = str2double(['3',num2str(data_table_out.site(i)),num2str(data_table_out.subject_id(i)),'00']);
    elseif numel(num2str(data_table_out.subject_id(i))) == 1
        data_table_out.subject_id(i) = str2double(['3',num2str(data_table_out.site(i)),num2str(data_table_out.subject_id(i)),'000']);
    else
        data_table_out.subject_id(i) = str2double(['3',num2str(data_table_out.site(i)),num2str(data_table_out.subject_id(i)),'0000']);
    end
end
%%
% extract the subject type or studygroup. NOTE How do I code for ROD ?
% recode the study group
data_table_out.subjecttype = nan(size(data_table,1),1);
data_table_out.subjecttype(strcmp(data_table.Studygroup,'HC'),1)  = 1;
data_table_out.subjecttype(strcmp(data_table.Studygroup,'CHR'),1) = 2;
%%
% extract age in months and assessment date
data_table_out.assessment_age = nan(size(data_table,1),1);
birthdate_var = data_table.BIRTHDATE;
data_table_out.assessment_date = nan(size(data_table,1),1);
data_table_out.assessment_date = data_table.Study_date_sMRI_T0;
%% TODO need to fix the query to export all ALL_QUEST_Examination_DATE for all questionnaires
quests={'SIPS_P','SPI_A_COGDIS','GAF','GF','TREATMENT','DROPOUT'};
visits = {'Screening','T0'};
for q = 1:numel(quests)
    for v = 1:numel(visits)
    if ismember({['ALL_QUEST_Examination_Date_',quests{q},'_',visits{v}]},data_table.Properties.VariableNames)
        data_table_out.assessment_date(cellfun(@isempty,data_table_out.assessment_date)) = data_table.(['ALL_QUEST_Examination_Date_',quests{q},'_',visits{v}])(cellfun(@isempty,MRI_var));
    end
    end
end

% create a mask for empty values
Mask = cellfun(@isempty,birthdate_var) | cellfun(@isempty,data_table_out.assessment_date);
% create a warning in case we are missing age
if sum(Mask) ~= 0
    warning('HARMONYRules:  some participants do not have age')
end

data_table_out.assessment_age(~Mask) = cell2mat(cellfun(@(x,y) split(caldiff([datetime(x,'InputFormat','dd-MM-yyyy'),datetime(y,'InputFormat','dd-MM-yyyy')],'months'),'mo'),birthdate_var(~Mask),data_table_out.assessment_date(~Mask),'UniformOutput',false));
%%
% visit_month == 0 because we are looking at baseline, expand this for
% follow up data
data_table_out.visit_month = nan(size(data_table,1),1);
data_table_out.visit_month(~Mask) = 0;
%%
% gender == 0
data_table_out.gender = nan(size(data_table,1),1);
data_table_out.gender(~Mask) = 0;
%%
% extract education
data_table_out.education_current = nan(size(data_table,1),1);
data_table_out.education_current(data_table.DEMOG_T0T1T2_38A2_WorkCurrent_Type_T0 == 2,1) = 1;
data_table_out.education_current(data_table.DEMOG_T0T1T2_38A2_WorkCurrent_Type_T0 ~= 2,1) = 2;
%% 
% education highest
data_table_out.education_highest = nan(size(data_table,1),1);
data_table_out.education_highest(data_table.DEMOG_T0T1T2_33_GraduationOther_T0 == 1,1) = 1;
data_table_out.education_highest(data_table.DEMOG_T0T1T2_34_GraduationWithout_T0 == 1,1) = 1;
data_table_out.education_highest(data_table.DEMOG_T0T1T2_32_GraduationUni_T0 == 1,1) = 2;
data_table_out.education_highest(data_table.DEMOG_T0T1T2_35_UniDegree_T0 == 1,1) = 5;
%%
% extract race_white
data_table_out.race_white = nan(size(data_table,1),1);
data_table_out.race_white(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[1:6]),1) = 1;
data_table_out.race_white(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[1:6]),1) = 2;
%%
% extract race_black
data_table_out.race_black = nan(size(data_table,1),1);
data_table_out.race_black(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[14:16]),1) = 1;
data_table_out.race_black(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[14:16]),1) = 2;
%%
% extract 'race_east_asian'
data_table_out.race_east_asian = nan(size(data_table,1),1);
data_table_out.race_east_asian(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[10:12]),1) = 1;
data_table_out.race_east_asian(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[10:12]),1) = 2;
%%
% extract 'race_asian_indian
data_table_out.race_asian_indian = nan(size(data_table,1),1);
data_table_out.race_asian_indian(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[7:9]),1) = 1;
data_table_out.race_asian_indian(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[7:9]),1) = 2;
%%
% extract 'race_middle_eastern'
data_table_out.race_middle_eastern = nan(size(data_table,1),1);
data_table_out.race_middle_eastern(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,13),1) = 1;
data_table_out.race_middle_eastern(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,13),1) = 2;
%%
% extract 'race_inter_racial'
data_table_out.race_inter_racial = nan(size(data_table,1),1);
data_table_out.race_inter_racial(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[17:20]),1) = 1;
data_table_out.race_inter_racial(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[17:20]),1) = 2;
%%
% extract 'race_other'
data_table_out.race_other = nan(size(data_table,1),1);
data_table_out.race_other(ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[21:23]),1) = 1;
data_table_out.race_other(~ismember(data_table.DEMOG_T0_02_Ethnicity_T0,[21:23]),1) = 2;
%%
% extract 'Immigrated'
data_table_out.Immigrated = nan(size(data_table,1),1);
data_table_out.Immigrated(~isnan(data_table.DEMOG_T0_01_AgeImmigration_T0),1) = 1;
data_table_out.Immigrated(isnan(data_table.DEMOG_T0_01_AgeImmigration_T0),1) = 2;
%%
% extract 'Referral_Source'
%% double check the numerical correspondance
data_table_out.Referral_Source = nan(size(data_table,1),1);
%% to code better ref 7 not existing
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_1,1),1) = 130;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_2,1),1) = 120;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_3,1),1) = 120;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_4,1),1) = 120;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_5,1),1) = 160;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_6,1),1) = 84;
% data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_7,1),1) = 170;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_8,1),1) = 140;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_12,1),1) = 140;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_9,1),1) = 100;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_10,1),1) = 110;
data_table_out.Referral_Source(ismember(data_table.REF_03_Referrer_Screening_11,1),1) = 180;
%% 
% Hospitalization
data_table_out.Hospitalization = nan(size(data_table,1),1);
data_table_out.Hospitalization(ismember(data_table.TREAT_HOSP_00_EverHospitalized_Screening,1),1) = 2;
data_table_out.Hospitalization(ismember(data_table.TREAT_HOSP_00_EverHospitalized_Screening,2),1) = 1;
%%
% extract 'caarms_e'
data_table_out.caarms_e = nan(size(data_table,1),1);
for i=1:size(data_table,1)
    %NOTE what happens if it is none of those
    % check if we can do that columnwise
    %% original, the name of the item does not match
    %     P.CAARMS_CHR_01_Q3_Group(i)==1 || P.CAARMS_CHR_01_Q2_Group(i)==1 || P.CAARMS_CHR_01_Q6_Group(i)==1
    % double check if this is correct
    if data_table.CAARMS_CHR_01_Q3_Group_Screening(i)==1 || data_table.CAARMS_CHR_02_Q2_Group_Screening(i)==1 || data_table.CAARMS_CHR_03_Q6_Group_Screening(i)==1
        data_table_out.caarms_e(i,1) = 1;
    else
        data_table_out.caarms_e(i,1) = 0;
    end
end
%%
% extract sips
SubSips = {'n','d','g'};
SubSipsPronia = {'N','D','G'};
for s = 1:numel(SubSips)
    for i = 1:6
        if i > 4
        if ismember(SubSips(s),{'d','g'})
            % skip if we have the items d and g and is bigger than 4, no
            % rules for taht
            continue
        end
        end
        data_table_out.([SubSips{s},num2str(i),'_onsetdatecode']) = nan(size(data_table,1),1);
        % 1="NA (Severity < 3)"
        data_table_out.([SubSips{s},num2str(i),'_onsetdatecode'])(data_table.(['SIPS_',SubSipsPronia{s},num2str(i),'_01_QUALY_B_00_0_SeverityScale_T0']) < 3) = 1;
        % 2="Onset Date available"
        data_table_out.([SubSips{s},num2str(i),'_onsetdatecode'])(data_table.(['SIPS_',SubSipsPronia{s},num2str(i),'_01_QUALY_C_00_Onset_T0']) == 3) = 2;
        % 3="Lifetime or cannot recall year of onset"
        data_table_out.([SubSips{s},num2str(i),'_onsetdatecode'])(data_table.(['SIPS_',SubSipsPronia{s},num2str(i),'_01_QUALY_C_00_Onset_T0']) == 1 | data_table.(['SIPS_',SubSipsPronia{s},num2str(i),'_01_QUALY_C_00_Onset_T0']) == 2) = 3;
    end
end
%%
% extract 'spi_a_sum'
data_table_out.('spi_a_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_A1_1_1_T0,data_table.SPI_A_A1_2_1_T0,data_table.SPI_A_A1_3_1_T0,...
    data_table.SPI_A_A2_1_1_T0,data_table.SPI_A_A2_2_1_T0,data_table.SPI_A_A3_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_a_sum') = sum(temp_spia,2);
%%
% extract 'spi_b_sum'
data_table_out.('spi_b_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_COGDIS_B1_1_Screening,data_table.SPI_A_B2_1_T0,data_table.SPI_A_B3_1_T0,...
    data_table.SPI_A_B4_1_T0,data_table.SPI_A_B5_1_T0,data_table.SPI_A_B6_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_b_sum') = sum(temp_spia,2);
%%
% extract 'spi_c_sum'
data_table_out.('spi_c_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_C1_1_T0,data_table.SPI_A_COGDIS_C2_1_Screening,data_table.SPI_A_COGDIS_C3_1_Screening,...
    data_table.SPI_A_COGDIS_C4_1_Screening,data_table.SPI_A_COGDIS_C5_1_Screening,data_table.SPI_A_C6_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_c_sum') = sum(temp_spia,2);
%%
% extract 'spi_d_sum'
data_table_out.('spi_d_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_D1_1_T0,data_table.SPI_A_D2_1_T0,data_table.SPI_A_COGDIS_D3_1_Screening,...
    data_table.SPI_A_COGDIS_D4_1_Screening,data_table.SPI_A_D5_1_Screening,data_table.SPI_A_A3_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_d_sum') = sum(temp_spia,2);
%%
% extract 'spi_e_sum'
data_table_out.('spi_e_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_E1_1_T0,data_table.SPI_A_E2_1_T0,data_table.SPI_A_E3_1_T0,...
    data_table.SPI_A_E4_1_T0,data_table.SPI_A_E5_1_T0,data_table.SPI_A_E6_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_e_sum') = sum(temp_spia,2);
%%
% extract 'spi_f_sum'
data_table_out.('spi_f_sum') = nan(size(data_table,1),1);
temp_spia = [data_table.SPI_A_F1_1_T0,data_table.SPI_A_F2_1_Screening,data_table.SPI_A_F3_1_Screening,...
    data_table.SPI_A_F4_1_T0,data_table.SPI_A_F5_1_Screening,data_table.SPI_A_F6_1_T0];
temp_spia(ismember(temp_spia,[7,8,9])) = 0;
data_table_out.('spi_f_sum') = sum(temp_spia,2);
%%
% extract 'ctq' see function CTQRules on how these are computed. NOTE some
% of the subscales are not there because the number of the CTQ differs from
% Pronia
% recoded items (add *1 to the item and reverse some of them)
for i = 1:9
    data_table_out.(['ctq_0',num2str(i)]) = nan(size(data_table,1),1);
    data_table_out.(['ctq_0',num2str(i)]) = data_table.(['CTQ_0',num2str(i),'_recod_T0']);
end
for i = 10:28
    data_table_out.(['ctq_',num2str(i)]) = nan(size(data_table,1),1);
    data_table_out.(['ctq_',num2str(i)]) = data_table.(['CTQ_',num2str(i),'_recod_T0']);
end
% extract ctq sexual abuse
data_table_out.ctqscore_sa = nan(size(data_table,1),1);
data_table_out.ctqscore_sa = data_table.CTQ_sabu_T0;
% extract ctq physical neglect
data_table_out.ctqscore_pn = nan(size(data_table,1),1);
data_table_out.ctqscore_pn = data_table.CTQ_pneg_T0;
% extract ctq validity, not in pronia rules might be worth it to add it
% CTQ validity items=CTQ_10 + CTQ_16 + CTQ_22
%         v = final_CTQ.(CTQ{10}) + final_CTQ.(CTQ{16}) + final_CTQ.(CTQ{22});
data_table_out.ctqscore_val = nan(size(data_table,1),1);
data_table_out.ctqscore_val = data_table.CTQ_10_recod_T0 + data_table.CTQ_16_recod_T0 + data_table.CTQ_22_recod_T0;
%%
% rename the items in the column recode
for d = 1:size(Dict,1)
    % check if it needs to be recoded, and if yes check that it is in the
    % data_table_out
    if isequal(Dict.Recode(d),1)
        if ~ismember(Dict.varName(d),data_table_out.Properties.VariableNames)
            disp([Dict.varName{d},' does not exist in data_table_out but the dictionary says that it is recoded'])
        end
        continue
    end
    % check if there is a corresponding variable name, if not set all the
    % values to nan
    if isequal(Dict.PRONIAvar(d),{'NaN'})
%         disp([Dict.varName{d},' has no corresponding variable name in dictionnary set the whole variable to empty'])
        data_table_out.(Dict.varName{d}) = nan(size(data_table,1),1);
        continue
    end
    % check if it is a date and transform into HARMONY date format, also
    % here there is a more elegant way to do that
    if isequal(Dict.Type(d),{'Date'})
        if ismember({[Dict.PRONIAvar{d},'_Screening']},data_table.Properties.VariableNames)
        TempDate = data_table.([Dict.PRONIAvar{d},'_Screening']);
        elseif ismember({[Dict.PRONIAvar{d},'_T0']},data_table.Properties.VariableNames)
        TempDate = data_table.([Dict.PRONIAvar{d},'_T0']);
        else
        disp([Dict.varName{d},' does not exist in data_table but has a corresponding name in dictionnary replace the whole column by nan'])
        data_table_out.(Dict.varName{d}) = nan(size(data_table,1),1);
        continue
        end
        data_table_out.(Dict.varName{d}) = cell(size(data_table,1),1);
        for p = 1:numel(TempDate)
            if ~isempty(TempDate{p})
                % there is certainly a more elegant way to do that
               data_table_out.(Dict.varName{d}){p} = TempDate{p}([4,5,3,1,2,6,7,8,9,10]);
            end
        end 
        continue
    end  
        
    % just rename the item
    
    if ismember({[Dict.PRONIAvar{d},'_Screening']},data_table.Properties.VariableNames)
        data_table_out.(Dict.varName{d}) = data_table.([Dict.PRONIAvar{d},'_Screening']);
    elseif ismember({[Dict.PRONIAvar{d},'_T0']},data_table.Properties.VariableNames)
        data_table_out.(Dict.varName{d}) = data_table.([Dict.PRONIAvar{d},'_T0']);
    else
        if isnan(Dict.varName{d})
            break
        end
        disp([Dict.varName{d},' does not exist in data_table but has a corresponding name in dictionnary replace the whole column by nan'])
        data_table_out.(Dict.varName{d}) = nan(size(data_table,1),1);
        continue
    end
end

%% need to replace nan by empty. Convert the whole table to cell and replace nans by empty

% script to generate dictionnary for mri HARMONY
% step 1. Define the subjects items such as te tr subject id and regions of
% interest
% step 2. Read the data automatically (MetaQualMRI and the text files
% storing the data)
% step 3.  create the dictionnary, depending on the number of regions of
% interest
% step 4. create the dicitonnary with the following columns:
% * Variable name
% * Variable description
% * Type (continuous, categorical, date)
% * Value (range, category, date format)

%% subject id old
% subjectkey
% src_subject_id
% interview_date
% interview_age
% gender

% data_table_out.consortium
% data_table_out.subject_id
% data_table_out.site
% data_table_out.assessment_date
% data_table_out.assessment_age
% data_table_out.gender

% mri parameters
% data_table_out.site
% acqu_sub_site? no need for this one except if we have a scanner upgrade
%%
% mri_brand (model name? Mark?)
% 'Manufacturer_sMRI_T0_PhilipsHealthcare';
% 'Manufacturer_sMRI_T0_PhilipsMedicalSystems';
% 'Manufacturer_sMRI_T0_SIEMENS'
% {'SiemensTrio PhilipsIngena'};
% offer of naming convention dictionnary
%
%% 
% initialise the out dictionnary
OutDict = Dict(1,:);
%%
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 'mri_brand';
TempDict.Values     = '1="PhilipsMedicalSystemsAchieva",2="PhilipsHealthcareIngenia",3="PhilipsMedicalSystemsIngenuity",4="SIEMENSPrisma",5="SIEMENSPrismafit",6="SIEMENSTrioTim",7="SIEMENSVerio"';
TempDict.Type       = 'categorical'; 
TempDict.PRONIAvar = {'Brand and scanner type combined'};

OutDict = [OutDict;TempDict];
% 1="PhilipsMedicalSystemsAchieva",2="PhilipsHealthcareIngenia",3="PhilipsMedicalSystemsIngenuity",4="SIEMENSPrisma",5="SIEMENSPrismafit",6="SIEMENSTrioTim",7="SIEMENSVerio"
BrandNames = {'Achieva','Ingenia','Ingenuity','Prisma','Prismafit','TrioTim','Verio'};
data_table_out.mri_brand = nan(size(data_table,1),1);
for b = 1:numel(BrandNames)
    data_table_out.mri_brand(data_table.(['Model_name_sMRI_T0_',BrandNames{b}])==1) = b;
end
%%
% mri_field_strength 3T by default
% offer of naming convention dictionnary
% 1="3T"
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 'mri_field_strength';
TempDict.Values     = '1="3T"';
TempDict.Type       = 'categorical'; 
TempDict.PRONIAvar = {'Scanner field strength'};

OutDict = [OutDict;TempDict];
data_table_out.mri_field_strength = nan(size(data_table,1),1);
data_table_out.mri_field_strength(~isempty(data_table_out.mri_brand)) = 1;
%%
% t1w_acqu_time TA
%%
% t1w_vox_size_x
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_vox_size_x';
TempDict.Values     = '0:3:4';
TempDict.Type       = 'continous'; 
TempDict.PRONIAvar = {'T1-weighted mri voxel size x direction'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_vox_size_x = nan(size(data_table,1),1);
data_table_out.t1w_vox_size_x = data_table.Voxel_size_x_mm_sMRI_T0;
%%
% t1w_vox_size_y
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_vox_size_y';
TempDict.Values     = '0:3:4';
TempDict.Type       = 'continous';
TempDict.PRONIAvar = {'T1-weighted mri voxel size y direction'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_vox_size_y = nan(size(data_table,1),1);
data_table_out.t1w_vox_size_y = data_table.Voxel_size_y_mm_sMRI_T0;
%%
% t1w_vox_size_z no value
% data_table_out.t1w_vox_size_z = nan(size(data_table,1),1);
% data_table_out.t1w_vox_size_z = data_table.Slice_thickness_mm_sMRI_T0;
%%
% t1w_slices
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_slices';
TempDict.Values     = '0:300:0';
TempDict.Type       = 'continous';
TempDict.PRONIAvar = {'T1-weighted number of slices'};

OutDict = [OutDict;TempDict];

data_table_out.t1w_slices = nan(size(data_table,1),1);
data_table_out.t1w_slices = data_table.Number_of_slices_sMRI_T0;
%%
%t1w_orientation
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_orientation';
TempDict.Values     = '1="coronal",2="sagittal",3="transversal"';
TempDict.Type       = 'categorical'; 
TempDict.PRONIAvar = {'T1-weighted orientation'};

OutDict = [OutDict;TempDict];
% dictionnary 1="coronal",2="sagittal",3="transversal";
OrientName = {'cor','sag','tra'};
data_table_out.t1w_orientation = nan(size(data_table,1),1);
for b = 1:numel(OrientName)
    data_table_out.t1w_orientation(data_table.(['Orientation_sMRI_T0_',OrientName{b}])==1) = b;
end
%%
% t1w_fov ?
% data_table_out.t1w_fov = nan(size(data_table,1),1);
% data_table_out.t1w_fov =
%%
% t1w_thickness
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_thickness';
TempDict.Values     = '0:10:4';
TempDict.Type       = 'continuous'; 
TempDict.PRONIAvar = {'T1-weighted slice thickness'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_thickness = nan(size(data_table,1),1);
data_table_out.t1w_thickness = data_table.Slice_thickness_mm_sMRI_T0;
%%
% t1w_tr
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_tr';
TempDict.Values     = '0:3:4';
TempDict.Type       = 'continuous'; 
TempDict.PRONIAvar = {'T1-weighted repetition time'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_tr = nan(size(data_table,1),1);
data_table_out.t1w_tr = data_table.Repetition_time_ms_sMRI_T0;
%%
% t1w_te
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_te';
TempDict.Values     = '0:6:4';
TempDict.Type       = 'continuous';
TempDict.PRONIAvar = {'T1-weighted echo time'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_te = nan(size(data_table,1),1);
data_table_out.t1w_te = data_table.Echo_time1_ms_sMRI_T0;
%%
% t1w_fa
TempDict = Dict(1,:);
TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
TempDict.varName   = 't1w_fa';
TempDict.Values     = '0:15:0';
TempDict.Type       = 'continuous'; 
TempDict.PRONIAvar = {'T1-weighted flip angle'};

OutDict = [OutDict;TempDict];
data_table_out.t1w_fa = nan(size(data_table,1),1);
data_table_out.t1w_fa = data_table.Flip_angle_sMRI_T0;
%%
% t1w_pulse_sequ?
%%
% actual mri stuff
% part of the code to integrate in the query system
% folder containing the files. WARNING: these data are not in the pronia
% format

DataFolder = '/volume/HARMONY/Qunex/freesurfer_tables';
TextNames = {'_T0_aparc_stats_hemi_lh.txt','_T0_aparc_stats_hemi_rh.txt','_T0_asegs_stats.txt'};
TempNamesAll = {};
for p = 1:size(data_table,1)
    % read cortex lh
    for t = 1:numel(TextNames)
        FileName = fullfile(DataFolder,[data_table.PSN{p},TextNames{t}]);
        if ~exist(FileName,'file')
            % if the file does not exist there is no point to continue
            continue
        end
        TempData = readtable(fullfile(DataFolder,[data_table.PSN{p},TextNames{t}]));
        TempNamesAll = [TempNamesAll,TempData.Properties.VariableNames];

        for d = 1:numel(TempData)
            if ismember(TempData.Properties.VariableNames(d),{'lh_aparc_area','rh_aparc_area','Measure_volume'})
                continue
            end
            if ~ismember(TempData.Properties.VariableNames(d),data_table_out.Properties.VariableNames)
                data_table_out.(TempData.Properties.VariableNames{d}) = nan(size(data_table,1),1);
            end
            data_table_out.(TempData.Properties.VariableNames{d})(p) = TempData.(TempData.Properties.VariableNames{d});
        end
    end
end

TempNamesAll = unique(TempNamesAll);
TempNamesAll = TempNamesAll(~ismember(TempNamesAll,{'lh_aparc_area','rh_aparc_area','Measure_volume'}));
% create the dictionnary

for i = 1:numel(TempNamesAll)
    
    TempDict = Dict(1,:);
    TempDict.Properties.VariableNames = Dict.Properties.VariableNames;
    TempDict.varName   = TempNamesAll{i};
    TempDict.Values     = ['0:',num2str(ceil(max(data_table_out.(TempNamesAll{i})))),':0'];
    TempDict.Type       = 'continuous'; 
    TempDict.PRONIAvar = TempDict.varName;

    OutDict = [OutDict;TempDict];
end

OutDict = OutDict(2:end,:);
OutDict.Properties.VariableNames(ismember(OutDict.Properties.VariableNames,{'PRONIAvar'})) = {'VariableDescription'};
OutDict(:,{'Recode','RacheleComments'}) = [];
%% 
% FS_EulerNumber
end