masks.hpp

// Copyright (c) 2023, INRAE.
// Distributed under the terms of the GPL-3 Licence.
// The full licence is in the file LICENCE, distributed with this software.

#ifndef EVALHYD_MASKS_HPP
#define EVALHYD_MASKS_HPP

#include <map>
#include <set>
#include <vector>
#include <array>
#include <string>
#include <regex>
#include <stdexcept>

#include <xtensor/xexpression.hpp>
#include <xtensor/xtensor.hpp>
#include <xtensor/xview.hpp>
#include <xtensor/xsort.hpp>
#include <xtensor/xindex_view.hpp>


typedef std::map<std::string, std::vector<std::vector<std::string>>> msk_tree;

namespace evalhyd
{
    namespace masks
    {
        /// Function to parse a string containing masking conditions.
        inline msk_tree parse_masking_conditions(std::string msk_str)
        {
            msk_tree subset;

            // pattern supported to specify conditions to generate masks on
            // observed or predicted (median or mean for probabilist) streamflow
            // e.g. q{>9.} q{<9} q{>=99.0} q{<=99} q{>9,<99} q{==9} q{!=9}
            std::regex exp_q (
                    R"((q_obs|q_prd_median|q_prd_mean)\{(((<|>|<=|>=|==|!=)(mean,?|median,?|qtl(0|1)\.(0|1|2|3|4|5|6|7|8|9)+,?|(0|1|2|3|4|5|6|7|8|9)+\.?(0|1|2|3|4|5|6|7|8|9)*,?))+)\})"
                    // NOTE: this should be `R"((q_obs|q_prd_median|q_prd_mean)\{(((<|>|<=|>=|==|!=)(mean,?|median,?|qtl[0-1]\.[0-9]+,?|[0-9]+\.?[0-9]*,?))+)\})"`
                    //       but there is a bug in the building chain for R packages
                    //       https://gitlab.irstea.fr/HYCAR-Hydro/evalhyd/evalhyd-r/-/issues/6
            );

            for (std::sregex_iterator i =
                    std::sregex_iterator(msk_str.begin(), msk_str.end(), exp_q);
                 i != std::sregex_iterator(); i++)
            {
                const std::smatch & mtc = *i;

                std::string var = mtc[1];
                std::string str = mtc[2];

                // process masking conditions on streamflow
                std::vector<std::vector<std::string>> conditions;

                // pattern supported to specify masking conditions based on streamflow
                std::regex ex (
                        R"((<|>|<=|>=|==|!=)(mean|median|qtl(0|1)\.(0|1|2|3|4|5|6|7|8|9)+|(0|1|2|3|4|5|6|7|8|9)+\.?(0|1|2|3|4|5|6|7|8|9)*))"
                        // NOTE: this should be `R"((<|>|<=|>=|==|!=)(mean|median|qtl[0-1]\.[0-9]+|[0-9]+\.?[0-9]*))"`
                        //       but there is a bug in the building chain for R packages
                        //       https://gitlab.irstea.fr/HYCAR-Hydro/evalhyd/evalhyd-r/-/issues/6
                );

                for (std::sregex_iterator j =
                        std::sregex_iterator(str.begin(), str.end(), ex);
                     j != std::sregex_iterator(); j++)
                {
                    const std::smatch & mt = *j;

                    if ((mt[2].str() == "median")