mirror of
https://github.com/babalae/better-genshin-impact.git
synced 2026-04-26 22:39:47 +08:00
6d4f8b80e8
* 记录一次对hutaofisher的访谈，帮助开发者理解其算法
* 本地化HelloWorld
* .csproj取消windows版本号，此处导致了IDE在新建代码文件和自动生成代码时，默认命名空间丢失的问题。已知VisualStudio和ReSharper存在这个问题。
* 优化扩展方法写法，改为从localizer扩展；Converter优化写法，避免冲突；新增两种语言，待测试ocr效果
* Revert ".csproj取消windows版本号，此处导致了IDE在新建代码文件和自动生成代码时，默认命名空间丢失的问题。已知VisualStudio和ReSharper存在这个问题。"
This reverts commit 8bd7ee74c5.
* localizer改为由构造函数传入以支持单元测试；一个英语上钩的单元测试
* 传送任务支持英语游戏界面；本地化参数挪至OtherConfig类下，但界面位置暂不挪动，待定
* 调整resx位置风格，放在直接使用字符串的类下；一条龙合成树脂及领取每日奖励支持游戏内中英双语
* 删除无用碎片文件
* 删去两个不必要的Sdcb包引用
* Paddle服务类去掉分类模型；检测和识别新增支持繁中和法语，配有单元测试；因小语种识别效果不理想，使用正则匹配替换多处识别文本相等或包含判断；钓鱼、一条龙合成树脂及领取每日奖励支持游戏内繁中和法语；
* 检查今日奖励任务的多语言化；右侧联机的P图标检测区域宽度缩减，避免英语角色名被误识别成P
* AutoDomainTask的游戏多语言化，由于我的游戏账号无法测试，仅配一些测试用例
* 修复有3个Mizuki导致异常的bug，临时用拼音代替新角色英文名，并为该数据初始化方法添加单元测试
* 瓦雷莎删去别名“牛牛”，因荒泷一斗已占用此别名；别名加载和读取优化
* 加个锁避免单元测试中多线程初始化paddle崩溃
208 lines
7.5 KiB
C#
208 lines
7.5 KiB
C#
using OpenCvSharp;
|
||
using System;
|
||
using System.Linq;
|
||
|
||
namespace BetterGenshinImpact.GameTask.AutoFishing;
|
||
|
||
/// <summary>
|
||
/// copy from https://github.com/myHuTao-qwq/HutaoFisher/blob/master/src/rodnet.cpp
|
||
///
|
||
/// 以下是hutaofisher的访谈：
|
||
///
|
||
/// 修改我的qq昵称: 03-20 00:22:48
|
||
/// 有任何问题可以问我 只要我在工位而不是在实验室里跟XXXX斗智斗勇就回答你
|
||
///
|
||
/// 额 我的算法有几个感觉挺符合直觉的假设
|
||
///
|
||
/// 就是 你抛竿的时候鱼是否会咬钩取决于鱼饵落点到鱼的距离
|
||
///
|
||
/// 这里的距离是鱼饵落点的(x, y, z)坐标与鱼的(x, y, z)坐标之差在xy平面(水平面)上的投影
|
||
///
|
||
/// 现在问题转化成了如何估计这两个坐标
|
||
///
|
||
/// 这里我们先约定x轴就是水平面上与屏幕长边平行的线 y轴就是屏幕短边的对于 剩下一个z就是游戏里面往天上的方向
|
||
///
|
||
/// 然后经验告诉我鱼饵落点环的x几乎为0
|
||
///
|
||
/// 剩下我们可以用鱼饵落点环的长和宽算出y和z
|
||
///
|
||
/// 再近似鱼的bouding box的中心就是鱼计算咬钩的那个点 同时近似一个大类里面的所有鱼的z是相同的(这个参数可学习) 于是我们也可以计算出鱼的xyz
|
||
///
|
||
/// 最后一步就是把对应的投影距离算出来 然后线性回归一下得到太近 刚好 太远三个类
|
||
///
|
||
/// tmd 今天我意识到 XXXX可不就是XXXX
|
||
///
|
||
/// 哦 到这一步以后剩下的就很弱智了 远了挪近一点 近了挪远一点 调调参差不多得了
|
||
/// </summary>
|
||
public class RodNet
|
||
{
|
||
const double alpha = 1734.34 / 2.5;
|
||
|
||
static readonly double[] dz = [ 0.561117562965, 0.637026851288, 0.705579317577,
|
||
1.062734463845, 0.949307580751, 1.015620474332,
|
||
1.797904203405, 1.513476738412, 1.013873007495,
|
||
1.159949954831, 1.353650974146, 1.302893195071 ];
|
||
|
||
static readonly double[,] theta = {
|
||
{-0.262674397633, 0.317025388945, -0.457150765450, 0.174522158281,
|
||
-0.957110676932, -0.095339800558, -0.119519564026, -0.139914755291,
|
||
-0.580893838475, 0.702302245305, 0.271575851220, 0.708473199472,
|
||
0.699108382380},
|
||
{-1.062702043060, -0.280779165943, -0.289891597384, 0.220173840594,
|
||
0.493463877037, -0.326492366566, 1.215859141832, 1.607133159643,
|
||
1.619199133672, 0.356402262447, 0.365385941958, 0.411869019381,
|
||
0.224962055122},
|
||
{0.460481782256, 0.048180392806, 0.475529271293, -0.150186412126,
|
||
0.135512307120, 0.087365984352, -1.317661146364, -1.882438208662,
|
||
-1.502483859283, -0.580228373556, -1.005821958682, -1.184199131739,
|
||
-1.285988918494}
|
||
};
|
||
|
||
static readonly double[] B = [1.241950004386, 3.051113640564, -3.848898190087];
|
||
|
||
static readonly double[] offset = [ 0.4, 0.2, 0.4, 0, 0.3, 0.3,
|
||
0.3, 0.15, 0.5, 0.5, 0.5, 0.5 ];
|
||
|
||
static void F(double[] dst, double[] x, double[] y)
|
||
{
|
||
double y0 = x[0], z0 = x[1], t = x[2];
|
||
double tmp = (y0 + t * z0) * (y0 + t * z0) - 1;
|
||
dst[0] = Math.Sqrt((1 + t * t) / tmp) - y[0];
|
||
dst[1] = (1 + t * t) * z0 / tmp - y[1];
|
||
dst[2] = ((t * t - 1) * y0 * z0 + t * (y0 * y0 - z0 * z0 - 1)) / tmp - y[2];
|
||
}
|
||
|
||
static void DfInv(double[] dst, double[] x)
|
||
{
|
||
double y0 = x[0], z0 = x[1], t = x[2];
|
||
double tmp1 = (y0 + t * z0) * (y0 + t * z0) - 1;
|
||
double tmp2 = 1 + t * t;
|
||
dst[0] = (1 - y0 * y0 + z0 * z0) / y0 * Math.Sqrt(tmp1 / tmp2);
|
||
dst[1] = -z0 * (y0 * y0 + t * (t * (1 + z0 * z0) + 2 * y0 * z0)) / tmp2 / y0;
|
||
dst[2] = ((t * t - 1) * y0 * z0 + t * (y0 * y0 - z0 * z0 - 1)) / y0 / tmp2;
|
||
dst[3] = -2 * z0 * Math.Sqrt(tmp1 / tmp2);
|
||
dst[4] = tmp1 / tmp2;
|
||
dst[5] = 0;
|
||
dst[6] = -z0 / y0 * Math.Sqrt(tmp1 / tmp2);
|
||
dst[7] = (y0 + t * z0) * (y0 + t * z0) / y0;
|
||
dst[8] = 1 + t * z0 / y0;
|
||
}
|
||
|
||
static bool NewtonRaphson(Action<double[], double[], double[]> f, Action<double[], double[]> dfInv, double[] dst, double[] y,
|
||
double[] init, int n, int maxIter, double eps)
|
||
{
|
||
double[] fEst = new double[n];
|
||
double[] dfInvMat = new double[n * n];
|
||
double[] x = new double[n];
|
||
double err;
|
||
|
||
Array.Copy(init, x, n);
|
||
|
||
for (int iter = 0; iter < maxIter; iter++)
|
||
{
|
||
err = 0;
|
||
f(fEst, x, y);
|
||
for (int i = 0; i < n; i++)
|
||
{
|
||
err += Math.Abs(fEst[i]);
|
||
}
|
||
|
||
if (err < eps)
|
||
{
|
||
Array.Copy(x, dst, n);
|
||
// printf("Newton-Raphson solver converge after %d steps, err: %lf !\n",
|
||
// iter, err);
|
||
return true;
|
||
}
|
||
|
||
dfInv(dfInvMat, x);
|
||
|
||
for (int i = 0; i < n; i++)
|
||
{
|
||
for (int j = 0; j < n; j++)
|
||
{
|
||
x[i] -= dfInvMat[n * i + j] * fEst[j];
|
||
}
|
||
}
|
||
}
|
||
return false;
|
||
}
|
||
|
||
static void Softmax(double[] dst, double[] x, int n)
|
||
{
|
||
double sum = 0;
|
||
for (int i = 0; i < n; i++)
|
||
{
|
||
dst[i] = Math.Exp(x[i]);
|
||
sum += dst[i];
|
||
}
|
||
for (int i = 0; i < n; i++)
|
||
{
|
||
dst[i] /= sum;
|
||
}
|
||
}
|
||
|
||
public static int GetRodState(RodInput input)
|
||
{
|
||
double a, b, v0, u, v;
|
||
|
||
a = (input.rod_x2 - input.rod_x1) / 2 / alpha;
|
||
b = (input.rod_y2 - input.rod_y1) / 2 / alpha;
|
||
|
||
if (a < b)
|
||
{
|
||
(b, a) = (a, b);
|
||
}
|
||
|
||
v0 = (288 - (input.rod_y1 + input.rod_y2) / 2) / alpha;
|
||
|
||
u = (input.fish_x1 + input.fish_x2 - input.rod_x1 - input.rod_x2) / 2 / alpha;
|
||
v = (288 - (input.fish_y1 + input.fish_y2) / 2) / alpha;
|
||
|
||
double[] y0z0t = new double[3];
|
||
double[] abv0 = [a, b, v0];
|
||
double[] init = [30, 15, 1];
|
||
|
||
bool solveSuccess = NewtonRaphson(F, DfInv, y0z0t, abv0, init, 3, 1000, 1e-6);
|
||
|
||
if (!solveSuccess)
|
||
{
|
||
return -1;
|
||
}
|
||
|
||
double y0 = y0z0t[0], z0 = y0z0t[1], t = y0z0t[2];
|
||
double x, y, dist;
|
||
x = u * (z0 + dz[input.fish_label]) * Math.Sqrt(1 + t * t) / (t - v);
|
||
y = (z0 + dz[input.fish_label]) * (1 + t * v) / (t - v);
|
||
dist = Math.Sqrt(x * x + (y - y0) * (y - y0));
|
||
|
||
double[] logits = new double[3];
|
||
for (int i = 0; i < 3; i++)
|
||
{
|
||
logits[i] = theta[i, 0] * dist + theta[i, 1 + input.fish_label] + B[i];
|
||
}
|
||
|
||
double[] pred = new double[3];
|
||
Softmax(pred, logits, 3);
|
||
pred[0] -= offset[input.fish_label];
|
||
|
||
return Array.IndexOf(pred, pred.Max());
|
||
}
|
||
|
||
public static int GetRodState(Rect rod, Rect fish, int fishTypeIndex)
|
||
{
|
||
RodInput input = new()
|
||
{
|
||
rod_x1 = rod.Left,
|
||
rod_x2 = rod.Right,
|
||
rod_y1 = rod.Top,
|
||
rod_y2 = rod.Bottom,
|
||
fish_x1 = fish.Left,
|
||
fish_x2 = fish.Right,
|
||
fish_y1 = fish.Top,
|
||
fish_y2 = fish.Bottom,
|
||
fish_label = fishTypeIndex
|
||
};
|
||
return GetRodState(input);
|
||
}
|
||
} |