%% AmPeps % the Entry Prepare System of AmP %% function AmPeps(infoAmPgui) % created 2020/06/09 by Bas Kooijman, modified 2021/04/26 %% Syntax % <../AmPeps.m *AmPeps*>(infoAmPgui) %% Description % The function is a shell around several other functions, including AmPgui, that eventually write: % mydata_my_pet.m, pars_init_my_pet.m, predict_my_pet.m and run_my_pet.m. % Guidance is presented at % % Input: % % * infoAmPgui: optional boolean for skip writing (0) or writing (>0) 4 source files % % - 0, species is in AmP, skip writing 4 source files % - 1, writing 4 source files with species recognized by CoL % - 2, writing 4 source files with species not recognized by CoL, but genus is in AmP % - 3, writing 4 source files with species not recognized by CoL, genus is not in AmP, but family is % - 4, writing 4 source files with species not recognized by CoL, family is not in AmP, but order is % - 5, writing 4 source files with species not recognized by CoL, order is not in AmP, but class is % - 6, writing 4 source files with species not recognized by CoL, class is not in AmP, but phylum is % - 7, writing 4 source files with species not recognized by CoL, phylum is not in AmP global data metaData txtData auxData pets hclimateLand hclimateSea hecozones hoceans my_pet_lineage % downloads below use the built-in websave (cross-platform, no external wget needed) if ~exist('infoAmPgui', 'var') % open webpages, show figures and start AmPgui web('https://debportal.debtheory.org/docs/AmPeps.html', '-browser'); web([set_path2server, 'add_my_pet/AmPeco.html'], '-browser'); hclimateLand = figure('Name','Land climate', 'Position',[300 450 500 300]); image(imread('climate_land.png')); set(gca, 'Visible', 'off'); hclimateSea = figure('Name','Sea climate', 'Position',[900 450 500 300]); image(imread('climate_sea.jpg')); set(gca, 'Visible', 'off'); hecozones = figure('Name','Land ecozone', 'Position',[300 50 500 300]); image(imread('ecozones.png')); set(gca, 'Visible', 'off'); hoceans = figure('Name','Sea ecozone', 'Position',[900 50 500 300]); image(imread('oceans.jpg')); set(gca, 'Visible', 'off'); AmPgui elseif infoAmPgui == 0 % skip the rest of AmPeps and proceed with opening source files in Matlab editor close all path = [set_path2server, 'add_my_pet/entries/', metaData.species, '/']; files = {['mydata_', metaData.species, '.m'], ['pars_init_', metaData.species, '.m'], ... ['predict_', metaData.species, '.m'], ['run_', metaData.species, '.m']}; for i = 1:numel(files) websave(files{i}, [path, files{i}]); end edit(['mydata_', metaData.species, '.m'], ... ['pars_init_', metaData.species, '.m'], ... ['predict_', metaData.species, '.m'], ... ['run_', metaData.species, '.m']) else % infoAmPgui > 0: proceed to writing 4 AmP source files for new species for AmP close all; % no return from here: write mydata and run files, deleting mat-files copyfile(['results_', metaData.species, '.mat'], ['results_', metaData.species, '_backup.mat'], 'f') if ~isempty(data.data_1) fld_data_1 = fields(data.data_1); % required to add data-specific scaled functional responses else fld_data_1 = {}; end [data, metaData] = AmPgui2mydata(data, metaData); % modify data and metaData to mydata format prt_mydata(data, auxData, metaData, txtData); % write mydata_my_pet.m file prt_run_my_pet(metaData.species); % write run_my_pet.m file % get pars_init file of a related species to produce structure par, metaPar, txtPar model_def = get_model(metaData.phylum, metaData.class, metaData.order); % default model for this taxon if strcmp(model_def, 'hep') && (isfield(data,'Lj') || isfield(data,'Wwj') || isfield(data,'Wdj') || isfield(data,'tj')) model_def = 'hex'; end % park data structures, because they will be overwritten by load(results_my_pet) data_my_pet = data; txtData_my_pet = txtData; auxData_my_pet = auxData; metaData_my_pet = metaData; switch infoAmPgui case 1 % species recognized by CoL, not in AmP Clade = clade(metaData.species); % identify clade to which species belongs case 2 % species not recognized by CoL, genus in AmP genus = strsplit(metaData.species,'_'); genus = genus{1}; Clade = select(genus); case 3 % species not recognized by CoL, genus not in AmP, family in AmP Clade = select(metaData.family); case 4 % species not recognized by CoL, family not in AmP, order in AmP Clade = select(metaData.order); case 5 % species not recognized by CoL, order not in AmP, class in AmP Clade = select(metaData.class); case 6 % species not recognized by CoL, class not in AmP, phylum in AmP Clade = select(metaData.phylum); case 7 % species not recognized by CoL, phylum not in AmP Clade = select; end Clade = Clade(~ismember(Clade,metaData.species)); % exclude the species itself n_Clade = length(Clade); Clade = Clade(1:min(10,n_Clade)); n_Clade = length(Clade); % set max clade members at 10 criterion = zeros(n_Clade,1); model_Clade = cell(n_Clade,1); resultsFn = cell(n_Clade,1); path = [set_path2server, 'add_my_pet/entries/']; % path for results_my_pet.mat files for i = 1:n_Clade % scan clade members resultsFn{i} = ['results_', Clade{i}, '.mat']; websave(resultsFn{i}, [path, Clade{i}, '/', resultsFn{i}]); load(resultsFn{i}); criterion(i) = metaData.COMPLETE / metaPar.MRE; model_Clade{i} = metaPar.model; end if isempty(model_def); model_def = get_model(metaData.phylum, metaData.class, metaData.order); end % default model for this taxon sel_Clade = ismember(model_Clade, model_def); if ~any(sel_Clade) fprintf(['Warning from AmPeps: none of the ', num2str(n_Clade), ' clade members has the required model ', model_def, '; one with another model is used\n']); sel_Clade(1) = true; % fall back to the first member end % among the selected members, pick the one with the largest criterion (COMPLETE/MRE) crit = criterion; crit(~sel_Clade) = -Inf; [~, i_best] = max(crit); load(resultsFn{i_best}); % use the best clade member for the initial parameter estimates fprintf(['Notice from AmPeps: AmP species ', Clade{i_best}, ' was used for initial parameter estimates with model ', model_Clade{i_best}, '\n']); for i = 1:n_Clade % remove all downloaded clade results files (results_my_pet_backup.mat is kept) if exist(resultsFn{i}, 'file'); delete(resultsFn{i}); end end data = data_my_pet; txtData = txtData_my_pet; auxData = auxData_my_pet; metaData = metaData_my_pet; % result data structures auxParFld = prt_predict(par, metaPar, data, auxData, metaData); % write predict file for model type taken from metaPar % auxParFld contains names of required auxiliary parameter % get auxiliary parameters parFields = fields(par); % current par fields EparFields = get_parfields(metaPar.model); % core primary parameter fields for model parFields = setdiff(parFields, {'T_ref'}); % remove parameter T_ref from current-list parFields = setdiff(parFields, EparFields); % non-core parameter fields chemParFields = fields(addchem([], [], [], [], metaData.phylum, metaData.class)); % par fields for chemical parameters otherParFields = setdiff(parFields, chemParFields); % current auxiliary parameter fields otherParFields = setdiff(otherParFields, {'free', 'f'}); selPar = ismember(otherParFields, auxParFld); par = rmfield(par, otherParFields(~selPar)); % remove unnecessary parameters addParFields = auxParFld(~ismember(auxParFld, otherParFields(selPar))); % parameters fields that must be added n_add = length(addParFields); if n_add > 0 % edit par & txtPar load auxPar for i = 1:n_add if ischar(auxPar.(addParFields{i}).value) par.(addParFields{i}) = eval(auxPar.(addParFields{i}).value); else par.(addParFields{i}) = auxPar.(addParFields{i}).value; end par.free.(addParFields{i}) = auxPar.(addParFields{i}).free; txtPar.units.(addParFields{i}) = auxPar.(addParFields{i}).units; txtPar.label.(addParFields{i}) = auxPar.(addParFields{i}).label; end end n_add = length(fld_data_1); % add scaled functional responses for i = 1:n_add nm = ['f_', fld_data_1{i}]; par.(nm) = 1; par.free.(nm) = 1; txtPar.units.(nm) = '-'; txtPar.label.(nm) = ['scaled functional response for ',fld_data_1{i}, ' data']; end % overwrite results_my_pet.mat in a way that it can no longer be used as substrate for AmPgui save(['results_', metaData.species, '.mat'], 'data', 'auxData', 'metaData', 'txtData', 'par', 'metaPar', 'txtPar'); pets = {metaData.species}; mat2pars_init; % this uses results_my_pet.mat to overwrite pars_init_my_pet.m % open 4 AmP source files in Matlab Editor edit(['mydata_', metaData.species, '.m'], ... ['pars_init_', metaData.species, '.m'], ... ['predict_', metaData.species, '.m'], ... ['run_', metaData.species, '.m']); end