@@ -1,185 +1,227 | |||
|
1 | 1 | #ifndef SCIQLOP_DATASERIESUTILS_H |
|
2 | 2 | #define SCIQLOP_DATASERIESUTILS_H |
|
3 | 3 | |
|
4 | 4 | #include "CoreGlobal.h" |
|
5 | 5 | |
|
6 | #include <Data/DataSeriesIterator.h> | |
|
7 | ||
|
6 | 8 | #include <QLoggingCategory> |
|
7 | 9 | |
|
8 | 10 | Q_DECLARE_LOGGING_CATEGORY(LOG_DataSeriesUtils) |
|
9 | 11 | |
|
10 | 12 | /** |
|
11 | 13 | * Utility class with methods for data series |
|
12 | 14 | */ |
|
13 | 15 | struct SCIQLOP_CORE_EXPORT DataSeriesUtils { |
|
14 | 16 | /** |
|
15 | 17 | * Define a meshs. |
|
16 | 18 | * |
|
17 | 19 | * A mesh is a regular grid representing cells of the same width (in x) and of the same height |
|
18 | 20 | * (in y). At each mesh point is associated a value. |
|
19 | 21 | * |
|
20 | 22 | * Each axis of the mesh is defined by a minimum value, a number of values is a mesh step. |
|
21 | 23 | * For example: if min = 1, nbValues = 5 and step = 2 => the axis of the mesh will be [1, 3, 5, |
|
22 | 24 | * 7, 9]. |
|
23 | 25 | * |
|
24 | 26 | * The values are defined in an array of size {nbX * nbY}. The data is stored along the X axis. |
|
25 | 27 | * |
|
26 | 28 | * For example, the mesh: |
|
27 | 29 | * Y = 2 [ 7 ; 8 ; 9 |
|
28 | 30 | * Y = 1 4 ; 5 ; 6 |
|
29 | 31 | * Y = 0 1 ; 2 ; 3 ] |
|
30 | 32 | * X = 0 X = 1 X = 2 |
|
31 | 33 | * |
|
32 | 34 | * will be represented by data [1, 2, 3, 4, 5, 6, 7, 8, 9] |
|
33 | 35 | */ |
|
34 | 36 | struct Mesh { |
|
35 | 37 | explicit Mesh() = default; |
|
36 | 38 | explicit Mesh(int nbX, double xMin, double xStep, int nbY, double yMin, double yStep) |
|
37 | 39 | : m_NbX{nbX}, |
|
38 | 40 | m_XMin{xMin}, |
|
39 | 41 | m_XStep{xStep}, |
|
40 | 42 | m_NbY{nbY}, |
|
41 | 43 | m_YMin{yMin}, |
|
42 | 44 | m_YStep{yStep}, |
|
43 | 45 | m_Data(nbX * nbY) |
|
44 | 46 | { |
|
45 | 47 | } |
|
46 | 48 | |
|
47 | 49 | inline bool isEmpty() const { return m_Data.size() == 0; } |
|
48 | 50 | inline double xMax() const { return m_XMin + (m_NbX - 1) * m_XStep; } |
|
49 | 51 | inline double yMax() const { return m_YMin + (m_NbY - 1) * m_YStep; } |
|
50 | 52 | |
|
51 | 53 | int m_NbX{0}; |
|
52 | 54 | double m_XMin{}; |
|
53 | 55 | double m_XStep{}; |
|
54 | 56 | int m_NbY{0}; |
|
55 | 57 | double m_YMin{}; |
|
56 | 58 | double m_YStep{}; |
|
57 | 59 | std::vector<double> m_Data{}; |
|
58 | 60 | }; |
|
59 | 61 | |
|
60 | 62 | /** |
|
61 | 63 | * Represents a resolution used to generate the data of a mesh on the x-axis or in Y. |
|
62 | 64 | * |
|
63 | 65 | * A resolution is represented by a value and flag indicating if it's in the logarithmic scale |
|
64 | 66 | * @sa Mesh |
|
65 | 67 | */ |
|
66 | 68 | struct Resolution { |
|
67 | 69 | double m_Val{std::numeric_limits<double>::quiet_NaN()}; |
|
68 | 70 | bool m_Logarithmic{false}; |
|
69 | 71 | }; |
|
70 | 72 | |
|
71 | 73 | /** |
|
72 | 74 | * Processes data from a data series to complete the data holes with a fill value. |
|
73 | 75 | * |
|
74 | 76 | * A data hole is determined by the resolution passed in parameter: if, between two continuous |
|
75 | 77 | * data on the x-axis, the difference between these data is greater than the resolution, then |
|
76 | 78 | * there is one or more holes between them. The holes are filled by adding: |
|
77 | 79 | * - for the x-axis, new data corresponding to the 'step resolution' starting from the first |
|
78 | 80 | * data; |
|
79 | 81 | * - for values, a default value (fill value) for each new data added on the x-axis. |
|
80 | 82 | * |
|
81 | 83 | * For example, with : |
|
82 | 84 | * - xAxisData = [0, 1, 5, 7, 14 ] |
|
83 | 85 | * - valuesData = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] (two components per x-axis data) |
|
84 | 86 | * - fillValue = NaN |
|
85 | 87 | * - and resolution = 2; |
|
86 | 88 | * |
|
87 | 89 | * For the x axis, we calculate as data holes: [3, 9, 11, 13]. These holes are added to the |
|
88 | 90 | * x-axis data, and NaNs (two per x-axis data) are added to the values: |
|
89 | 91 | * => xAxisData = [0, 1, 3, 5, 7, 9, 11, 13, 14 ] |
|
90 | 92 | * => valuesData = [0, 1, 2, 3, NaN, NaN, 4, 5, 6, 7, NaN, NaN, NaN, NaN, NaN, NaN, 8, 9] |
|
91 | 93 | * |
|
92 | 94 | * It is also possible to set bounds for the data series. If these bounds are defined and exceed |
|
93 | 95 | * the limits of the data series, data holes are added to the series at the beginning and/or the |
|
94 | 96 | * end. |
|
95 | 97 | * |
|
96 | 98 | * The generation of data holes at the beginning/end of the data series is performed starting |
|
97 | 99 | * from the x-axis series limit and adding data holes at each 'resolution step' as long as the |
|
98 | 100 | * new bound is not reached. |
|
99 | 101 | * |
|
100 | 102 | * For example, with : |
|
101 | 103 | * - xAxisData = [3, 4, 5, 6, 7 ] |
|
102 | 104 | * - valuesData = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] |
|
103 | 105 | * - fillValue = NaN |
|
104 | 106 | * - minBound = 0 |
|
105 | 107 | * - maxBound = 12 |
|
106 | 108 | * - and resolution = 2; |
|
107 | 109 | * |
|
108 | 110 | * => Starting from 3 and decreasing 2 by 2 until reaching 0 : a data hole at value 1 will be |
|
109 | 111 | * added to the beginning of the series |
|
110 | 112 | * => Starting from 7 and increasing 2 by 2 until reaching 12 : data holes at values 9 and 11 |
|
111 | 113 | * will be added to the end of the series |
|
112 | 114 | * |
|
113 | 115 | * So : |
|
114 | 116 | * => xAxisData = [1, 3, 4, 5, 6, 7, 9, 11 ] |
|
115 | 117 | * => valuesData = [NaN, NaN, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, NaN, NaN, NaN, NaN] |
|
116 | 118 | * |
|
117 | 119 | * @param xAxisData the x-axis data of the data series |
|
118 | 120 | * @param valuesData the values data of the data series |
|
119 | 121 | * @param resolution the resoultion (on x-axis) used to determinate data holes |
|
120 | 122 | * @param fillValue the fill value used for data holes in the values data |
|
121 | 123 | * @param minBound the limit at which to start filling data holes for the series. If set to NaN, |
|
122 | 124 | * the limit is not used |
|
123 | 125 | * @param maxBound the limit at which to end filling data holes for the series. If set to NaN, |
|
124 | 126 | * the limit is not used |
|
125 | 127 | * |
|
126 | 128 | * @remarks There is no control over the consistency between x-axis data and values data. The |
|
127 | 129 | * method considers that the data is well formed (the total number of values data is a multiple |
|
128 | 130 | * of the number of x-axis data) |
|
129 | 131 | */ |
|
130 | 132 | static void fillDataHoles(std::vector<double> &xAxisData, std::vector<double> &valuesData, |
|
131 | 133 | double resolution, |
|
132 | 134 | double fillValue = std::numeric_limits<double>::quiet_NaN(), |
|
133 | 135 | double minBound = std::numeric_limits<double>::quiet_NaN(), |
|
134 | 136 | double maxBound = std::numeric_limits<double>::quiet_NaN()); |
|
135 | 137 | /** |
|
136 | 138 | * Computes the resolution of a dataset passed as a parameter. |
|
137 | 139 | * |
|
138 | 140 | * The resolution of a dataset is the minimum difference between two values that follow in the |
|
139 | 141 | * set. |
|
140 | 142 | * For example: |
|
141 | 143 | * - for the set [0, 2, 4, 8, 10, 11, 13] => the resolution is 1 (difference between 10 and 11). |
|
142 | 144 | * |
|
143 | 145 | * A resolution can be calculated on the logarithmic scale (base of 10). In this case, the |
|
144 | 146 | * dataset is first converted to logarithmic values. |
|
145 | 147 | * For example: |
|
146 | 148 | * - for the set [10, 100, 10000, 1000000], the values are converted to [1, 2, 4, 6] => the |
|
147 | 149 | * logarithmic resolution is 1 (difference between 1 and 2). |
|
148 | 150 | * |
|
149 | 151 | * @param begin the iterator pointing to the beginning of the dataset |
|
150 | 152 | * @param end the iterator pointing to the end of the dataset |
|
151 | 153 | * @param logarithmic computes a logarithmic resolution or not |
|
152 | 154 | * @return the resolution computed |
|
153 | 155 | * @warning the method considers the dataset as sorted and doesn't control it. |
|
154 | 156 | */ |
|
155 | 157 | template <typename Iterator> |
|
156 | 158 | static Resolution resolution(Iterator begin, Iterator end, bool logarithmic = false); |
|
159 | ||
|
160 | /** | |
|
161 | * Computes a regular mesh for a data series, according to resolutions for x-axis and y-axis | |
|
162 | * passed as parameters. | |
|
163 | * | |
|
164 | * The mesh is created from the resolutions in x and y and the boundaries delimiting the data | |
|
165 | * series. If the resolutions do not allow to obtain a regular mesh, they are recalculated. | |
|
166 | * | |
|
167 | * For example : | |
|
168 | * Let x-axis data = [0, 1, 3, 5, 9], its associated values ββ= [0, 10, 30, 50, 90] and | |
|
169 | * xResolution = 2. | |
|
170 | * Based on the resolution, the mesh would be [0, 2, 4, 6, 8, 10] and would be invalid because | |
|
171 | * it exceeds the maximum bound of the data. The resolution is thus recalculated so that the | |
|
172 | * mesh holds between the data terminals. | |
|
173 | * So => resolution is 1.8 and the mesh is [0, 1.8, 3.6, 5.4, 7.2, 9]. | |
|
174 | * | |
|
175 | * Once the mesh is generated in x and y, the values ββare associated with each mesh point, | |
|
176 | * based on the data in the series, finding the existing data at which the mesh point would be | |
|
177 | * or would be closest to, without exceeding it. | |
|
178 | * | |
|
179 | * In the example, we determine the value of each mesh point: | |
|
180 | * - x = 0 => value = 0 (existing x in the data series) | |
|
181 | * - x = 1.8 => value = 10 (the closest existing x: 1) | |
|
182 | * - x = 3.6 => value = 30 (the closest existing x: 3) | |
|
183 | * - x = 5.4 => value = 50 (the closest existing x: 5) | |
|
184 | * - x = 7.2 => value = 50 (the closest existing x: 5) | |
|
185 | * - x = 9 => value = 90 (existing x in the data series) | |
|
186 | * | |
|
187 | * Same algorithm is applied for y-axis. | |
|
188 | * | |
|
189 | * @param begin the iterator pointing to the beginning of the data series | |
|
190 | * @param end the iterator pointing to the end of the data series | |
|
191 | * @param xResolution the resolution expected for the mesh's x-axis | |
|
192 | * @param yResolution the resolution expected for the mesh's y-axis | |
|
193 | * @return the mesh created, an empty mesh if the input data do not allow to generate a regular | |
|
194 | * mesh (empty data, null resolutions, logarithmic x-axis) | |
|
195 | * @warning the method considers the dataset as sorted and doesn't control it. | |
|
196 | */ | |
|
197 | static Mesh regularMesh(DataSeriesIterator begin, DataSeriesIterator end, | |
|
198 | Resolution xResolution, Resolution yResolution); | |
|
157 | 199 | }; |
|
158 | 200 | |
|
159 | 201 | template <typename Iterator> |
|
160 | 202 | DataSeriesUtils::Resolution DataSeriesUtils::resolution(Iterator begin, Iterator end, |
|
161 | 203 | bool logarithmic) |
|
162 | 204 | { |
|
163 | 205 | // Retrieves data into a work dataset |
|
164 | 206 | using ValueType = typename Iterator::value_type; |
|
165 | 207 | std::vector<ValueType> values{}; |
|
166 | 208 | std::copy(begin, end, std::back_inserter(values)); |
|
167 | 209 | |
|
168 | 210 | // Converts data if logarithmic flag is activated |
|
169 | 211 | if (logarithmic) { |
|
170 | 212 | std::for_each(values.begin(), values.end(), |
|
171 | 213 | [logarithmic](auto &val) { val = std::log10(val); }); |
|
172 | 214 | } |
|
173 | 215 | |
|
174 | 216 | // Computes the differences between the values in the dataset |
|
175 | 217 | std::adjacent_difference(values.begin(), values.end(), values.begin()); |
|
176 | 218 | |
|
177 | 219 | // Retrieves the smallest difference |
|
178 | 220 | auto resolutionIt = std::min_element(values.begin(), values.end()); |
|
179 | 221 | auto resolution |
|
180 | 222 | = resolutionIt != values.end() ? *resolutionIt : std::numeric_limits<double>::quiet_NaN(); |
|
181 | 223 | |
|
182 | 224 | return Resolution{resolution, logarithmic}; |
|
183 | 225 | } |
|
184 | 226 | |
|
185 | 227 | #endif // SCIQLOP_DATASERIESUTILS_H |
@@ -1,81 +1,116 | |||
|
1 | 1 | #include "Data/DataSeriesUtils.h" |
|
2 | 2 | |
|
3 | #include <cmath> | |
|
4 | ||
|
3 | 5 | Q_LOGGING_CATEGORY(LOG_DataSeriesUtils, "DataSeriesUtils") |
|
4 | 6 | |
|
5 | 7 | void DataSeriesUtils::fillDataHoles(std::vector<double> &xAxisData, std::vector<double> &valuesData, |
|
6 | 8 | double resolution, double fillValue, double minBound, |
|
7 | 9 | double maxBound) |
|
8 | 10 | { |
|
9 | 11 | if (resolution == 0. || std::isnan(resolution)) { |
|
10 | 12 | qCWarning(LOG_DataSeriesUtils()) |
|
11 | 13 | << "Can't fill data holes with a null resolution, no changes will be made"; |
|
12 | 14 | return; |
|
13 | 15 | } |
|
14 | 16 | |
|
15 | 17 | if (xAxisData.empty()) { |
|
16 | 18 | qCWarning(LOG_DataSeriesUtils()) |
|
17 | 19 | << "Can't fill data holes for empty data, no changes will be made"; |
|
18 | 20 | return; |
|
19 | 21 | } |
|
20 | 22 | |
|
21 | 23 | // Gets the number of values per x-axis data |
|
22 | 24 | auto nbComponents = valuesData.size() / xAxisData.size(); |
|
23 | 25 | |
|
24 | 26 | // Generates fill values that will be used to complete values data |
|
25 | 27 | std::vector<double> fillValues(nbComponents, fillValue); |
|
26 | 28 | |
|
27 | 29 | // Checks if there are data holes on the beginning of the data and generates the hole at the |
|
28 | 30 | // extremity if it's the case |
|
29 | 31 | auto minXAxisData = xAxisData.front(); |
|
30 | 32 | if (!std::isnan(minBound) && minBound < minXAxisData) { |
|
31 | 33 | auto holeSize = static_cast<int>((minXAxisData - minBound) / resolution); |
|
32 | 34 | if (holeSize > 0) { |
|
33 | 35 | xAxisData.insert(xAxisData.begin(), minXAxisData - holeSize * resolution); |
|
34 | 36 | valuesData.insert(valuesData.begin(), fillValues.begin(), fillValues.end()); |
|
35 | 37 | } |
|
36 | 38 | } |
|
37 | 39 | |
|
38 | 40 | // Same for the end of the data |
|
39 | 41 | auto maxXAxisData = xAxisData.back(); |
|
40 | 42 | if (!std::isnan(maxBound) && maxBound > maxXAxisData) { |
|
41 | 43 | auto holeSize = static_cast<int>((maxBound - maxXAxisData) / resolution); |
|
42 | 44 | if (holeSize > 0) { |
|
43 | 45 | xAxisData.insert(xAxisData.end(), maxXAxisData + holeSize * resolution); |
|
44 | 46 | valuesData.insert(valuesData.end(), fillValues.begin(), fillValues.end()); |
|
45 | 47 | } |
|
46 | 48 | } |
|
47 | 49 | |
|
48 | 50 | // Generates other data holes |
|
49 | 51 | auto xAxisIt = xAxisData.begin(); |
|
50 | 52 | while (xAxisIt != xAxisData.end()) { |
|
51 | 53 | // Stops at first value which has a gap greater than resolution with the value next to it |
|
52 | 54 | xAxisIt = std::adjacent_find( |
|
53 | 55 | xAxisIt, xAxisData.end(), |
|
54 | 56 | [resolution](const auto &a, const auto &b) { return (b - a) > resolution; }); |
|
55 | 57 | |
|
56 | 58 | if (xAxisIt != xAxisData.end()) { |
|
57 | 59 | auto nextXAxisIt = xAxisIt + 1; |
|
58 | 60 | |
|
59 | 61 | // Gets the values that has a gap greater than resolution between them |
|
60 | 62 | auto lowValue = *xAxisIt; |
|
61 | 63 | auto highValue = *nextXAxisIt; |
|
62 | 64 | |
|
63 | 65 | // Completes holes between the two values by creating new values (according to the |
|
64 | 66 | // resolution) |
|
65 | 67 | for (auto i = lowValue + resolution; i < highValue; i += resolution) { |
|
66 | 68 | // Gets the iterator of values data from which to insert fill values |
|
67 | 69 | auto nextValuesIt = valuesData.begin() |
|
68 | 70 | + std::distance(xAxisData.begin(), nextXAxisIt) * nbComponents; |
|
69 | 71 | |
|
70 | 72 | // New value is inserted before nextXAxisIt |
|
71 | 73 | nextXAxisIt = xAxisData.insert(nextXAxisIt, i) + 1; |
|
72 | 74 | |
|
73 | 75 | // New values are inserted before nextValuesIt |
|
74 | 76 | valuesData.insert(nextValuesIt, fillValues.begin(), fillValues.end()); |
|
75 | 77 | } |
|
76 | 78 | |
|
77 | 79 | // Moves to the next value to continue loop on the x-axis data |
|
78 | 80 | xAxisIt = nextXAxisIt; |
|
79 | 81 | } |
|
80 | 82 | } |
|
81 | 83 | } |
|
84 | DataSeriesUtils::Mesh DataSeriesUtils::regularMesh(DataSeriesIterator begin, DataSeriesIterator end, | |
|
85 | Resolution xResolution, Resolution yResolution) | |
|
86 | { | |
|
87 | // Checks preconditions | |
|
88 | if (xResolution.m_Val == 0. || std::isnan(xResolution.m_Val) || yResolution.m_Val == 0. | |
|
89 | || std::isnan(yResolution.m_Val)) { | |
|
90 | qCWarning(LOG_DataSeriesUtils()) << "Can't generate mesh with a null resolution"; | |
|
91 | return Mesh{}; | |
|
92 | } | |
|
93 | ||
|
94 | if (xResolution.m_Logarithmic) { | |
|
95 | qCWarning(LOG_DataSeriesUtils()) | |
|
96 | << "Can't generate mesh with a logarithmic x-axis resolution"; | |
|
97 | return Mesh{}; | |
|
98 | } | |
|
99 | ||
|
100 | if (std::distance(begin, end) == 0) { | |
|
101 | qCWarning(LOG_DataSeriesUtils()) << "Can't generate mesh for empty data"; | |
|
102 | return Mesh{}; | |
|
103 | } | |
|
104 | ||
|
105 | auto yData = begin->y(); | |
|
106 | if (yData.empty()) { | |
|
107 | qCWarning(LOG_DataSeriesUtils()) << "Can't generate mesh for data with no y-axis"; | |
|
108 | return Mesh{}; | |
|
109 | } | |
|
110 | ||
|
111 | // Converts y-axis and its resolution to logarithmic values | |
|
112 | if (yResolution.m_Logarithmic) { | |
|
113 | std::for_each(yData.begin(), yData.end(), [](auto &val) { val = std::log10(val); }); | |
|
114 | } | |
|
115 | ||
|
116 | } |
General Comments 0
You need to be logged in to leave comments.
Login now