UniParser.h

Go to the documentation of this file.

// -----------------------------------------------------------------------------------------
// -----  Functions to parse TXT, XML, CSV files and write the values to the database  -----
// -----                     Created 18/08/15  by K. Gertsenberger                     -----
// -----------------------------------------------------------------------------------------
//
// It uses *.xslt scheme with rules to parse given format:
// <skip line_count=""/> - skip given count of text lines in file
// <search name="tag name"/> - search tag with given name
// <move down="recursively child number"/> - move to child tag recursively
// <cycle child="parent name" table_name="database table to write parsing values" delimiter="delimeters between
// elements" skip="number of skipping text lines">
//   - cycle for elements separated by delimiters and write them to the database (childs of parent with given name if it
//   was set)
// <element action=["skip", "update", "write", "multi", "parse"], where:
// action="skip" - skip current element
// action="update" column_name="database column name" type="type name"/> - update database table row with database
// column name being equal the cuurent element action="write" column_name="database column name" type="type name"/> -
// convert current element to the given type (int, double, string, datetime, hex, binary) and write to the given
// database table column action="multi"> <subelement action...> <subelement action...> ... </element> action="parse"
// start_index="parse from start symbol" parse_type=["counter" - current element number from 1, "value:fixed_value",
// "parsed type name + 'int_array' or 'double_array'"] delimiter="delimeter symbol if parsed array"
// column_name="database column name" type="C++ type"/>
//   - converted value from start symbol (default, 0) OR special values to write to the database table column
// action="parse"  parse_type=["counter" - current element number from 1, "value:fixed_value"] column_name="database
// column name" type="type name"/> - special values to write to the database table column
// </cycle>
// possible data types: "int", "hex", "double", "string", "datetime", "binary"
 
#ifndef UNIPARSER_H
#define UNIPARSER_H 1
 
#include "TDatime.h"
#include "TObjArray.h"
#include "TString.h"
 
#include <boost/any.hpp>
#include <vector>
using namespace std;
 
struct structParseRow
{
    TString strColumnName;
    // int, hex, double, string, datetime, binary
    TString strStatementType;
    bool isParse;
    int iStartIndex;
    TString strParseType;
    TString strDelimiter;
 
    structParseRow(TString column_name,
                   TString statement_type,
                   bool is_parse = false,
                   int start_index = 0,
                   TString parse_type = "",
                   TString delimiter = "")
    {
        strColumnName = column_name;
        strStatementType = statement_type;
        isParse = is_parse;
        iStartIndex = start_index;
        strParseType = parse_type;
        strDelimiter = delimiter;
    }
};
 
struct structParseSchema
{
    // skip element
    bool isSkip;
    bool isUpdate;
    vector<structParseRow> vecRows;
 
    structParseSchema() { isUpdate = false; }
    structParseSchema(bool is_skip, TString column_name = "", TString statement_type = "")
    {
        isUpdate = false;
        isSkip = is_skip;
        if (is_skip == false) {
            structParseRow row(column_name, statement_type);
            vecRows.push_back(row);
        }
    }
    ~structParseSchema()
    {
        if (isSkip == false)
            vecRows.clear();
    }
};
 
struct structParseValue
{
    TDatime dtSpillEnd;
    vector<boost::any> arrValues;
};
 
class UniParser
{
  public:
    UniParser();
    virtual ~UniParser();
 
    // common parsing functions for the most cases
    int ParseXml2Db(TString xmlName, TString schemaPath, bool isUpdate = false);
    int ParseCsv2Db(TString csvName, TString schemaPath, bool isUpdate = false);
    int ParseTxt2Db(TString txtName, TString schemaPath, bool isUpdate = false);
 
    // parse text file not in a common format (should be rewritten to a common case later)
    int ParseTxtNoise2Db(int period_number, TString txtName, TString schemaPath);
 
    // parse DB fields to write to other fields (temporary function)
    int ParseDb2Db();
 
    // parse text file to structure with values
    int ParseTxt2Struct(TString txtName,
                        TString schemaPath,
                        vector<structParseValue*>& parse_values,
                        vector<structParseSchema>& vecElements,
                        int iVerbose = 1);
 
    // save text ELOG of the BM@N experiemnt in CSV format to new Elog DB (temporary function)
    int ConvertElogCsv(TString csvName = "parse_schemes/elog.csv", char separate_symbol = ';');
 
    ClassDef(UniParser, 1)
};
 
#endif