添加工具类

jeecg-module-data-analyze/src/main/java/org/jeecg/util/DistributionAnalysisToolkit.java

@@ -1,11 +1,12 @@
 package org.jeecg.util;
 
 import lombok.Data;
+import lombok.extern.slf4j.Slf4j;
 import org.jeecg.entity.NuclideActConcIntvl;
 
 import java.util.*;
 import java.util.stream.Collectors;
-
+@Slf4j
 public class DistributionAnalysisToolkit {
     /**
      * 区间统计结果
@@ -16,6 +17,7 @@ public class DistributionAnalysisToolkit {
         private final int count;
         private final List<Double> values;
         private final Map<Integer, Integer> levelDistribution;
+
         public IntervalStat(String interval, List<NuclideActConcIntvl> nuclideData) {
             values = new ArrayList<>();
             levelDistribution = new TreeMap<>();
@@ -50,8 +52,13 @@ public class DistributionAnalysisToolkit {
             this.density = density;
         }
 
-        public double getX() { return x; }
+        public double getX() {
-        public double getDensity() { return density; }
+            return x;
+        }
+
+        public double getDensity() {
+            return density;
+        }
     }
 
     /**
@@ -66,15 +73,19 @@ public class DistributionAnalysisToolkit {
             this.cumulativeProb = cumulativeProb;
         }
 
-        public double getValue() { return value; }
+        public double getValue() {
-        public double getCumulativeProb() { return cumulativeProb; }
+            return value;
         }
+
+        public double getCumulativeProb() {
+            return cumulativeProb;
+        }
+    }
+
     /**
      * 数据区间统计
      *
      * @param nuclideData 原始数据
-     * @param start 起始值
-     * @param step 区间宽度
      * @return 区间统计结果列表
      */
     //region
@@ -112,36 +123,69 @@ public class DistributionAnalysisToolkit {
 //    }
 //
     //endregion
-
+    public static List<IntervalStat> calculateIntervalStats(List<NuclideActConcIntvl> nuclideData) {
-    public static List<IntervalStat> calculateIntervalStats(List<NuclideActConcIntvl> nuclideData, double start, double step) {
         if (nuclideData == null || nuclideData.isEmpty()) {
             throw new IllegalArgumentException("数据不能为空");
         }
 
-        // 计算结束边界
+        int numBins = calculateOptimalBins(nuclideData);
-        double maxConc = nuclideData.stream()
+        //int numBins = 50;
-                .mapToDouble(NuclideActConcIntvl::getConc)
+        log.info("分箱个数",numBins);
-                .max()
+        //统计 min/max/count
-                .orElse(0.0);
+        DoubleSummaryStatistics statsData = nuclideData.stream()
-        double end = Math.ceil(maxConc / step) * step + step;
+                .map(nuclideActConcIntvl -> nuclideActConcIntvl.getConc())
+                .filter(Objects::nonNull)
+                .filter(d -> !d.isNaN() && !d.isInfinite())
+                .mapToDouble(Double::doubleValue)
+                .summaryStatistics();
+        if (statsData.getCount() == 0) {
+            return Collections.emptyList();
+        }
+        double min = statsData.getMin();
+        double max = statsData.getMax();
+        log.info("数据min:",min);
+        log.info("数据max:",max);
 
+        //2. 处理所有值相同或极小范围（关键！避免 binWidth = 0）
+        if (max - min < 1e-8) {
+            // 取值向下取整为起点，向上扩展一个单位
+            min = Math.floor(min);
+            max = min + Math.max(1.0, Math.abs(min) * 0.1); // 避免 min=-1000 时扩展太小
+        }
+
+        double binWidth = (max - min) / numBins;
+        if (binWidth <= 0) {
+            binWidth = 1.0; // 极端防御
+        }
+
+        List<String> bins = new ArrayList<>();
         // 初始化区间映射
         Map<String, List<NuclideActConcIntvl>> intervalMap = new TreeMap<>();
-        for (double lower = start; lower < end; lower += step) {
+        for (int i = 0; i < numBins; i++) {
-            double upper = lower + step;
+            double start1 = min + i * binWidth;
-            String key = String.format("[%.1f, %.1f)", lower, upper);
+            double end1 = min + (i + 1) * binWidth;
+
+            // 最后一个区间使用闭区间，语义更清晰
+            String key = i == numBins - 1
+                    ? String.format("[%.4f ~ %.4f]", start1, max)
+                    : String.format("[%.4f ~ %.4f)", start1, end1);
+            bins.add(key);
             intervalMap.put(key, new ArrayList<>());
         }
 
         // 分配数据到区间
         for (NuclideActConcIntvl nuclide : nuclideData) {
             double value = nuclide.getConc();
-            double lower = Math.floor(value / step) * step;
+            // 计算当前对象属于哪个区间索引
-            String key = String.format("[%.1f, %.1f)", lower, lower + step);
+            int index = (int) ((value - min) / binWidth);
-
+            if (index < 0) {
-            intervalMap.get(key).add(nuclide);
+                index = 0;
+            }
+            if (index >= numBins) {
+                index = numBins - 1;
+            }
+            intervalMap.get(bins.get(index)).add(nuclide);
         }
-
         // 转换为统计结果对象
         List<IntervalStat> stats = new ArrayList<>();
         for (Map.Entry<String, List<NuclideActConcIntvl>> entry : intervalMap.entrySet()) {
@@ -152,9 +196,6 @@ public class DistributionAnalysisToolkit {
     }
 
 
-
-
-
     /**
      * 计算95%累积线
      *
@@ -218,7 +259,6 @@ public class DistributionAnalysisToolkit {
     }
 
 
-
     /**
      * 核函数接口
      */
@@ -228,10 +268,12 @@ public class DistributionAnalysisToolkit {
     }
 
     // 常用核函数实现
-    public static final KernelFunction GAUSSIAN_KERNEL = u -> Math.exp(-0.5 * u * u) / Math.sqrt(2 * Math.PI);
+    public static final KernelFunction GAUSSIAN_KERNEL =
-    public static final KernelFunction EPANECHNIKOV_KERNEL = u -> (Math.abs(u) <= 1) ? 0.75 * (1 - u * u) : 0;
+            u -> Math.exp(-0.5 * u * u) / Math.sqrt(2 * Math.PI);
-    public static final KernelFunction TRIANGULAR_KERNEL = u -> (Math.abs(u) <= 1) ? 1 - Math.abs(u) : 0;
+    public static final KernelFunction EPANECHNIKOV_KERNEL =
-
+            u -> (Math.abs(u) <= 1) ? 0.75 * (1 - u * u) : 0;
+    public static final KernelFunction TRIANGULAR_KERNEL =
+            u -> (Math.abs(u) <= 1) ? 1 - Math.abs(u) : 0;
 
 
     /**
@@ -239,7 +281,9 @@ public class DistributionAnalysisToolkit {
      */
     public static double calculateBandwidthSilverman(List<Double> data) {
         int n = data.size();
-        if (n <= 1) return 1.0;
+        if (n <= 1) {
+            return 1.0;
+        }
 
         // 计算标准差
         double mean = data.stream().mapToDouble(Double::doubleValue).average().orElse(0.0);
@@ -310,13 +354,13 @@ public class DistributionAnalysisToolkit {
     }
 
 
-
     //获取浓度值集合
     public static List<Double> convertConcToDoubleList(List<NuclideActConcIntvl> nuclideList) {
         return nuclideList.stream()               // 创建流
                 .map(NuclideActConcIntvl::getConc) // 提取conc值
                 .collect(Collectors.toList());    // 收集为List<Double>
     }
+
     //累积和
     public static List<Double> cumulativeSum(List<Double> data) {
         List<Double> result = new ArrayList<>();
@@ -331,15 +375,32 @@ public class DistributionAnalysisToolkit {
     }
 
 
-
+    public static int calculateOptimalBins(List<NuclideActConcIntvl> data) {
-
+        if (Objects.isNull(data)) {
-
+            return 5;
-
+        }
-
+        List<Double> values = data.stream()
-
+                .map(NuclideActConcIntvl::getConc)
-
+                .filter(Objects::nonNull)
-
+                .filter(d -> !d.isNaN() && !d.isInfinite())
-
+                .sorted()
+                .collect(Collectors.toList());
+        int n = values.size();
+        if (n < 2) {
+            return 5;
+        }
+        double q1 = values.get(n / 4);
+        double q3 = values.get((3 * n) / 4);
+        double iqr = q3 - q1;
+        if (iqr <= 0) {
+            return (int) Math.max(3, Math.ceil(Math.sqrt(n)));
+        }
+        double binWidth = 2 * iqr * Math.pow(n, -1.0 / 3.0);
+        double dataRange = values.get(n - 1) - values.get(0);
+        int bins = (int) Math.ceil(dataRange / binWidth);
+        bins = Math.max(3, Math.min(40, bins));
+        return bins;
+    }
 
 
 }