Initial commit

2025-11-30 08:35:28 +08:00
commit 3073463598
13 changed files with 735 additions and 0 deletions
--- a/skills/hydro-forecast/SKILL.md
+++ b/skills/hydro-forecast/SKILL.md
@@ -0,0 +1,125 @@
+---
+name: hydro-forecast
+description: 
+---
+
+# 1 运行环境说明
+
+- 在Julia中运行
+
+- 在julia中首先加载包，`using HydroTools`
+
+- 若没有包加载出错，则安装之，`using Pkg; Pkg.add("HydroTools")`
+
+
+## 说明
+
+先不要立即执行该skills，提醒用户输入的数据的格式。用户需要整理好的数据路径即可。
+
+```
+
+```
+
+`model`模型选择
+
+ MarrMot
+ XAJ
+ TCN
+ LSTM
+ KAN
+
+如果复杂、参数比较多的模型：要求用户输入模型参数`json`文件。
+按照如下示例
+
+```json
+{
+    "clumping_index": 0.62,
+    "LAI_max_o": 4.5,
+    "LAI_max_u": 2.4,
+    "z00": 1.33,
+    "mass_overstory": 35,
+    "mass_understory": 10,
+    "root_depth": 0.6,
+    "α_canopy_vis": 0.035,
+    "α_canopy_nir": 0.23,
+    "r_root_decay": 0.95,
+    "minimum_stomatal_resistance": 150,
+    "z_canopy_o": 20,
+    "z_canopy_u": 3,
+    "g1_w": 8,
+    "VCmax25": 62.5,
+    "leaf_resp_co": 0.0015,
+    "stem_resp_co": 0.0020,
+    "root_resp_co": 0.0020,
+    "fine_root_resp_co": 0.003,
+    "N_leaf": 4.45,
+    "slope_Vc": 0.33152
+}
+```
+
+## 1.1 任务说明
+
+### 1.1.1 `framework`：
+
+```julia
+
+function hydro_forecast(f; model, outdir)
+  res = ...
+
+  fwrite(res.output, ...)
+  fwrite(res.gof, ...)
+  fwrite(res.info_flood, ...)
+  fwrite(res.dat_flood, ...)
+  fwrite(res.evaluation, ...)
+end
+
+function hydro_forecast(X::AbstractArray, Y::AbstractArray; model::Function, outdir = "OUTPUT")
+  mkpath(outdir)
+  res; # return a NamedTuple
+end
+```
+
+**输入**：X, Y, model
+
+**输出**：Qsim, GOF, Pass_rate
+
+  + `output`: 三类数据集的输出，A DataFrame with columns of `date`, `Qsim`, 
+  + `gof`: 三类数据的拟合优度
+  + `info_flood`: 洪水场次信息，`id`, `time_beg`, `time_end`, `duration`, `Q_peak`, `Q_min`
+  + `dat_flood`：洪水场次的驱动数据，
+  + `evaluation`: 每个洪水场次上的模拟优度, csv
+
+**绘图**：
+
+  + 交给他绘图的函数，数据
+
+**总结**：
+
+  + `evaluation`总结模型预报精度 （`AI执行`）
+
+
+**内部模块设计**：
+
+  + `flood_division`: 采用R语言，划分洪水场次
+
+  + `划分数据集`：train, test, valid
+
+  + `loss`: 根据拟合优度指标去设计loss，例如KGE, NSE, RMSE，注意loss越小越优。根据loss去优选模型参数。
+ 
+  + `evaluation`: 在三种数据集，train, test, valid。每个洪水场次的洪峰、峰现时间合格率。
+
+
+### 1.1.2 `model`：水文模型、LSTM、TCN、KAN
+
+  ```julia
+  Ysim = Model(X, Y; params, state) # Lux的设计哲学
+  ```
+
+### 1.1.3 文件保存
+
+文件保存采用Julia包`DataFrames`，`RTableTools`
+
+```julia
+using RTableTools
+fwrite(df, "out.csv") # df is a DataFrame
+```
--- a/skills/hydro-forecast/examples.jl
+++ b/skills/hydro-forecast/examples.jl
@@ -0,0 +1,22 @@
+using HydroTools
+using Dates
+
+lat = 20.0
+doy = 120
+
+ws = HourAngleSunSet(lat, doy)
+# doy
+cal_Rsi_toa(lat, doy)
+
+# date
+date = Date(2010, 6, 12)
+doy = dayofyear(date)
+cal_Rsi_toa(lat, doy)
+
+# datetime
+time = DateTime(2010, 6, 12)
+doy = dayofyear(date)
+Rsi = cal_Rsi_toa(lat, doy) # [MJ d-1 m-2]
+
+
+MJ2W(Rsi) # [MJ d-1 m-2] to [W m-2]
--- a/skills/julia-hydrotools/SKILL.md
+++ b/skills/julia-hydrotools/SKILL.md
@@ -0,0 +1,32 @@
+---
+name: julia-hydrotools
+description: 计算短波辐射、长波辐射、潜在蒸散发、日出日落时间、湿度的基本变量处理。
+---
+
+# 1 运行环境说明
+
+- 在Julia中运行
+
+- 在julia中首先加载包，`using HydroTools`
+
+- 若没有包加载出错，则安装之，`using Pkg; Pkg.add("HydroTools")`
+
+
+## 1.1 函数说明
+
+- `cal_Rsi_toa(lat, J)`: daily extraterrestrial radiation in MJ m-2 day-1
+  + `lat`: latitude in deg
+  + `J`: doy of year
+  > 注意lat和J是scalar
+  > 如果是vector，按照Julia的语法，采用`cal_Rsi_toa.(lat, J)`调用
+  + 默认返回单位是`MJ d-1`，若想转为`W m-2`，需要调用[MJ2W]函数，告诉用户返回的数字单位
+
+
+## 1.2 文件保存
+
+文件保存采用Julia包`DataFrames`，`RTableTools`
+
+```julia
+using RTableTools
+fwrite(df, "out.csv") # df is a DataFrame
+```
--- a/skills/julia-hydrotools/examples.jl
+++ b/skills/julia-hydrotools/examples.jl
@@ -0,0 +1,22 @@
+using HydroTools
+using Dates
+
+lat = 20.0
+doy = 120
+
+ws = HourAngleSunSet(lat, doy)
+# doy
+cal_Rsi_toa(lat, doy)
+
+# date
+date = Date(2010, 6, 12)
+doy = dayofyear(date)
+cal_Rsi_toa(lat, doy)
+
+# datetime
+time = DateTime(2010, 6, 12)
+doy = dayofyear(date)
+Rsi = cal_Rsi_toa(lat, doy) # [MJ d-1 m-2]
+
+
+MJ2W(Rsi) # [MJ d-1 m-2] to [W m-2]
--- a/skills/julia-numerical/SKILL.md
+++ b/skills/julia-numerical/SKILL.md
@@ -0,0 +1,120 @@
+---
+name: julia-numerical
+description: Execute numerical calculations and mathematical computations using Julia. Use this skill for matrix operations, linear algebra, numerical integration, optimization, statistics, and scientific computing tasks.
+---
+
+# Julia Numerical Calculation Skill
+
+This skill enables you to execute numerical calculations using Julia, a high-performance programming language designed for numerical and scientific computing.
+
+## When to Use
+
+Use this skill when you need to:
+- Perform matrix operations and linear algebra
+- Solve differential equations
+- Execute numerical integration or optimization
+- Calculate statistical measures
+- Handle large-scale numerical computations
+- Work with complex mathematical operations
+
+## Setup
+
+Before using this skill, ensure Julia is installed on your system:
+
+```bash
+# On macOS (using Homebrew)
+brew install julia
+
+# On Linux (Ubuntu/Debian)
+sudo apt-get install julia
+
+# On Windows (using Chocolatey)
+choco install julia
+
+# Or download from https://julialang.org/downloads/
+```
+
+## Basic Examples
+
+### Linear Algebra
+
+```julia
+using LinearAlgebra
+
+# Create matrices
+A = [1 2; 3 4]
+B = [5 6; 7 8]
+
+# Matrix multiplication
+C = A * B
+
+# Eigenvalues and eigenvectors
+eigenvals, eigenvecs = eigen(A)
+
+# Matrix inverse
+A_inv = inv(A)
+```
+
+### Numerical Integration
+
+```julia
+using QuadGK
+
+# Define a function
+f(x) = sin(x) * exp(-x)
+
+# Integrate from 0 to ∞
+result, error = quadgk(f, 0, Inf)
+```
+
+### Optimization
+
+```julia
+using Optim
+
+# Define objective function
+f(x) = (x[1] - 2)^2 + (x[2] - 3)^2
+
+# Minimize
+result = optimize(f, [0.0, 0.0])
+```
+
+### Statistics
+
+```julia
+using Statistics
+
+data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
+
+# Statistical measures
+mean_val = mean(data)
+std_val = std(data)
+var_val = var(data)
+median_val = median(data)
+```
+
+## How to Use This Skill
+
+When you ask me to perform a numerical calculation:
+1. I'll identify the appropriate Julia packages needed
+2. Write Julia code to solve the problem
+3. Execute the code
+4. Return results and explanations
+
+## Common Julia Packages
+
+- **LinearAlgebra**: Matrix operations and linear algebra
+- **Statistics**: Statistical functions
+- **QuadGK**: Numerical integration
+- **Optim**: Optimization algorithms
+- **DifferentialEquations**: Solving differential equations
+- **Plots**: Visualization
+- **Distributions**: Probability distributions
+- **Random**: Random number generation
+
+## Notes
+
+- Julia is JIT-compiled, so first runs may include compilation time
+- Use `.jl` files for organizing longer scripts
+- Install packages with `using Pkg; Pkg.add("PackageName")`
+- Results are returned as Julia objects that are converted to readable format
--- a/skills/julia-numerical/examples.jl
+++ b/skills/julia-numerical/examples.jl
@@ -0,0 +1,155 @@
+# Julia Numerical Calculation Examples
+# This file contains common numerical computation patterns
+
+# ============================================================================
+# Linear Algebra Examples
+# ============================================================================
+
+function linear_algebra_examples()
+    using LinearAlgebra
+
+    println("=== Linear Algebra Examples ===")
+
+    # Matrix creation and basic operations
+    A = [1 2 3; 4 5 6; 7 8 10]
+    b = [1, 2, 3]
+
+    println("Matrix A:")
+    println(A)
+
+    # Solve linear system Ax = b
+    x = A \ b
+    println("\nSolution to Ax = b:")
+    println(x)
+
+    # Eigenvalues
+    eigenvals, eigenvecs = eigen(A)
+    println("\nEigenvalues:")
+    println(eigenvals)
+
+    # Singular value decomposition
+    U, S, V = svd(A)
+    println("\nSingular values:")
+    println(S)
+
+    # Determinant and norm
+    println("\nDeterminant: ", det(A))
+    println("Frobenius norm: ", norm(A))
+end
+
+
+# ============================================================================
+# Numerical Integration Examples
+# ============================================================================
+
+function integration_examples()
+    using QuadGK
+
+    println("\n=== Numerical Integration Examples ===")
+
+    # Integrate sin(x) from 0 to π
+    f1(x) = sin(x)
+    result1, error1 = quadgk(f1, 0, π)
+    println("∫sin(x)dx from 0 to π = ", result1)
+    println("Estimated error: ", error1)
+
+    # Integrate exp(-x^2) from -∞ to ∞ (Gaussian)
+    f2(x) = exp(-x^2)
+    result2, error2 = quadgk(f2, -Inf, Inf)
+    println("\n∫exp(-x²)dx from -∞ to ∞ = ", result2)
+    println("Theoretical value: ", sqrt(π))
+
+    # Integrate 1/(1+x^2) from 0 to 1
+    f3(x) = 1/(1 + x^2)
+    result3, error3 = quadgk(f3, 0, 1)
+    println("\n∫1/(1+x²)dx from 0 to 1 = ", result3)
+    println("Theoretical value (π/4): ", π/4)
+end
+
+# ============================================================================
+# Optimization Examples
+# ============================================================================
+
+function optimization_examples()
+    using Optim
+
+    println("\n=== Optimization Examples ===")
+
+    # Simple quadratic function
+    f(x) = (x[1] - 2)^2 + (x[2] - 3)^2
+
+    result = optimize(f, [0.0, 0.0])
+    println("Minimize f(x,y) = (x-2)² + (y-3)²")
+    println("Minimum found at: ", Optim.minimizer(result))
+    println("Minimum value: ", Optim.minimum(result))
+
+    # Rosenbrock function (more challenging)
+    rosenbrock(x) = (1 - x[1])^2 + 100(x[2] - x[1]^2)^2
+
+    result2 = optimize(rosenbrock, [0.0, 0.0])
+    println("\nMinimize Rosenbrock function")
+    println("Minimum found at: ", Optim.minimizer(result2))
+    println("Minimum value: ", Optim.minimum(result2))
+end
+
+# ============================================================================
+# Statistics Examples
+# ============================================================================
+
+function statistics_examples()
+    using Statistics
+
+    println("\n=== Statistics Examples ===")
+
+    data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20]
+
+    println("Data: ", data)
+    println("\nStatistical measures:")
+    println("Mean: ", mean(data))
+    println("Median: ", median(data))
+    println("Standard deviation: ", std(data))
+    println("Variance: ", var(data))
+    println("Minimum: ", minimum(data))
+    println("Maximum: ", maximum(data))
+    println("Range: ", maximum(data) - minimum(data))
+
+    # Quantiles
+    println("\nQuantiles:")
+    println("25th percentile: ", quantile(data, 0.25))
+    println("50th percentile: ", quantile(data, 0.50))
+    println("75th percentile: ", quantile(data, 0.75))
+end
+
+# ============================================================================
+# Root Finding Examples
+# ============================================================================
+
+function root_finding_examples()
+    using Roots
+
+    println("\n=== Root Finding Examples ===")
+
+    # Find root of f(x) = x^3 - 2
+    f(x) = x^3 - 2
+    root = find_zero(f, 1.0)
+    println("Root of x³ - 2 = 0: ", root)
+    println("Verification: f(root) = ", f(root))
+
+    # Find root of f(x) = sin(x) - 0.5
+    f2(x) = sin(x) - 0.5
+    root2 = find_zero(f2, 0.5)
+    println("\nRoot of sin(x) - 0.5 = 0: ", root2)
+    println("Verification: f(root) = ", f2(root2))
+end
+
+# ============================================================================
+# Main execution
+# ============================================================================
+
+if abspath(PROGRAM_FILE) == @__FILE__
+    linear_algebra_examples()
+    integration_examples()
+    optimization_examples()
+    statistics_examples()
+    root_finding_examples()
+end
--- a/skills/julia-numerical/test_basic.jl
+++ b/skills/julia-numerical/test_basic.jl
@@ -0,0 +1,34 @@
+# Basic Julia numerical test
+using LinearAlgebra
+using Statistics
+
+println("Testing Julia Numerical Calculation Skill")
+println("==========================================\n")
+
+# Test 1: Basic arithmetic
+println("Test 1: Basic Arithmetic")
+result = 2 + 2 * 3
+println("2 + 2 * 3 = ", result)
+
+# Test 2: Vector operations
+println("\nTest 2: Vector Operations")
+v1 = [1, 2, 3]
+v2 = [4, 5, 6]
+dot_product = dot(v1, v2)
+println("dot([1,2,3], [4,5,6]) = ", dot_product)
+
+# Test 3: Matrix operations
+println("\nTest 3: Matrix Operations")
+A = [1 2; 3 4]
+println("Matrix A:")
+println(A)
+println("det(A) = ", det(A))
+
+# Test 4: Statistics
+println("\nTest 4: Statistics")
+data = [10, 20, 30, 40, 50]
+println("Data: ", data)
+println("mean = ", mean(data))
+println("std = ", std(data))
+
+println("\n✓ All basic tests passed!")
--- a/skills/typst-physica/SKILL.md
+++ b/skills/typst-physica/SKILL.md
@@ -0,0 +1,53 @@
+---
+name: typst-physica
+description: typst公式中的微分、偏微分方程编写，latex公式转typst。
+---
+
+
+# 引用包
+
+应在typst文档的开头，引用包。公式编写、文档排版，依赖`modern-cug-report`。
+
+用于如果已经引用了`modern-cug-report`，则无需再重复添加了。
+
+```typst
+#import "@local/modern-cug-report:0.1.3": *
+#show: doc => template(doc, footer: "CUG水文气象学2025", header: "")
+```
+
+
+# 偏微分方程
+
+- `(∂ theta) / (∂ t)`
+
+  `\frac{\partial \theta}{\partial t}`采用typst编写会非常简单，`pdv(theta, t)`
+
+  ```typst
+  (partial.diff theta) / (partial.diff t) // 是错误写法
+  pdv(theta, t)                           // 正确写法
+  ```
+
+- `(d theta) / (d t)`则是：`dv(theta, t)`
+
+
+# text
+
+typst公式中的本文需要使用引号：
+
+```typst
+q_(infiltration)   // 错误
+q_("infiltration") // 正确
+```
+
+# fraction
+
+- latex的`\frac{y}{x}`，写成typst则是`y/x`；
+  
+  若分子、分母有多个变量，则用括号括起来。例如latex的`\frac{y z}{x}`，写成typst则是`(y z) / x`
+
+
+# 排版
+
+- 一级标题之前空两行，凸显章节的层次感。
+
+- 第一个一级标题，不用空两行。
--- a/skills/typst-physica/example_physica.typ
+++ b/skills/typst-physica/example_physica.typ
@@ -0,0 +1,40 @@
+// Copyright 2023 Leedehai
+// Use of this code is governed by a MIT license in the LICENSE.txt file.
+// For a manual on this package, see physica-manual.pdf.
+
+#import "@local/modern-cug-report:0.1.3": *
+#show: doc => template(doc, footer: "CUG水文气象学2025", header: "")
+
+// #import "physica.typ": *
+
+#show: super-T-as-transpose // Render "..^T" as transposed matrix
+
+$
+  A^T, curl vb(E) = - pdv(vb(B), t),
+  quad
+  tensor(Lambda, +mu, -nu) = dmat(1, RR),
+  quad
+  f(x,y) dd(x, y),
+  quad
+  dd(vb(x), y, [3]),
+  quad
+  dd(x, y, 2, d: Delta, p: and),
+  quad
+  dv(phi, t, d: upright(D)) = pdv(phi, t) + vb(u) grad phi \
+  H(f) = hmat(f; x, y; delim: "[", big: #true),
+  quad
+  vb(v^a) = sum_(i=1)^n alpha_i vu(u^i),
+  quad
+  Set((x, y), pdv(f, x, y, [2,1]) + pdv(f, x, y, [1,2]) < epsilon) \
+  -1/c^2 pdv(, t, 2)psi + laplacian psi = (m^2c^2) / hbar^2 psi,
+  quad
+  ket(n^((1))) = sum_(k in.not D) mel(k^((0)), V, n^((0))) / (E_n^((0)) - E_k^((0))) ket(k^((0))),
+  quad
+  integral_V dd(V) (pdv(cal(L), phi) - partial_mu (pdv(cal(L), (partial_mu phi)))) = 0 \
+  dd(s, 2) = -(1-(2G M)/r) dd(t, 2) + (1-(2G M)/r)^(-1) dd(r, 2) + r^2 dd(Omega, 2)
+$
+
+$
+  "clk:" & signals("|1....|0....|1....|0....|1....|0....|1....|0..", step: #0.5em) \
+  "bus:" & signals(" #.... X=... ..... ..... X=... ..... ..... X#.", step: #0.5em)
+$
--- a/skills/typst-physica/examples.typ
+++ b/skills/typst-physica/examples.typ
@@ -0,0 +1,36 @@
+#import "@local/modern-cug-report:0.1.3": *
+#show: doc => template(doc, footer: "CUG水文气象学2025", header: "")
+
+
+== 1 Richards方程
+
+Richards方程：
+
+$ pdv(theta, t) = nabla dot [K(theta) nabla H] + S $
+
+其中：
+- $theta$：体积含水量 [L^3/L^3]
+- $t$：时间 [T]
+- $S$：源汇项 [1/T]
+
+总水头 $H$ 由基质势 $h$ 和重力势 $z$ 组成：
+$ H = h + z $
+
+
+== 2 质量守恒定律
+对于土壤控制体积，质量守恒方程为：
+$ pdv(rho theta, t) + nabla dot (rho q) = rho S $
+
+假设水密度 $rho$ 为常数，简化为：
+$ pdv(theta, t) + nabla dot q = S $
+
+
+== 3 上边界层条件
+
+上边界通常受大气条件控制，主要包括：
+
+*降雨入渗条件：*
+$ -K(theta) pdv(H, z) |_(z=0) = q_("infiltration") $
+
+*蒸发条件：*
+$ -K(theta) pdv(H, z) |_(z=0) = q_("evaporation") $