annotate data/fileio/CSVFormat.h @ 1021:1888ca033a84

If the first column of a CSV file has zero or one non-empty values in it, during the first 5 rows, then ignore it by default
author Chris Cannam
date Mon, 01 Dec 2014 10:18:55 +0000
parents e369dd281cf2
children 26cf6d5251ec
rev   line source
Chris@392 1 /* -*- c-basic-offset: 4 indent-tabs-mode: nil -*- vi:set ts=8 sts=4 sw=4: */
Chris@392 2
Chris@392 3 /*
Chris@392 4 Sonic Visualiser
Chris@392 5 An audio file viewer and annotation editor.
Chris@392 6 Centre for Digital Music, Queen Mary, University of London.
Chris@392 7 This file copyright 2006 Chris Cannam.
Chris@392 8
Chris@392 9 This program is free software; you can redistribute it and/or
Chris@392 10 modify it under the terms of the GNU General Public License as
Chris@392 11 published by the Free Software Foundation; either version 2 of the
Chris@392 12 License, or (at your option) any later version. See the file
Chris@392 13 COPYING included with this distribution for more information.
Chris@392 14 */
Chris@392 15
Chris@392 16 #ifndef _CSV_FORMAT_H_
Chris@392 17 #define _CSV_FORMAT_H_
Chris@392 18
Chris@392 19 #include <QString>
Chris@392 20 #include <QStringList>
Chris@392 21
Chris@392 22 class CSVFormat
Chris@392 23 {
Chris@392 24 public:
Chris@392 25 enum ModelType {
Chris@392 26 OneDimensionalModel,
Chris@392 27 TwoDimensionalModel,
Chris@628 28 TwoDimensionalModelWithDuration,
Chris@897 29 TwoDimensionalModelWithDurationAndPitch,
Chris@392 30 ThreeDimensionalModel
Chris@392 31 };
Chris@392 32
Chris@392 33 enum TimingType {
Chris@392 34 ExplicitTiming,
Chris@392 35 ImplicitTiming
Chris@392 36 };
Chris@628 37
Chris@392 38 enum TimeUnits {
Chris@392 39 TimeSeconds,
Chris@990 40 TimeMilliseconds,
Chris@392 41 TimeAudioFrames,
Chris@990 42 TimeWindows,
Chris@392 43 };
Chris@392 44
Chris@629 45 enum ColumnPurpose {
Chris@629 46 ColumnUnknown,
Chris@629 47 ColumnStartTime,
Chris@629 48 ColumnEndTime,
Chris@629 49 ColumnDuration,
Chris@629 50 ColumnValue,
Chris@897 51 ColumnPitch,
Chris@629 52 ColumnLabel
Chris@629 53 };
Chris@629 54
Chris@629 55 enum ColumnQuality {
Chris@1021 56 ColumnNumeric = 1,
Chris@1021 57 ColumnIntegral = 2,
Chris@1021 58 ColumnIncreasing = 4,
Chris@1021 59 ColumnLarge = 8,
Chris@1021 60 ColumnNearEmpty = 16,
Chris@629 61 };
Chris@629 62 typedef unsigned int ColumnQualities;
Chris@392 63
Chris@392 64 CSVFormat() : // arbitrary defaults
Chris@392 65 m_modelType(TwoDimensionalModel),
Chris@392 66 m_timingType(ExplicitTiming),
Chris@392 67 m_timeUnits(TimeSeconds),
Chris@392 68 m_separator(","),
Chris@392 69 m_sampleRate(44100),
Chris@392 70 m_windowSize(1024),
Chris@629 71 m_columnCount(0),
Chris@629 72 m_variableColumnCount(false),
Chris@629 73 m_allowQuoting(true),
Chris@629 74 m_maxExampleCols(0)
Chris@392 75 { }
Chris@629 76
Chris@629 77 CSVFormat(QString path); // guess format
Chris@629 78
Chris@629 79 /**
Chris@629 80 * Guess the format of the given CSV file, setting the fields in
Chris@629 81 * this object accordingly. If the current separator is the empty
Chris@629 82 * string, the separator character will also be guessed; otherwise
Chris@629 83 * the current separator will be used. The other properties of
Chris@629 84 * this object will be set according to guesses from the file.
Chris@629 85 */
Chris@629 86 void guessFormatFor(QString path);
Chris@628 87
Chris@628 88 ModelType getModelType() const { return m_modelType; }
Chris@628 89 TimingType getTimingType() const { return m_timingType; }
Chris@628 90 TimeUnits getTimeUnits() const { return m_timeUnits; }
Chris@929 91 int getSampleRate() const { return m_sampleRate; }
Chris@929 92 int getWindowSize() const { return m_windowSize; }
Chris@630 93 int getColumnCount() const { return m_columnCount; }
Chris@631 94 bool getAllowQuoting() const { return m_allowQuoting; }
Chris@631 95 QChar getSeparator() const {
Chris@631 96 if (m_separator == "") return ' ';
Chris@631 97 else return m_separator[0];
Chris@631 98 }
Chris@630 99
Chris@628 100 void setModelType(ModelType t) { m_modelType = t; }
Chris@628 101 void setTimingType(TimingType t) { m_timingType = t; }
Chris@628 102 void setTimeUnits(TimeUnits t) { m_timeUnits = t; }
Chris@631 103 void setSeparator(QChar s) { m_separator = s; }
Chris@1009 104 void setSampleRate(int r) { m_sampleRate = r; }
Chris@1009 105 void setWindowSize(int s) { m_windowSize = s; }
Chris@630 106 void setColumnCount(int c) { m_columnCount = c; }
Chris@631 107 void setAllowQuoting(bool q) { m_allowQuoting = q; }
Chris@392 108
Chris@631 109 QList<ColumnPurpose> getColumnPurposes() const { return m_columnPurposes; }
Chris@629 110 void setColumnPurposes(QList<ColumnPurpose> cl) { m_columnPurposes = cl; }
Chris@631 111
Chris@631 112 ColumnPurpose getColumnPurpose(int i);
Chris@631 113 ColumnPurpose getColumnPurpose(int i) const;
Chris@631 114 void setColumnPurpose(int i, ColumnPurpose p);
Chris@392 115
Chris@629 116 // read-only; only valid if format has been guessed:
Chris@629 117 QList<ColumnQualities> getColumnQualities() const { return m_columnQualities; }
Chris@629 118
Chris@629 119 // read-only; only valid if format has been guessed:
Chris@392 120 QList<QStringList> getExample() const { return m_example; }
Chris@392 121 int getMaxExampleCols() const { return m_maxExampleCols; }
Chris@631 122
Chris@392 123 protected:
Chris@628 124 ModelType m_modelType;
Chris@628 125 TimingType m_timingType;
Chris@628 126 TimeUnits m_timeUnits;
Chris@628 127 QString m_separator;
Chris@929 128 int m_sampleRate;
Chris@929 129 int m_windowSize;
Chris@392 130
Chris@629 131 int m_columnCount;
Chris@629 132 bool m_variableColumnCount;
Chris@629 133
Chris@629 134 QList<ColumnQualities> m_columnQualities;
Chris@629 135 QList<ColumnPurpose> m_columnPurposes;
Chris@629 136
Chris@629 137 QList<float> m_prevValues;
Chris@629 138
Chris@629 139 bool m_allowQuoting;
Chris@392 140
Chris@392 141 QList<QStringList> m_example;
Chris@392 142 int m_maxExampleCols;
Chris@629 143
Chris@629 144 void guessSeparator(QString line);
Chris@629 145 void guessQualities(QString line, int lineno);
Chris@629 146 void guessPurposes();
Chris@629 147
Chris@629 148 void guessFormatFor_Old(QString path);
Chris@629 149
Chris@392 150 };
Chris@392 151
Chris@392 152 #endif