reader_internal_csv.cc

Go to the documentation of this file.

/*!
 *
 * Copyright (C) 2015 Technical University of Liberec.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify it under
 * the terms of the GNU General Public License version 3 as published by the
 * Free Software Foundation. (http://www.gnu.org/licenses/gpl-3.0.en.html)
 *
 * This program is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 * FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
 *
 *
 * @file    reader_internal_csv.cc
 * @brief
 */
 
 
#include "input/reader_internal_csv.hh"
#include "input/input_type.hh"
#include "input/csv_tokenizer.hh"
 
#include "system/asserts.hh"                           // for Assert, ASSERT
#include "system/file_path.hh"                         // for FilePath, File...
#include "system/logger.hh"                            // for operator<<
 
namespace Input {
 
using namespace std;
 
/*******************************************************************
 * implementation of ReaderInternalCsvInclude
 */
 
ReaderInternalCsvInclude::ReaderInternalCsvInclude()
{}
 
StorageBase * ReaderInternalCsvInclude::read_storage(PathBase &p, const Type::Array *array)
{
    if ( p.is_record_type() ) { // sub-type must be Record or Abstract type
        // load path to CSV file
        std::string included_file;
        if ( p.down("file") ) {
            included_file = get_included_file(p);
            p.up();
        } else {
            this->generate_input_error(p, array, "Missing key 'file' defines including input file.", false);
        }
 
        // number of head lines to skip
        unsigned int n_head_lines = 0;
        if ( p.down("n_head_lines") ) {
            try {
                n_head_lines = p.get_int_value();
            }
            catch (ExcInputError & e) {
                complete_input_error(e, p, ValueTypes::int_type);
                e << EI_InputType("number of lines to skip");
                throw;
            }
            p.up();
        }
 
        // separator characters
        stringstream separator;
        if ( p.down("separator") ) {
            try {
                separator << p.get_string_value() << " \t";
            }
            catch (ExcInputError & e) {
                complete_input_error(e, p, ValueTypes::str_type);
                e << EI_InputType("invalid separator of included CSV file");
                throw;
            }
            p.up();
        } else {
            separator << ", \t";
        }
 
        // open CSV file, get number of lines, skip head lines
        FilePath fp((included_file), FilePath::input_file);
        CSVTokenizer tok( fp, separator.str() );
 
        const Type::Abstract * abstract_type = dynamic_cast<const Type::Abstract *>(&array->get_sub_type());
        std::string record_name = p.get_record_tag();
        if ((abstract_type != NULL) && (record_name.size() <= 12)) {
            // missing record name for Abstract
            this->generate_input_error(p, array,
                "Missing record descendant in definition of tag in CSV include of Abstract type. Tag must be in format '!include_csv:Descendant_Name'!",
                false);
        }
        const Type::TypeBase &sub_type = ( abstract_type != NULL ) ?  // sub-type of array
                (abstract_type->get_descendant(record_name.substr(12))) : (array->get_sub_type());
 
        StorageBase *item_storage = nullptr; // storage of sub-type record of included array
        csv_columns_map_.clear();
        if ( p.down("format") ) {
            try {
                csv_subtree_depth_ = p.path_.size();
                item_storage = make_storage(p, &sub_type);
            } catch (ExcMultipleDefinitionCsvColumn &e) {
                e << EI_File(tok.f_name());
                throw;
            }
            p.up();
        } else {
            this->generate_input_error(p, array, "Missing key 'format' defines mapping column of CSV file to input subtree.", false);
        }
 
        // get value of maximal column index in map
        map<unsigned int, IncludeCsvData>::iterator it;
        it = csv_columns_map_.end(); --it;
        unsigned int max_column_index = it->first;
 
        unsigned int n_lines = tok.get_n_lines() - n_head_lines;
        tok.skip_header(n_head_lines);
        StorageArray *storage_array = new StorageArray(n_lines);
        std::set<unsigned int> unused_columns;
        for( unsigned int arr_item=0; arr_item < n_lines; ++arr_item) {
            unsigned int i_col;
            tok.next_line();
            for (i_col=0; !tok.eol(); ++i_col, ++tok) {
                it = csv_columns_map_.find(i_col);
                if (it != csv_columns_map_.end()) {
                    switch (it->second.data_type) {
                        case IncludeDataTypes::type_int: {
                            int val;
                            try {
                                val = tok.get_int_val();
                            } catch (ExcWrongCsvFormat &e) {
                                e << EI_Specification("Wrong integer value");
                                e << EI_ErrorAddress(p.as_string());
                                throw;
                            }
 
                            const Type::Integer *int_type = static_cast<const Type::Integer *>(it->second.type);
                            if ( !int_type->match(val) ) {
                                THROW( ExcWrongCsvFormat() << EI_Specification("Integer value out of bounds")
                                        << EI_TokenizerMsg(tok.position_msg()) << EI_ErrorAddress(p.as_string()) );
                            }
                            set_storage_from_csv( i_col, item_storage, new StorageInt(val) );
                            break;
                        }
                        case IncludeDataTypes::type_double: {
                            double val;
                            try {
                                val = tok.get_double_val();
                            } catch (ExcWrongCsvFormat &e) {
                                e << EI_ErrorAddress(p.as_string());
                                throw;
                            }
 
                            const Type::Double *double_type = static_cast<const Type::Double *>(it->second.type);
                            if ( !double_type->match(val) ) {
                                THROW( ExcWrongCsvFormat() << EI_Specification("Double value out of bounds")
                                        << EI_TokenizerMsg(tok.position_msg()) << EI_ErrorAddress(p.as_string()) );
                            }
                            set_storage_from_csv( i_col, item_storage, new StorageDouble(val) );
                            break;
                        }
                        case IncludeDataTypes::type_bool: {
                            int val;
                            try {
                                val = tok.get_int_val();
                            } catch (ExcWrongCsvFormat &e) {
                                e << EI_Specification("Wrong boolean value");
                                e << EI_ErrorAddress(p.as_string());
                                throw;
                            }
                            set_storage_from_csv( i_col, item_storage, new StorageBool(val) );
                            break;
                        }
                        case IncludeDataTypes::type_string: {
                            try {
                                set_storage_from_csv( i_col, item_storage, new StorageString(tok.get_string_val()) );
                            } catch (ExcWrongCsvFormat &e) {
                                e << EI_Specification("Wrong string value");
                                e << EI_ErrorAddress(p.as_string());
                                throw;
                            }
                            break;
                        }
                        case IncludeDataTypes::type_sel: {
                            const Type::Selection *selection = static_cast<const Type::Selection *>(it->second.type);
                            try {
                                std::string item_name = tok.get_string_val();
                                int val = selection->name_to_int( item_name );
                                set_storage_from_csv( i_col, item_storage, new StorageInt(val) );
                            } catch (ExcWrongCsvFormat &e) {
                                e << EI_Specification("Wrong selection value");
                                e << EI_ErrorAddress(p.as_string());
                                throw;
                            } catch (Type::Selection::ExcSelectionKeyNotFound &exc) {
                                THROW( ExcWrongCsvFormat() << EI_Specification("Wrong selection value")
                                        << EI_TokenizerMsg(tok.position_msg()) << EI_ErrorAddress(p.as_string()) );
                            }
                            break;
                        }
                    }
                } else {
                    // add index of unused column
                    unused_columns.insert(i_col);
                }
            }
            if ( max_column_index > (i_col-1) ) {
                this->generate_input_error(p, array, "Count of columns in CSV file is less than expected index, defined on input.", false);
            }
            ASSERT_PTR(item_storage);
            storage_array->new_item(arr_item, item_storage->deep_copy() );
        }
 
        if (unused_columns.size()) { // print warning with indexes of unused columns
            stringstream ss;
            for (std::set<unsigned int>::iterator it=unused_columns.begin(); it!=unused_columns.end(); ++it)
                ss << (*it) << " ";
            WarningOut().fmt("Unused columns: {}\nin imported CSV input file: {}\n", ss.str(), tok.f_name());
        }
        return storage_array;
 
    } else {
        this->generate_input_error(p, array, "Invalid definition of CSV include.", false);
    }
    return NULL;
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::Array *array)
{
    int arr_size;
    if ( (arr_size = p.get_array_size()) != -1 ) {
        return this->make_array_storage(p, array, arr_size);
    } else {
        this->generate_input_error(p, array, "Invalid type in CSV-included part of IST. Expected is Array!\n", false);
    }
 
    return NULL;
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::Selection *selection)
{
    int pos;
    if ( check_and_read_position_index(p, pos) ) {
        IncludeCsvData include_data;
        include_data.data_type = IncludeDataTypes::type_sel;
        include_data.storage_indexes = create_indexes_vector(p);
        include_data.type = selection;
        if (csv_columns_map_.find(pos)!=csv_columns_map_.end()) {
            THROW( ExcMultipleDefinitionCsvColumn() << EI_ColumnIndex(pos) << EI_ErrorAddress(p.as_string()) );
        } else {
            csv_columns_map_[pos] = include_data;
        }
 
        return new StorageInt( 0 );
    } else {
        string item_name = read_string_value(p, selection);
        try {
            int value = selection->name_to_int( item_name );
            return new StorageInt( value );
        } catch (Type::Selection::ExcSelectionKeyNotFound &exc) {
            this->generate_input_error(p, selection, "Wrong value '" + item_name + "' of the Selection.", false);
        }
    }
    return NULL;
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::Bool *bool_type)
{
    int pos;
    if ( check_and_read_position_index(p, pos) ) {
        IncludeCsvData include_data;
        include_data.data_type = IncludeDataTypes::type_bool;
        include_data.storage_indexes = create_indexes_vector(p);
        include_data.type = bool_type;
        if (csv_columns_map_.find(pos)!=csv_columns_map_.end()) {
            THROW( ExcMultipleDefinitionCsvColumn() << EI_ColumnIndex(pos) << EI_ErrorAddress(p.as_string()) );
        } else {
            csv_columns_map_[pos] = include_data;
        }
 
        return new StorageInt( 0 );
    } else {
        return new StorageInt( read_bool_value(p, bool_type) );
    }
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::Integer *int_type)
{
    int pos;
    if ( check_and_read_position_index(p, pos) ) {
        IncludeCsvData include_data;
        include_data.data_type = IncludeDataTypes::type_int;
        include_data.storage_indexes = create_indexes_vector(p);
        include_data.type = int_type;
        if (csv_columns_map_.find(pos)!=csv_columns_map_.end()) {
            THROW( ExcMultipleDefinitionCsvColumn() << EI_ColumnIndex(pos) << EI_ErrorAddress(p.as_string()) );
        } else {
            csv_columns_map_[pos] = include_data;
        }
 
        return new StorageInt( 0 );
    } else {
        return new StorageInt( read_int_value(p, int_type) );
    }
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::Double *double_type)
{
    int pos;
    if ( check_and_read_position_index(p, pos) ) {
        IncludeCsvData include_data;
        include_data.data_type = IncludeDataTypes::type_double;
        include_data.storage_indexes = create_indexes_vector(p);
        include_data.type = double_type;
        if (csv_columns_map_.find(pos)!=csv_columns_map_.end()) {
            THROW( ExcMultipleDefinitionCsvColumn() << EI_ColumnIndex(pos) << EI_ErrorAddress(p.as_string()) );
        } else {
            csv_columns_map_[pos] = include_data;
        }
 
        return new StorageDouble( 0.0 );
    } else {
        return new StorageDouble( read_double_value(p, double_type) );
    }
}
 
StorageBase * ReaderInternalCsvInclude::make_sub_storage(PathBase &p, const Type::String *string_type)
{
    int pos;
    if ( check_and_read_position_index(p, pos) ) {
        IncludeCsvData include_data;
        include_data.data_type = IncludeDataTypes::type_string;
        include_data.storage_indexes = create_indexes_vector(p);
        include_data.type = string_type;
        if (csv_columns_map_.find(pos)!=csv_columns_map_.end()) {
            THROW( ExcMultipleDefinitionCsvColumn() << EI_ColumnIndex(pos) << EI_ErrorAddress(p.as_string()) );
        } else {
            csv_columns_map_[pos] = include_data;
        }
 
        return new StorageString("");
    } else {
        return new StorageString( read_string_value(p, string_type) );
    }
}
 
vector<unsigned int> ReaderInternalCsvInclude::create_indexes_vector(PathBase &p)
{
    vector<unsigned int> csv_storage_indexes( p.path_.size()-csv_subtree_depth_ );
    for (unsigned int i_source=csv_subtree_depth_, i_target=0; i_source<p.path_.size(); ++i_source, ++i_target ) {
        ASSERT_GE(p.path_[i_source].first, 0).error();
        csv_storage_indexes[i_target] = p.path_[i_source].first;
    }
    return csv_storage_indexes;
}
 
void ReaderInternalCsvInclude::set_storage_from_csv(unsigned int column_index, StorageBase * item_storage, StorageBase * new_storage)
{
    map<unsigned int, IncludeCsvData>::iterator it = csv_columns_map_.find(column_index);
    ASSERT(it!=csv_columns_map_.end()).error();
 
    unsigned int i;
    StorageBase *loop_storage = item_storage;
    for (i=0; i<it->second.storage_indexes.size()-1; ++i) loop_storage = loop_storage->get_item( it->second.storage_indexes[i] );
    loop_storage->set_item( it->second.storage_indexes[i], new_storage );
}
 
bool ReaderInternalCsvInclude::check_and_read_position_index(PathBase &p, int &pos)
{
    string value;
    try {
        value = p.get_string_value();
    } catch (ExcInputError &) {
        // value is not string, return false
        return false;
    }
 
    // value must start with '$', follows nonnegative number
    if ( value.size() && (value.substr(0,1) == "$") ) {
        try {
            pos = std::stoi( value.substr(1) );
            return (pos >= 0);
        } catch (std::invalid_argument &) {
            return false;
        }
    } else {
        return false;
    }
}
 
 
} // namespace Input