1 /***********************************************************************
2  * Copyright (c) 2013-2024 Commonwealth Computer Research, Inc.
3  * All rights reserved. This program and the accompanying materials
4  * are made available under the terms of the Apache License, Version 2.0
5  * which accompanies this distribution and is available at
6  * http://www.opensource.org/licenses/apache2.0.php.
7  ***********************************************************************/
8 
9 package org.locationtech.geomesa.tools.stats
10 
11 import com.beust.jcommander.ParameterException
12 import org.locationtech.geomesa.index.geotools.GeoMesaDataStore
13 import org.locationtech.geomesa.index.metadata.GeoMesaMetadata
14 import org.locationtech.geomesa.tools.stats.StatsAnalyzeCommand.StatsAnalyzeParams
15 import org.locationtech.geomesa.tools.{CatalogParam, Command, DataStoreCommand, TypeNameParam}
16 import org.locationtech.geomesa.utils.stats._
17 import org.locationtech.jts.geom.Geometry
18 
19 trait StatsAnalyzeCommand[DS <: GeoMesaDataStore[DS]] extends DataStoreCommand[DS] {
20 
21   import org.locationtech.geomesa.utils.geotools.RichSimpleFeatureType.RichSimpleFeatureType
22 
23   import scala.collection.JavaConverters._
24 
25   override val name = "stats-analyze"
26   override def params: StatsAnalyzeParams
27 
28   override def execute(): Unit = withDataStore(analyze)
29 
30   private def analyze(ds: DS): Unit = {
31     val sft = ds.getSchema(params.featureName)
32     if (sft == null) {
33       throw new ParameterException(s"Schema '${params.featureName}' does not exist")
34     }
35 
36     def noStats = ds.metadata.read(sft.getTypeName, GeoMesaMetadata.StatsGenerationKey, cache = false).isEmpty
37 
38     Command.user.info(s"Running stat analysis for feature type ${sft.getTypeName}...")
39     val stats = if (noStats) {
40       // this is the first time anyone has accessed this data store since stats were implemented
41       // stats should be kicked off and run in the background, we just have to wait for it to finish
42       // note: if something goes wrong this will just hang... but this is kind of a fringe case
43       while (noStats) {
44         Thread.sleep(1000)
45       }
46       // sleep an additional bit so that the stat table gets configured
47       Thread.sleep(1000)
48       val queries = sft.getAttributeDescriptors.asScala.map(d => Stat.MinMax(d.getLocalName))
49       ds.stats.getStat[CountStat](sft, Stat.Count()) ++ ds.stats.getSeqStat[MinMax[Any]](sft, queries.toSeq)
50     } else {
51       ds.stats.writer.analyze(sft)
52     }
53 
54     val strings = stats.collect {
55       case s: CountStat   => s"Total features: ${s.count}"
56       case s: MinMax[Any] =>
57         val bounds = if (s.isEmpty) {
58           "[ no matching data ]"
59         } else if (sft.getGeomField == s.property) {
60           val e = s.min.asInstanceOf[Geometry].getEnvelopeInternal
61           e.expandToInclude(s.max.asInstanceOf[Geometry].getEnvelopeInternal)
62           s"[ ${e.getMinX}, ${e.getMinY}, ${e.getMaxX}, ${e.getMaxY} ] cardinality: ${s.cardinality}"
63         } else {
64           val stringify = Stat.stringifier(sft.getDescriptor(s.property).getType.getBinding)
65           s"[ ${stringify(s.min)} to ${stringify(s.max)} ] cardinality: ${s.cardinality}"
66         }
67         s"Bounds for ${s.property}: $bounds"
68     }
69 
70     Command.user.info("Stats analyzed:")
71     Command.output.info(strings.mkString("  ", "\n  ", ""))
72     Command.user.info("Use 'stats-histogram', 'stats-top-k' or 'stats-count' commands for more details")
73   }
74 }
75 
76 object StatsAnalyzeCommand {
77   // @Parameters(commandDescription = "Analyze statistics on a GeoMesa feature type")
78   trait StatsAnalyzeParams extends CatalogParam with TypeNameParam
79 }