Classifier (#1357)

* the best version so far * matche python mostly * before major change, kinda works * long one, all numpy stuff * classfier v1 * Seems to be working * stlye check * removed some not used stuff * add additional convert methods * some test fixes * rehaul * changes from code review comments * starting to work * trying new ncml * new tests * all seems to work * all seems to be working! * applied style * final style update * fix the shape of class_specs to 17 * fix pr comments * fix the flaky test * delete tests, modify constructor * remove unused ncml, add int and float test * update variable length to correct num * Suppress spurious/not relevant jfreechart CVEs (#1358) finalize reviews --------- Co-authored-by: Tara Drwenski <[email protected]> Co-authored-by: Hailey Johnson <[email protected]>
Unidata · Jul 1, 2024 · da644a8 · da644a8
1 parent c2058b2
commit da644a8
Show file tree

Hide file tree

Showing 5 changed files with 165 additions and 220 deletions.
diff --git a/cdm/core/src/main/java/ucar/nc2/filter/Classifier.java b/cdm/core/src/main/java/ucar/nc2/filter/Classifier.java
@@ -1,45 +1,63 @@
 package ucar.nc2.filter;
 
-import java.io.IOException;
+
 import ucar.ma2.Array;
-import ucar.ma2.DataType;
 import ucar.ma2.IndexIterator;
-import ucar.nc2.dataset.VariableDS;
+import ucar.nc2.Variable;
+import ucar.nc2.constants.CDM;
+import ucar.nc2.Attribute;
+import ucar.nc2.util.Misc;
+import java.util.ArrayList;
+import java.util.List;
+
 
 public class Classifier implements Enhancement {
-  private Classifier classifier = null;
-  private static Classifier emptyClassifier;
-  private int classifiedVal;
-  private int[] classifiedArray;
-
-  public static Classifier createFromVariable(VariableDS var) {
-    try {
-      Array arr = var.read();
-      // DataType type = var.getDataType();
-      return emptyClassifier();
-    } catch (IOException e) {
-      return emptyClassifier();
-    }
+  private String[] AttCat;
+  private List<int[]> rules = new ArrayList<>();
+
+  public Classifier() {
+    this.AttCat = new String[0];
+    this.rules = new ArrayList<>();
   }
 
-  public static Classifier emptyClassifier() {
-    emptyClassifier = new Classifier();
-    return emptyClassifier;
+  // Constructor with attributes
+  public Classifier(String[] AttCat) {
+    this.AttCat = AttCat;
+    this.rules = loadClassificationRules();
   }
 
-  /** Enough of a constructor */
-  public Classifier() {}
+  // Factory method to create a Classifier from a Variable
+  public static Classifier createFromVariable(Variable var) {
+    List<Attribute> attributes = var.attributes().getAttributes();
+
+    for (Attribute attribute : attributes) {
+      // check like this, or something else?
+      if (attribute == var.attributes().findAttribute(CDM.CLASSIFY)) {
+        String[] sets = attribute.getStringValue().split(";");
+        for (int i = 0; i < sets.length; i++) {
+          // trim and clean so it's ready
+          sets[i] = sets[i].trim();
+        }
+        return new Classifier(sets);
+      }
+    }
+
+    return new Classifier();
+
+  }
+
+
 
-  /** Classify double array */
-  public int[] classifyDoubleArray(Array arr) {
+  public int[] classifyWithAttributes(Array arr) {
     int[] classifiedArray = new int[(int) arr.getSize()];
-    int i = 0;
     IndexIterator iterArr = arr.getIndexIterator();
+    int i = 0;
     while (iterArr.hasNext()) {
       Number value = (Number) iterArr.getObjectNext();
       if (!Double.isNaN(value.doubleValue())) {
-
-        classifiedArray[i] = classifyArray(value.doubleValue());
+        classifiedArray[i] = classifyArrayAttribute(value.doubleValue());
+      } else {
+        classifiedArray[i] = Integer.MIN_VALUE;
       }
       i++;
     }
@@ -48,23 +66,54 @@ public int[] classifyDoubleArray(Array arr) {
 
 
 
-  /** for a single double */
-  public int classifyArray(double val) {
-    if (val >= 0) {
-      classifiedVal = 1;
-    } else {
-      classifiedVal = 0;
+  public int classifyArrayAttribute(double val) {
+    for (int[] rule : rules) {
+      if (val > rule[0] && val <= rule[1] + Misc.defaultMaxRelativeDiffFloat) {
+        return rule[2];
+      }
     }
+    // Return min possible int if no rule matches
+    return Integer.MIN_VALUE;
+  }
 
-    return classifiedVal;
+  // Method to load classification rules from the attributes
+  private List<int[]> loadClassificationRules() {
+    for (String rules : this.AttCat) {
+      int[] rule = stringToIntArray(rules);
+      this.rules.add(rule);
+    }
+    return rules;
   }
 
   @Override
   public double convert(double val) {
-    return emptyClassifier.classifyArray(val);
+    return classifyArrayAttribute(val);
   }
 
+  public static int[] stringToIntArray(String str) {
+    String[] stringArray = str.split(" "); // Split the string by spaces
+    int[] intArray = new int[stringArray.length]; // Create an array to hold the parsed integers
+
+    for (int i = 0; i < stringArray.length; i++) {
+
+      double value = Double.parseDouble(stringArray[i]); // Parse each string to a double
+
+      if (Double.isNaN(value)) {
+        // Check if the entry is NaN and assign Integer.MIN_VALUE or Integer.MAX_VALUE based on the index
+        if (i == 0) {
+          intArray[i] = Integer.MIN_VALUE;
+        } else if (i == 1) {
+          intArray[i] = Integer.MAX_VALUE;
+        } else {
+          intArray[i] = -99999; // Default value for other indices if needed
+        }
+      } else {
+        intArray[i] = (int) value; // Convert the value to int if it is not NaN
+      }
 
-}
+    }
 
+    return intArray;
+  }
 
+}
diff --git a/cdm/core/src/test/data/ncml/enhance/testAddToClassifier.ncml b/cdm/core/src/test/data/ncml/enhance/testAddToClassifier.ncml
@@ -0,0 +1,29 @@
+<?xml version="1.0" encoding="UTF-8"?>
+
+<!--
+  ~ Copyright (c) 1998-2020 John Caron and University Corporation for Atmospheric Research/Unidata
+  ~  See LICENSE for license information.
+  -->
+
+<netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2" enhance="all">
+
+    <variable name="class_specs" shape="17" type="double">
+        <attribute name="classify" value="NaN 0 0; 0 10 10; 10 100 100; 100 NaN 1000"/>
+        <values>-500000 NaN -10 0 1 2 3 11 25 29 NaN 100 150 121 102 199999 12211</values>
+    </variable>
+
+    <variable name="classify_ints" shape="5" type="int">
+        <attribute name="classify" value="NaN 0 0; 0 NaN 1"/>
+        <values>1 -2 0 4 -5</values>
+    </variable>
+
+    <variable name="classify_floats" shape="5" type="float">
+        <attribute name="classify" value="NaN 0 0; 0 NaN 1"/>
+        <values>1.0 -2.0 0.0 4.0 -5.0</values>
+    </variable>
+
+
+
+
+
+</netcdf>
diff --git a/cdm/core/src/test/data/ncml/enhance/testClassifier.ncml b/cdm/core/src/test/data/ncml/enhance/testClassifier.ncml
diff --git a/cdm/core/src/test/java/ucar/nc2/filter/TestClassifier.java b/cdm/core/src/test/java/ucar/nc2/filter/TestClassifier.java