模仿学习、离线轨迹与 DAgger

import java.util.*;
public class Main {
    static void near(double expected,double actual) {
        if (!Double.isFinite(actual) || Math.abs(expected-actual)>1e-9*Math.max(1,Math.abs(expected)))
            throw new AssertionError(expected+" != "+actual);
    }
    static void check(boolean ok) { if(!ok) throw new AssertionError(); }
    static void bad(Runnable r) {
        try { r.run(); } catch(IllegalArgumentException e) { return; }
        throw new AssertionError("invalid input accepted");
    }
record Example(int state,int action,long weight) {}
    static Map<Integer,Integer> fit(List<Example> data,int actions) {
        if(data==null||actions<1||actions>10000) throw new IllegalArgumentException("shape");
        Map<Integer,long[]> counts=new HashMap<>();
        for(Example e:data) {
            if(e==null||e.state()<0||e.action()<0||e.action()>=actions||e.weight()<=0)
                throw new IllegalArgumentException("example");
            long[] row=counts.computeIfAbsent(e.state(),k->new long[actions]);
            try { row[e.action()]=Math.addExact(row[e.action()],e.weight()); }
            catch(ArithmeticException ex) { throw new IllegalArgumentException("weight overflow",ex); }
            // 保留全部轮次的标签权重，而非用最后一次动作覆盖历史。
        }
        Map<Integer,Integer> result=new HashMap<>();
        for(var entry:counts.entrySet()) {
            int best=0;long[] row=entry.getValue();
            for(int a=1;a<actions;a++) if(row[a]>row[best]) best=a; // 同票时固定选较小编号。
            result.put(entry.getKey(),best); // 每个状态独立最小化加权零一错误。
        }
        return Map.copyOf(result);
    }
    public static void main(String[] args) {
check(fit(List.of(new Example(0,1,2),new Example(0,0,1)),2).get(0)==1);
        check(fit(List.of(new Example(3,0,1),new Example(3,1,1)),2).get(3)==0);
        check(fit(List.of(new Example(0,1,1),new Example(3,0,1)),2).size()==2);
        check(fit(List.of(),2).isEmpty());
        bad(()->fit(List.of(new Example(0,2,1)),2));
        bad(()->fit(List.of(new Example(0,0,0)),2));
        bad(()->fit(List.of(new Example(0,0,Long.MAX_VALUE),new Example(0,0,1)),2));
        check(!fit(List.of(new Example(0,1,1)),2).containsKey(4));
        System.out.println("PASS 8 cases");
    }
}