QGIS API Documentation 3.40.0-Bratislava (b56115d8743)
Loading...
Searching...
No Matches
qgsalgorithmkmeansclustering.cpp
Go to the documentation of this file.
1/***************************************************************************
2 qgsalgorithmkmeansclustering.cpp
3 ---------------------
4 begin : June 2018
5 copyright : (C) 2018 by Nyall Dawson
6 email : nyall dot dawson at gmail dot com
7 ***************************************************************************/
8
9/***************************************************************************
10 * *
11 * This program is free software; you can redistribute it and/or modify *
12 * it under the terms of the GNU General Public License as published by *
13 * the Free Software Foundation; either version 2 of the License, or *
14 * (at your option) any later version. *
15 * *
16 ***************************************************************************/
17
19#include <unordered_map>
20
22
23const int KMEANS_MAX_ITERATIONS = 1000;
24
25QString QgsKMeansClusteringAlgorithm::name() const
26{
27 return QStringLiteral( "kmeansclustering" );
28}
29
30QString QgsKMeansClusteringAlgorithm::displayName() const
31{
32 return QObject::tr( "K-means clustering" );
33}
34
35QStringList QgsKMeansClusteringAlgorithm::tags() const
36{
37 return QObject::tr( "clustering,clusters,kmeans,points" ).split( ',' );
38}
39
40QString QgsKMeansClusteringAlgorithm::group() const
41{
42 return QObject::tr( "Vector analysis" );
43}
44
45QString QgsKMeansClusteringAlgorithm::groupId() const
46{
47 return QStringLiteral( "vectoranalysis" );
48}
49
50void QgsKMeansClusteringAlgorithm::initAlgorithm( const QVariantMap & )
51{
52 addParameter( new QgsProcessingParameterFeatureSource( QStringLiteral( "INPUT" ),
53 QObject::tr( "Input layer" ), QList< int >() << static_cast< int >( Qgis::ProcessingSourceType::VectorAnyGeometry ) ) );
54 addParameter( new QgsProcessingParameterNumber( QStringLiteral( "CLUSTERS" ), QObject::tr( "Number of clusters" ),
56
57 auto fieldNameParam = std::make_unique<QgsProcessingParameterString>( QStringLiteral( "FIELD_NAME" ),
58 QObject::tr( "Cluster field name" ), QStringLiteral( "CLUSTER_ID" ) );
59 fieldNameParam->setFlags( fieldNameParam->flags() | Qgis::ProcessingParameterFlag::Advanced );
60 addParameter( fieldNameParam.release() );
61 auto sizeFieldNameParam = std::make_unique<QgsProcessingParameterString>( QStringLiteral( "SIZE_FIELD_NAME" ),
62 QObject::tr( "Cluster size field name" ), QStringLiteral( "CLUSTER_SIZE" ) );
63 sizeFieldNameParam->setFlags( sizeFieldNameParam->flags() | Qgis::ProcessingParameterFlag::Advanced );
64 addParameter( sizeFieldNameParam.release() );
65
66 addParameter( new QgsProcessingParameterFeatureSink( QStringLiteral( "OUTPUT" ), QObject::tr( "Clusters" ), Qgis::ProcessingSourceType::VectorAnyGeometry ) );
67}
68
69QString QgsKMeansClusteringAlgorithm::shortHelpString() const
70{
71 return QObject::tr( "Calculates the 2D distance based k-means cluster number for each input feature.\n\n"
72 "If input geometries are lines or polygons, the clustering is based on the centroid of the feature." );
73}
74
75QgsKMeansClusteringAlgorithm *QgsKMeansClusteringAlgorithm::createInstance() const
76{
77 return new QgsKMeansClusteringAlgorithm();
78}
79
80QVariantMap QgsKMeansClusteringAlgorithm::processAlgorithm( const QVariantMap &parameters, QgsProcessingContext &context, QgsProcessingFeedback *feedback )
81{
82 std::unique_ptr< QgsProcessingFeatureSource > source( parameterAsSource( parameters, QStringLiteral( "INPUT" ), context ) );
83 if ( !source )
84 throw QgsProcessingException( invalidSourceError( parameters, QStringLiteral( "INPUT" ) ) );
85
86 int k = parameterAsInt( parameters, QStringLiteral( "CLUSTERS" ), context );
87
88 QgsFields outputFields = source->fields();
89 QgsFields newFields;
90 const QString clusterFieldName = parameterAsString( parameters, QStringLiteral( "FIELD_NAME" ), context );
91 newFields.append( QgsField( clusterFieldName, QMetaType::Type::Int ) );
92 const QString clusterSizeFieldName = parameterAsString( parameters, QStringLiteral( "SIZE_FIELD_NAME" ), context );
93 newFields.append( QgsField( clusterSizeFieldName, QMetaType::Type::Int ) );
94 outputFields = QgsProcessingUtils::combineFields( outputFields, newFields );
95
96 QString dest;
97 std::unique_ptr< QgsFeatureSink > sink( parameterAsSink( parameters, QStringLiteral( "OUTPUT" ), context, dest, outputFields, source->wkbType(), source->sourceCrs() ) );
98 if ( !sink )
99 throw QgsProcessingException( invalidSinkError( parameters, QStringLiteral( "OUTPUT" ) ) );
100
101 // build list of point inputs - if it's already a point, use that. If not, take the centroid.
102 feedback->pushInfo( QObject::tr( "Collecting input points" ) );
103 const double step = source->featureCount() > 0 ? 50.0 / source->featureCount() : 1;
104 int i = 0;
105 int n = 0;
106 int featureWithGeometryCount = 0;
107 QgsFeature feat;
108
109 std::vector< Feature > clusterFeatures;
110 QgsFeatureIterator features = source->getFeatures( QgsFeatureRequest().setNoAttributes() );
111 QHash< QgsFeatureId, int > idToObj;
112 while ( features.nextFeature( feat ) )
113 {
114 i++;
115 if ( feedback->isCanceled() )
116 {
117 break;
118 }
119
120 feedback->setProgress( i * step );
121 if ( !feat.hasGeometry() )
122 continue;
123 featureWithGeometryCount++;
124
125 QgsPointXY point;
127 point = QgsPointXY( *qgsgeometry_cast< const QgsPoint * >( feat.geometry().constGet() ) );
128 else
129 {
130 const QgsGeometry centroid = feat.geometry().centroid();
131 if ( centroid.isNull() )
132 continue; // centroid failed, e.g. empty linestring
133
134 point = QgsPointXY( *qgsgeometry_cast< const QgsPoint * >( centroid.constGet() ) );
135 }
136
137 n++;
138
139 idToObj[ feat.id() ] = clusterFeatures.size();
140 clusterFeatures.emplace_back( Feature( point ) );
141 }
142
143 if ( n < k )
144 {
145 feedback->reportError( QObject::tr( "Number of geometries is less than the number of clusters requested, not all clusters will get data" ) );
146 k = n;
147 }
148
149 if ( k > 1 )
150 {
151 feedback->pushInfo( QObject::tr( "Calculating clusters" ) );
152
153 // cluster centers
154 std::vector< QgsPointXY > centers( k );
155
156 initClusters( clusterFeatures, centers, k, feedback );
157 calculateKMeans( clusterFeatures, centers, k, feedback );
158 }
159
160 // cluster size
161 std::unordered_map< int, int> clusterSize;
162 for ( const int obj : idToObj )
163 clusterSize[ clusterFeatures[ obj ].cluster ]++;
164
165 features = source->getFeatures();
166 i = 0;
167 while ( features.nextFeature( feat ) )
168 {
169 i++;
170 if ( feedback->isCanceled() )
171 {
172 break;
173 }
174
175 feedback->setProgress( 50 + i * step );
176 QgsAttributes attr = feat.attributes();
177 const auto obj = idToObj.find( feat.id() );
178 if ( !feat.hasGeometry() || obj == idToObj.end() )
179 {
180 attr << QVariant() << QVariant();
181 }
182 else if ( k <= 1 )
183 {
184 attr << 0 << featureWithGeometryCount;
185 }
186 else
187 {
188 const int cluster = clusterFeatures[ *obj ].cluster;
189 attr << cluster << clusterSize[ cluster ];
190 }
191 feat.setAttributes( attr );
192 if ( !sink->addFeature( feat, QgsFeatureSink::FastInsert ) )
193 throw QgsProcessingException( writeFeatureError( sink.get(), parameters, QStringLiteral( "OUTPUT" ) ) );
194 }
195
196 QVariantMap outputs;
197 outputs.insert( QStringLiteral( "OUTPUT" ), dest );
198 return outputs;
199}
200
201// ported from https://github.com/postgis/postgis/blob/svn-trunk/liblwgeom/lwkmeans.c
202
203void QgsKMeansClusteringAlgorithm::initClusters( std::vector<Feature> &points, std::vector<QgsPointXY> &centers, const int k, QgsProcessingFeedback *feedback )
204{
205 const std::size_t n = points.size();
206 if ( n == 0 )
207 return;
208
209 if ( n == 1 )
210 {
211 for ( int i = 0; i < k; i++ )
212 centers[ i ] = points[ 0 ].point;
213 return;
214 }
215
216 long duplicateCount = 1;
217 // initially scan for two most distance points from each other, p1 and p2
218 std::size_t p1 = 0;
219 std::size_t p2 = 0;
220 double distanceP1 = 0;
221 double distanceP2 = 0;
222 double maxDistance = -1;
223 for ( std::size_t i = 1; i < n; i++ )
224 {
225 distanceP1 = points[i].point.sqrDist( points[p1].point );
226 distanceP2 = points[i].point.sqrDist( points[p2].point );
227
228 // if this point is further then existing candidates, replace our choice
229 if ( ( distanceP1 > maxDistance ) || ( distanceP2 > maxDistance ) )
230 {
231 maxDistance = std::max( distanceP1, distanceP2 );
232 if ( distanceP1 > distanceP2 )
233 p2 = i;
234 else
235 p1 = i;
236 }
237
238 // also record count of duplicate points
239 if ( qgsDoubleNear( distanceP1, 0 ) || qgsDoubleNear( distanceP2, 0 ) )
240 duplicateCount++;
241 }
242
243 if ( feedback && duplicateCount > 1 )
244 {
245 feedback->pushInfo( QObject::tr( "There are at least %n duplicate input(s), the number of output clusters may be less than was requested", nullptr, duplicateCount ) );
246 }
247
248 // By now two points should be found and be not the same
249 // Q_ASSERT( p1 != p2 && maxDistance >= 0 );
250
251 // Accept these two points as our initial centers
252 centers[0] = points[p1].point;
253 centers[1] = points[p2].point;
254
255 if ( k > 2 )
256 {
257 // array of minimum distance to a point from accepted cluster centers
258 std::vector< double > distances( n );
259
260 // initialize array with distance to first object
261 for ( std::size_t j = 0; j < n; j++ )
262 {
263 distances[j] = points[j].point.sqrDist( centers[0] );
264 }
265 distances[p1] = -1;
266 distances[p2] = -1;
267
268 // loop i on clusters, skip 0 and 1 as found already
269 for ( int i = 2; i < k; i++ )
270 {
271 std::size_t candidateCenter = 0;
272 double maxDistance = std::numeric_limits<double>::lowest();
273
274 // loop j on points
275 for ( std::size_t j = 0; j < n; j++ )
276 {
277 // accepted clusters are already marked with distance = -1
278 if ( distances[j] < 0 )
279 continue;
280
281 // update minimal distance with previously accepted cluster
282 distances[j] = std::min( points[j].point.sqrDist( centers[i - 1] ), distances[j] );
283
284 // greedily take a point that's farthest from any of accepted clusters
285 if ( distances[j] > maxDistance )
286 {
287 candidateCenter = j;
288 maxDistance = distances[j];
289 }
290 }
291
292 // checked earlier by counting entries on input, just in case
293 Q_ASSERT( maxDistance >= 0 );
294
295 // accept candidate to centers
296 distances[candidateCenter] = -1;
297 // copy the point coordinates into the initial centers array
298 centers[i] = points[candidateCenter].point;
299 }
300 }
301}
302
303// ported from https://github.com/postgis/postgis/blob/svn-trunk/liblwgeom/lwkmeans.c
304
305void QgsKMeansClusteringAlgorithm::calculateKMeans( std::vector<QgsKMeansClusteringAlgorithm::Feature> &objs, std::vector<QgsPointXY> &centers, int k, QgsProcessingFeedback *feedback )
306{
307 int converged = false;
308 bool changed = false;
309
310 // avoid reallocating weights array for every iteration
311 std::vector< uint > weights( k );
312
313 uint i = 0;
314 for ( i = 0; i < KMEANS_MAX_ITERATIONS && !converged; i++ )
315 {
316 if ( feedback && feedback->isCanceled() )
317 break;
318
319 findNearest( objs, centers, k, changed );
320 updateMeans( objs, centers, weights, k );
321 converged = !changed;
322 }
323
324 if ( !converged && feedback )
325 feedback->reportError( QObject::tr( "Clustering did not converge after %n iteration(s)", nullptr, i ) );
326 else if ( feedback )
327 feedback->pushInfo( QObject::tr( "Clustering converged after %n iteration(s)", nullptr, i ) );
328}
329
330// ported from https://github.com/postgis/postgis/blob/svn-trunk/liblwgeom/lwkmeans.c
331
332void QgsKMeansClusteringAlgorithm::findNearest( std::vector<QgsKMeansClusteringAlgorithm::Feature> &points, const std::vector<QgsPointXY> &centers, const int k, bool &changed )
333{
334 changed = false;
335 const std::size_t n = points.size();
336 for ( std::size_t i = 0; i < n; i++ )
337 {
338 Feature &point = points[i];
339
340 // Initialize with distance to first cluster
341 double currentDistance = point.point.sqrDist( centers[0] );
342 int currentCluster = 0;
343
344 // Check all other cluster centers and find the nearest
345 for ( int cluster = 1; cluster < k; cluster++ )
346 {
347 const double distance = point.point.sqrDist( centers[cluster] );
348 if ( distance < currentDistance )
349 {
350 currentDistance = distance;
351 currentCluster = cluster;
352 }
353 }
354
355 // Store the nearest cluster this object is in
356 if ( point.cluster != currentCluster )
357 {
358 changed = true;
359 point.cluster = currentCluster;
360 }
361 }
362}
363
364// ported from https://github.com/postgis/postgis/blob/svn-trunk/liblwgeom/lwkmeans.c
365
366void QgsKMeansClusteringAlgorithm::updateMeans( const std::vector<Feature> &points, std::vector<QgsPointXY> &centers, std::vector<uint> &weights, const int k )
367{
368 const uint n = points.size();
369 std::fill( weights.begin(), weights.end(), 0 );
370 for ( int i = 0; i < k; i++ )
371 {
372 centers[i].setX( 0.0 );
373 centers[i].setY( 0.0 );
374 }
375 for ( uint i = 0; i < n; i++ )
376 {
377 const int cluster = points[i].cluster;
378 centers[cluster] += QgsVector( points[i].point.x(),
379 points[i].point.y() );
380 weights[cluster] += 1;
381 }
382 for ( int i = 0; i < k; i++ )
383 {
384 centers[i] /= weights[i];
385 }
386}
387
388
390
391
@ VectorAnyGeometry
Any vector layer with geometry.
@ Advanced
Parameter is an advanced parameter which should be hidden from users by default.
A vector of attributes.
Wrapper for iterator of features from vector data provider or vector layer.
bool nextFeature(QgsFeature &f)
Fetch next feature and stores in f, returns true on success.
This class wraps a request for features to a vector layer (or directly its vector data provider).
@ FastInsert
Use faster inserts, at the cost of updating the passed features to reflect changes made at the provid...
The feature class encapsulates a single feature including its unique ID, geometry and a list of field...
Definition qgsfeature.h:58
QgsAttributes attributes
Definition qgsfeature.h:67
QgsFeatureId id
Definition qgsfeature.h:66
void setAttributes(const QgsAttributes &attrs)
Sets the feature's attributes.
QgsGeometry geometry
Definition qgsfeature.h:69
bool hasGeometry() const
Returns true if the feature has an associated geometry.
bool isCanceled() const
Tells whether the operation has been canceled already.
Definition qgsfeedback.h:53
void setProgress(double progress)
Sets the current progress for the feedback object.
Definition qgsfeedback.h:61
Encapsulate a field in an attribute table or data source.
Definition qgsfield.h:53
Container of fields for a vector layer.
Definition qgsfields.h:46
bool append(const QgsField &field, Qgis::FieldOrigin origin=Qgis::FieldOrigin::Provider, int originIndex=-1)
Appends a field.
Definition qgsfields.cpp:69
A geometry is the spatial representation of a feature.
const QgsAbstractGeometry * constGet() const
Returns a non-modifiable (const) reference to the underlying abstract geometry primitive.
QgsGeometry centroid() const
Returns the center of mass of a geometry.
Qgis::WkbType wkbType() const
Returns type of the geometry as a WKB type (point / linestring / polygon etc.)
A class to represent a 2D point.
Definition qgspointxy.h:60
Contains information about the context in which a processing algorithm is executed.
Custom exception class for processing related exceptions.
Base class for providing feedback from a processing algorithm.
virtual void pushInfo(const QString &info)
Pushes a general informational message from the algorithm.
virtual void reportError(const QString &error, bool fatalError=false)
Reports that the algorithm encountered an error while executing.
A feature sink output for processing algorithms.
An input feature source (such as vector layers) parameter for processing algorithms.
A numeric parameter for processing algorithms.
static QgsFields combineFields(const QgsFields &fieldsA, const QgsFields &fieldsB, const QString &fieldsBPrefix=QString())
Combines two field lists, avoiding duplicate field names (in a case-insensitive manner).
A class to represent a vector.
Definition qgsvector.h:30
static Qgis::WkbType flatType(Qgis::WkbType type)
Returns the flat type for a WKB type.
bool qgsDoubleNear(double a, double b, double epsilon=4 *std::numeric_limits< double >::epsilon())
Compare two doubles (but allow some difference)
Definition qgis.h:5917