CSV import. Replaced old ASCII parser by the new RifCsvUserDataParser

This commit is contained in:
Bjørn Erik Jensen
2017-11-28 08:51:13 +01:00
parent 3b1775ad4d
commit 6d4820a691
12 changed files with 177 additions and 301 deletions

View File

@@ -16,7 +16,6 @@ ${CEE_CURRENT_LIST_DIR}RifPerforationIntervalReader.h
${CEE_CURRENT_LIST_DIR}RifReaderEclipseInput.h
${CEE_CURRENT_LIST_DIR}RifReaderEclipseOutput.h
${CEE_CURRENT_LIST_DIR}RifSummaryReaderInterface.h
${CEE_CURRENT_LIST_DIR}RifColumnBasedAsciiParser.h
${CEE_CURRENT_LIST_DIR}RifEclipseUserDataParserTools.h
${CEE_CURRENT_LIST_DIR}RifColumnBasedUserDataParser.h
${CEE_CURRENT_LIST_DIR}RifKeywordVectorParser.h
@@ -63,7 +62,6 @@ ${CEE_CURRENT_LIST_DIR}RifPerforationIntervalReader.cpp
${CEE_CURRENT_LIST_DIR}RifReaderEclipseInput.cpp
${CEE_CURRENT_LIST_DIR}RifReaderEclipseOutput.cpp
${CEE_CURRENT_LIST_DIR}RifSummaryReaderInterface.cpp
${CEE_CURRENT_LIST_DIR}RifColumnBasedAsciiParser.cpp
${CEE_CURRENT_LIST_DIR}RifEclipseUserDataParserTools.cpp
${CEE_CURRENT_LIST_DIR}RifColumnBasedUserDataParser.cpp
${CEE_CURRENT_LIST_DIR}RifKeywordVectorParser.cpp

View File

@@ -1,129 +0,0 @@
/////////////////////////////////////////////////////////////////////////////////
//
// Copyright (C) 2017- Statoil ASA
//
// ResInsight is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// ResInsight 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 at <http://www.gnu.org/licenses/gpl.html>
// for more details.
//
/////////////////////////////////////////////////////////////////////////////////
#include "RifColumnBasedAsciiParser.h"
#include "RiaLogging.h"
#include "cvfAssert.h"
#include <QString>
#include <QStringList>
#include <QTextStream>
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
RifColumnBasedAsciiParser::RifColumnBasedAsciiParser(QString& data, const QString dateFormat, QLocale decimalLocale, QString cellSeparator)
{
parseData(data, dateFormat, decimalLocale, cellSeparator);
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
const std::vector<QString>& RifColumnBasedAsciiParser::headers() const
{
return m_data.m_headers;
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
const std::vector<QDateTime>& RifColumnBasedAsciiParser::timeSteps() const
{
return m_data.m_timeSteps;
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
const std::vector<double>& RifColumnBasedAsciiParser::columnValues(size_t columnIndex) const
{
CVF_TIGHT_ASSERT(columnIndex < m_data.m_values.size());
return m_data.m_values[columnIndex];
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
void RifColumnBasedAsciiParser::parseData(QString& data, QString dateFormat, QLocale decimalLocale, QString cellSeparator)
{
QTextStream tableData(&data);
QString header;
do {
header = tableData.readLine();
} while (header.isEmpty() && !tableData.atEnd());
// No header row found
if (header.isEmpty()) return;
QStringList columnHeaders = header.split(cellSeparator);
for (int i = 1; i < columnHeaders.size(); ++i)
{
m_data.m_headers.push_back(columnHeaders[i]);
}
// No columns found
if (m_data.m_headers.empty()) return;
int numColumns = static_cast<int>(m_data.m_headers.size());
m_data.m_values.resize(numColumns);
size_t row = 0;
while (!tableData.atEnd())
{
++row;
QString line = tableData.readLine();
// Skip empty lines
if (line.isEmpty()) continue;
QStringList columns = line.split(cellSeparator);
if (columns.size() != numColumns + 1)
{
RiaLogging::warning(QString("Invalid number of columns in row %1").arg(row));
continue;
}
QDateTime date = QDateTime::fromString(columns[0], dateFormat);
if (!date.isValid())
{
RiaLogging::warning(QString("First column of row %1 could not be parsed as a date: %2").arg(row).arg(columns[0]));
continue;
}
m_data.m_timeSteps.push_back(date);
for (int col = 1; col < columns.size(); ++col)
{
bool ok;
m_data.m_values[col - 1].push_back(decimalLocale.toDouble(columns[col], &ok));
if (!ok)
{
RiaLogging::warning(QString("Could not parse value at row %1 column %2 as double: %3. Defaulting to 0.0").arg(row).arg(col).arg(columns[col]));
}
}
}
}

View File

@@ -1,53 +0,0 @@
/////////////////////////////////////////////////////////////////////////////////
//
// Copyright (C) 2017- Statoil ASA
//
// ResInsight is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// ResInsight 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 at <http://www.gnu.org/licenses/gpl.html>
// for more details.
//
/////////////////////////////////////////////////////////////////////////////////
#pragma once
#include <QDateTime>
#include <QLocale>
#include <QString>
#include <vector>
//==================================================================================================
///
//==================================================================================================
class RifColumnBasedAsciiParser
{
public:
RifColumnBasedAsciiParser(QString& data, const QString dateFormat, QLocale decimalLocale, QString cellSeparator);
const std::vector<QString>& headers() const;
const std::vector<QDateTime>& timeSteps() const;
const std::vector<double>& columnValues(size_t columnIndex) const;
private:
void parseData(QString& data, QString dateFormat, QLocale decimalLocale, QString cellSeparator);
private:
struct AsciiData
{
std::vector<QString> m_headers;
std::vector<QDateTime> m_timeSteps;
std::vector< std::vector<double> > m_values;
};
AsciiData m_data;
};

View File

@@ -79,6 +79,21 @@ const ColumnInfo* RifCsvUserDataParser::columnInfo(size_t columnIndex) const
return &(m_tableData.columnInfos()[columnIndex]);
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
const ColumnInfo* RifCsvUserDataParser::dateTimeColumn() const
{
for (const ColumnInfo& col : m_tableData.columnInfos())
{
if (col.dataType == ColumnInfo::DATETIME)
{
return &col;
}
}
return nullptr;
}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------

View File

@@ -47,6 +47,7 @@ public:
const TableData& tableData() const;
const ColumnInfo* columnInfo(size_t columnIndex) const;
const ColumnInfo* dateTimeColumn() const;
bool parseColumnInfo(const QString& cellSeparator);
QString previewText(int lineCount, const QString& cellSeparator);

View File

@@ -918,24 +918,13 @@ bool RifEclipseUserDataParserTools::isScalingText(const std::string& word)
return word.find_first_of('*') != std::string::npos;
}
////--------------------------------------------------------------------------------------------------
/////
////--------------------------------------------------------------------------------------------------
//ColumnInfo::~ColumnInfo()
//{
// if (values)
// {
// delete values;
// }
// if (textValues)
// {
// delete textValues;
// }
// if (dateTimeValues)
// {
// delete dateTimeValues;
// }
//}
//--------------------------------------------------------------------------------------------------
///
//--------------------------------------------------------------------------------------------------
std::string ColumnInfo::columnName() const
{
return summaryAddress.quantityName();
}
//--------------------------------------------------------------------------------------------------
///

View File

@@ -56,6 +56,7 @@ public:
{
}
std::string columnName() const;
size_t itemCount() const;
public:

View File

@@ -18,8 +18,10 @@
#include "RifReaderObservedData.h"
#include "RifColumnBasedAsciiParser.h"
#include "RifEclipseSummaryAddress.h"
#include "RifCsvUserDataParser.h"
#include "SummaryPlotCommands/RicPasteAsciiDataToSummaryPlotFeatureUi.h"
#include <QDateTime>
#include <QTextStream>
@@ -47,9 +49,10 @@ bool RifReaderObservedData::open(const QString& headerFileName,
const QString& identifierName,
RifEclipseSummaryAddress::SummaryVarCategory summaryCategory)
{
QString dateFormat = "yyyy-MM-dd";
QString cellSeparator = "\t";
QLocale decimalLocale = QLocale::Norwegian;
AsciiDataParseOptions parseOptions;
parseOptions.dateFormat = "yyyy-MM-dd";
parseOptions.cellSeparator = "\t";
parseOptions.locale = QLocale::Norwegian;
QString data;
QTextStream out(&data);
@@ -60,27 +63,29 @@ bool RifReaderObservedData::open(const QString& headerFileName,
out << "1994-05-23" << "\t" << "40" << "\t" << "4" << "\n";
m_asciiParser = std::unique_ptr<RifColumnBasedAsciiParser>(new RifColumnBasedAsciiParser(data, dateFormat, decimalLocale, cellSeparator));
m_asciiParser = std::unique_ptr<RifCsvUserDataParser>(new RifCsvUserDataPastedTextParser(data));
m_timeSteps.clear();
if (m_asciiParser)
if (m_asciiParser->parse(parseOptions))
{
for (QDateTime timeStep : m_asciiParser->timeSteps())
if (m_asciiParser && m_asciiParser->dateTimeColumn())
{
time_t t = timeStep.toTime_t();
m_timeSteps.push_back(t);
for (QDateTime timeStep : m_asciiParser->dateTimeColumn()->dateTimeValues)
{
time_t t = timeStep.toTime_t();
m_timeSteps.push_back(t);
}
m_allResultAddresses.clear();
for (auto s : m_asciiParser->tableData().columnInfos())
{
m_allResultAddresses.push_back(s.summaryAddress);
}
}
m_allResultAddresses.clear();
for (auto s : m_asciiParser->headers())
{
m_allResultAddresses.push_back(address(s, identifierName, summaryCategory));
}
return true;
}
if (!m_asciiParser) return false;
return true;
return false;
}
//--------------------------------------------------------------------------------------------------
@@ -100,9 +105,13 @@ bool RifReaderObservedData::values(const RifEclipseSummaryAddress& resultAddress
if (columnIndex != m_allResultAddresses.size())
{
for (auto& v : m_asciiParser->columnValues(columnIndex))
const ColumnInfo* col = m_asciiParser->columnInfo(columnIndex);
if (col && col->dataType == ColumnInfo::NUMERIC)
{
values->push_back(v);
for (auto& v : col->values)
{
values->push_back(v);
}
}
}

View File

@@ -30,6 +30,7 @@ class QString;
class RifColumnBasedAsciiParser;
class RifEclipseSummaryAddress;
class RifCsvUserDataParser;
//==================================================================================================
//
@@ -58,7 +59,7 @@ private:
RifEclipseSummaryAddress::SummaryVarCategory summaryCategory);
private:
std::unique_ptr<RifColumnBasedAsciiParser> m_asciiParser;
std::unique_ptr<RifCsvUserDataParser> m_asciiParser;
std::vector<time_t> m_timeSteps;
};