csv-reader.cc

Go to the documentation of this file.

/*
 * Copyright (c) 2019 Lawrence Livermore National Laboratory
 *
 * SPDX-License-Identifier: GPL-2.0-only
 *
 * Author: Mathew Bielejeski <bielejeski1@llnl.gov>
 */
 
#include "csv-reader.h"
 
#include "ns3/log.h"
 
#include <algorithm>
#include <cctype>
#include <fstream>
#include <iterator>
#include <limits>
#include <sstream>
#include <vector>
 
/**
 * \file
 * \ingroup csvreader
 *
 * ns3::CsvReader implementation
 */
 
NS_LOG_COMPONENT_DEFINE("CsvReader");
 
namespace
{
 
/**
 * Convert a string into another type.
 *
 * Uses a stringstream to deserialize the value stored in \p input
 * to a value of type T and writes the deserialized value to \p output.
 *
 * \tparam T Data type of output.
 * \param input String containing serialized data.
 * \param output Place to store deserialized value.
 *
 * \return \c true if deserialization was successful, \c false otherwise.
 */
template <typename T>
bool
GenericTransform(std::string input, T& output)
{
    NS_LOG_FUNCTION(input);
 
    std::istringstream stream(input);
 
    stream >> output;
 
    return static_cast<bool>(stream);
}
 
} // unnamed namespace
 
namespace ns3
{
 
CsvReader::CsvReader(const std::string& filepath, char delimiter /* =',' */)
    : m_delimiter(delimiter),
      m_rowsRead(0),
      m_fileStream(filepath),
      m_stream(&m_fileStream)
{
    NS_LOG_FUNCTION(this << filepath);
}
 
CsvReader::CsvReader(std::istream& stream, char delimiter /* =',' */)
    : m_delimiter(delimiter),
      m_rowsRead(0),
      m_fileStream(),
      m_stream(&stream)
{
    NS_LOG_FUNCTION(this);
}
 
CsvReader::~CsvReader()
{
}
 
std::size_t
CsvReader::ColumnCount() const
{
    NS_LOG_FUNCTION(this);
 
    return m_columns.size();
}
 
std::size_t
CsvReader::RowNumber() const
{
    NS_LOG_FUNCTION(this);
 
    return m_rowsRead;
}
 
char
CsvReader::Delimiter() const
{
    NS_LOG_FUNCTION(this);
 
    return m_delimiter;
}
 
bool
CsvReader::FetchNextRow()
{
    NS_LOG_FUNCTION(this);
 
    std::string line;
 
    if (m_stream->eof())
    {
        NS_LOG_LOGIC("Reached end of stream");
        return false;
    }
 
    NS_LOG_LOGIC("Reading line " << m_rowsRead + 1);
 
    std::getline(*m_stream, line);
 
    if (m_stream->fail())
    {
        NS_LOG_ERROR("Reading line " << m_rowsRead + 1 << " failed");
 
        return false;
    }
 
    ++m_rowsRead;
 
    ParseLine(line);
 
    return true;
}
 
bool
CsvReader::IsBlankRow() const
{
    return m_blankRow;
}
 
bool
CsvReader::GetValueAs(std::string input, double& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, float& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, signed char& value) const
{
    typedef signed char byte_type;
 
    NS_LOG_FUNCTION(this << input);
 
    std::istringstream tempStream(input);
 
    int16_t tempOutput = 0;
    tempStream >> tempOutput;
 
    if (tempOutput >= std::numeric_limits<byte_type>::min() &&
        tempOutput <= std::numeric_limits<byte_type>::max())
    {
        value = static_cast<byte_type>(tempOutput);
    }
 
    bool success = static_cast<bool>(tempStream);
 
    NS_LOG_DEBUG("Input='" << input << "', output=" << tempOutput << ", result=" << success);
 
    return success;
}
 
bool
CsvReader::GetValueAs(std::string input, short& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, int& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, long& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, long long& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, std::string& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    value = input;
 
    return true;
}
 
bool
CsvReader::GetValueAs(std::string input, unsigned char& value) const
{
    typedef unsigned char byte_type;
 
    NS_LOG_FUNCTION(this << input);
 
    std::istringstream tempStream(input);
 
    uint16_t tempOutput = 0;
    tempStream >> tempOutput;
 
    if (tempOutput >= std::numeric_limits<byte_type>::min() &&
        tempOutput <= std::numeric_limits<byte_type>::max())
    {
        value = static_cast<byte_type>(tempOutput);
    }
 
    bool success = static_cast<bool>(tempStream);
 
    NS_LOG_DEBUG("Input='" << input << "', output=" << tempOutput << ", result=" << success);
 
    return success;
}
 
bool
CsvReader::GetValueAs(std::string input, unsigned short& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, unsigned int& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, unsigned long& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::GetValueAs(std::string input, unsigned long long& value) const
{
    NS_LOG_FUNCTION(this << input);
 
    return GenericTransform(std::move(input), value);
}
 
bool
CsvReader::IsDelimiter(char c) const
{
    NS_LOG_FUNCTION(this << c);
 
    return c == m_delimiter;
}
 
void
CsvReader::ParseLine(const std::string& line)
{
    NS_LOG_FUNCTION(this << line);
 
    std::string value;
    m_columns.clear();
 
    auto start_col = line.begin();
    auto end_col = line.end();
 
    while (start_col != line.end())
    {
        std::tie(value, end_col) = ParseColumn(start_col, line.end());
 
        NS_LOG_DEBUG("ParseColumn() returned: " << value);
 
        m_columns.push_back(std::move(value));
 
        if (end_col != line.end())
        {
            ++end_col;
        }
 
        start_col = end_col;
    }
    m_blankRow = (m_columns.size() == 1) && (m_columns[0].empty());
    NS_LOG_LOGIC("blank row: " << m_blankRow);
}
 
std::tuple<std::string, std::string::const_iterator>
CsvReader::ParseColumn(std::string::const_iterator begin, std::string::const_iterator end)
{
    NS_LOG_FUNCTION(this << std::string(begin, end));
 
    enum class State
    {
        BEGIN,
        END_QUOTE,
        FIND_DELIMITER,
        QUOTED_STRING,
        UNQUOTED_STRING,
        END
    };
 
    State state = State::BEGIN;
    std::string buffer;
    auto iter = begin;
 
    while (state != State::END)
    {
        if (iter == end)
        {
            NS_LOG_DEBUG("Found end iterator, switching to END state");
 
            state = State::END;
            continue;
        }
 
        auto c = *iter;
 
        NS_LOG_DEBUG("Next character: '" << c << "'");
 
        // handle common cases here to avoid duplicating logic
        if (state != State::QUOTED_STRING)
        {
            if (IsDelimiter(c))
            {
                NS_LOG_DEBUG("Found field delimiter, switching to END state");
 
                if (state == State::UNQUOTED_STRING)
                {
                    NS_LOG_DEBUG("Removing trailing whitespace from unquoted field: '" << buffer
                                                                                       << "'");
                    auto len = buffer.size();
 
                    // remove trailing whitespace from the field
                    while (!buffer.empty() &&
                           std::isspace(static_cast<unsigned char>(buffer.back())))
                    {
                        buffer.pop_back();
                    }
 
                    auto finalLen = buffer.size();
 
                    NS_LOG_DEBUG("Removed " << (len - finalLen)
                                            << " trailing whitespace characters");
                }
 
                state = State::END;
 
                continue;
            }
            else if (c == '#')
            {
                NS_LOG_DEBUG("Found start of comment, switching to END state");
 
                // comments consume the rest of the line, set iter to end
                // to reflect that fact.
                iter = end;
                state = State::END;
 
                continue;
            }
        }
 
        switch (state)
        {
        case State::BEGIN: {
            if (c == '"')
            {
                NS_LOG_DEBUG("Switching state: BEGIN -> QUOTED_STRING");
 
                state = State::QUOTED_STRING;
            }
            else if (!std::isspace(c))
            {
                NS_LOG_DEBUG("Switching state: BEGIN -> UNQUOTED_STRING");
 
                state = State::UNQUOTED_STRING;
                buffer.push_back(c);
            }
        }
        break;
        case State::QUOTED_STRING: {
            if (c == '"')
            {
                NS_LOG_DEBUG("Switching state: QUOTED_STRING -> END_QUOTE");
                state = State::END_QUOTE;
            }
            else
            {
                buffer.push_back(c);
            }
        }
        break;
        case State::END_QUOTE: {
            if (c == '"')
            {
                NS_LOG_DEBUG("Switching state: END_QUOTE -> QUOTED_STRING");
 
                // an escape quote instead of an end quote
                state = State::QUOTED_STRING;
                buffer.push_back(c);
            }
            else
            {
                NS_LOG_DEBUG("Switching state: END_QUOTE -> FIND_DELIMITER");
                state = State::FIND_DELIMITER;
            }
        }
        break;
        case State::UNQUOTED_STRING: {
            buffer.push_back(c);
        }
        break;
        case State::FIND_DELIMITER:
        case State::END:
            break;
        }
 
        ++iter;
    }
 
    NS_LOG_DEBUG("Field value: " << buffer);
 
    return std::make_tuple(buffer, iter);
}
 
} // namespace ns3