diff --git a/NumSharp_Implementation_Documentation.md b/NumSharp_Implementation_Documentation.md new file mode 100644 index 00000000..c1e24a66 --- /dev/null +++ b/NumSharp_Implementation_Documentation.md @@ -0,0 +1,410 @@ +# NumSharp使用指南及核心组件详解 + +本文档详细介绍了NumSharp数值计算库的核心组件NDArray、Slice和Shape的使用方法，包含创建、修改、访问和操作数组的具体示例。 + +## 目录 + +1. [NDArray基本操作](#ndarray基本操作) +2. [Slice切片操作](#slice切片操作) +3. [Shape形状管理](#shape形状管理) +4. [高级操作示例](#高级操作示例) + +## NDArray基本操作 + +NDArray是NumSharp中的核心数据结构，代表多维同质数组。 + +### 创建NDArray + +#### 从数组创建 +```csharp +using NumSharp; + +// 从C#数组创建 +int[] arr1D = {1, 2, 3, 4, 5}; +NDArray nd1 = new NDArray(arr1D); + +double[,] arr2D = {{1.0, 2.0}, {3.0, 4.0}}; +NDArray nd2 = new NDArray(arr2D); + +// 使用np创建函数 +NDArray zeros = np.zeros(3, 4); // 3x4零矩阵 +NDArray ones = np.ones(2, 3, 4); // 2x3x4全1矩阵 +NDArray range = np.arange(10); // [0, 1, 2, ..., 9] +NDArray linspace = np.linspace(0, 10, 5); // [0, 2.5, 5, 7.5, 10] +``` + +#### 指定数据类型创建 +```csharp +// 指定数据类型和形状创建 +NDArray intArr = new NDArray(typeof(int), new Shape(3, 4)); // 3x4整数数组 +NDArray floatArr = new NDArray(NPTypeCode.Single, new Shape(2, 3)); // 2x3单精度浮点数组 +``` + +### 访问和修改数据 + +#### 获取数组信息 +```csharp +NDArray arr = np.arange(12).reshape(3, 4); + +Console.WriteLine($"数据类型: {arr.dtype}"); // 数据类型: System.Int32 +Console.WriteLine($"形状: [{string.Join(", ", arr.shape)}]"); // 形状: [3, 4] +Console.WriteLine($"维度数: {arr.ndim}"); // 维度数: 2 +Console.WriteLine($"元素总数: {arr.size}"); // 元素总数: 12 +Console.WriteLine($"数据大小: {arr.dtypesize}"); // 数据大小: 4 (字节) +``` + +#### 元素访问 +```csharp +NDArray arr = np.array(new int[] {10, 20, 30, 40}); + +// 访问单个元素 +int val = arr.GetInt32(2); // val = 30 +Console.WriteLine($"第三元素: {val}"); + +// 二维数组访问 +NDArray mat = np.array(new int[,] {{1, 2, 3}, {4, 5, 6}}); +int val2d = mat.GetInt32(1, 2); // val2d = 6 (第二行第三列) +Console.WriteLine($"(1,2)位置值: {val2d}"); + +// 通用访问方法 +var genericVal = mat.GetValue(0, 1); // genericVal = 2 +``` + +#### 元素修改 +```csharp +NDArray arr = np.arange(5); + +// 修改单个元素 +arr.SetValue(99, 2); // 将索引2处的元素改为99 +Console.WriteLine($"修改后数组: {arr}"); + +// 批量修改 +arr.SetData(new int[] {100, 200}, 1, 3); // 修改索引1和3处的值 + +// 修改二维数组 +NDArray mat = np.zeros(2, 3); +mat.SetInt32(42, 0, 1); // 将(0,1)位置的值设为42 +Console.WriteLine($"修改后矩阵:\n{mat}"); +``` + +#### 批量数据操作 +```csharp +// 获取内部数据的引用 +NDArray arr = np.array(new double[] {1.1, 2.2, 3.3}); +ArraySlice data = arr.Data(); +Console.WriteLine($"数据: [{string.Join(", ", data)}]"); + +// 批量设置数据 +NDArray target = new NDArray(typeof(double), 3); +target.SetData(new double[] {5.5, 6.6, 7.7}); +Console.WriteLine($"目标数组: {target}"); +``` + +## Slice切片操作 + +Slice类提供了类似NumPy的切片功能，可以方便地对数组进行部分访问和修改。 + +### 基本切片操作 +```csharp +NDArray arr = np.arange(10); // [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] + +// 使用Slice对象 +NDArray slice1 = arr[new Slice(2, 7)]; // [2, 3, 4, 5, 6] - 从索引2到6 +NDArray slice2 = arr[new Slice(null, 5)]; // [0, 1, 2, 3, 4] - 从开始到索引4 +NDArray slice3 = arr[new Slice(3, null)]; // [3, 4, 5, 6, 7, 8, 9] - 从索引3到结束 + +Console.WriteLine($"原数组: {arr}"); +Console.WriteLine($"切片[2:7]: {slice1}"); +Console.WriteLine($"切片[:5]: {slice2}"); +Console.WriteLine($"切片[3:]: {slice3}"); +``` + +### 带步长的切片 +```csharp +NDArray arr = np.arange(10); // [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] + +// 步长切片 +NDArray stepSlice = arr[new Slice(0, null, 2)]; // [0, 2, 4, 6, 8] - 步长为2 +NDArray reverseSlice = arr[new Slice(null, null, -1)]; // [9, 8, 7, 6, 5, 4, 3, 2, 1, 0] - 反转 +NDArray reverseStep = arr[new Slice(null, null, -2)]; // [9, 7, 5, 3, 1] - 反转步长为2 + +Console.WriteLine($"步长为2的切片: {stepSlice}"); +Console.WriteLine($"反向切片: {reverseSlice}"); +Console.WriteLine($"反向步长为2的切片: {reverseStep}"); +``` + +### 多维数组切片 +```csharp +// 创建3x4数组 +NDArray mat = np.arange(12).reshape(3, 4); +/* +mat = +[[ 0, 1, 2, 3], + [ 4, 5, 6, 7], + [ 8, 9, 10, 11]] +*/ + +// 行切片 +NDArray rowSlice = mat[new Slice(1, 3), Slice.All]; // 第1到2行 +Console.WriteLine($"第1-2行:\n{rowSlice}"); + +// 列切片 +NDArray colSlice = mat[Slice.All, new Slice(0, 2)]; // 第0到1列 +Console.WriteLine($"第0-1列:\n{colSlice}"); + +// 组合切片 +NDArray subMat = mat[new Slice(0, 2), new Slice(1, 3)]; // 前2行的第1-2列 +Console.WriteLine($"子矩阵:\n{subMat}"); + +// 单个索引（会减少一个维度） +NDArray row = mat[1, Slice.All]; // 第1行，结果是1维数组 +Console.WriteLine($"第1行: {row}"); + +NDArray col = mat[Slice.All, 2]; // 第2列，结果是1维数组 +Console.WriteLine($"第2列: {col}"); +``` + +### 使用字符串切片表示法 +```csharp +NDArray arr = np.arange(10); + +// 使用字符串表示法 +NDArray slice1 = arr["2:7"]; // 等同于 new Slice(2, 7) +NDArray slice2 = arr["::2"]; // 等同于 new Slice(null, null, 2) +NDArray slice3 = arr["::-1"]; // 等同于 new Slice(null, null, -1) + +Console.WriteLine($"字符串表示法切片'2:7': {slice1}"); +Console.WriteLine($"字符串表示法切片'::2': {slice2}"); +Console.WriteLine($"字符串表示法切片'::-1': {slice3}"); +``` + +### 切片修改操作 +```csharp +NDArray arr = np.arange(10); +Console.WriteLine($"原始数组: {arr}"); + +// 修改切片会影响原数组（因为切片是视图） +NDArray slice = arr[new Slice(2, 5)]; +slice.fill(99); // 将切片中的所有元素设置为99 +Console.WriteLine($"修改后的原数组: {arr}"); // [0, 1, 99, 99, 99, 5, 6, 7, 8, 9] + +// 二维数组切片修改 +NDArray mat = np.zeros(3, 4); +mat[new Slice(1, 3), new Slice(1, 3)] = np.ones(2, 2) * 5; +Console.WriteLine($"修改后的矩阵:\n{mat}"); +``` + +## Shape形状管理 + +Shape类管理数组的形状信息，包括维度、大小和内存布局。 + +### Shape基本操作 +```csharp +// 创建Shape对象 +Shape shape1 = new Shape(3, 4); // 3x4形状 +Shape shape2 = new Shape(2, 3, 4); // 2x3x4形状 +Shape shape3 = new Shape(new int[] {5, 6, 7}); // 从数组创建 + +Console.WriteLine($"shape1: {shape1}, 维度: {shape1.NDim}, 大小: {shape1.Size}"); +Console.WriteLine($"shape2: {shape2}, 维度: {shape2.NDim}, 大小: {shape2.Size}"); +Console.WriteLine($"shape3: {shape3}, 维度: {shape3.NDim}, 大小: {shape3.Size}"); + +// Shape的维度访问 +Console.WriteLine($"shape1的第0维: {shape1[0]}"); // 3 +Console.WriteLine($"shape1的第1维: {shape1[1]}"); // 4 +``` + +### Shape转换操作 +```csharp +NDArray arr = np.arange(12); + +// 重塑形状 +NDArray reshaped1 = arr.reshape(3, 4); // 3x4矩阵 +NDArray reshaped2 = arr.reshape(2, 6); // 2x6矩阵 +NDArray reshaped3 = arr.reshape(2, 2, 3); // 2x2x3三维数组 + +Console.WriteLine($"原数组形状: {arr.shape}"); +Console.WriteLine($"重塑为3x4:\n{reshaped1}"); +Console.WriteLine($"重塑为2x6:\n{reshaped2}"); +Console.WriteLine($"重塑为2x2x3:\n{reshaped3}"); + +// 扁平化 +NDArray flat = reshaped1.flatten(); // 转换为一维数组 +Console.WriteLine($"扁平化后: {flat}"); +``` + +### Shape的高级操作 +```csharp +NDArray arr = np.arange(24).reshape(2, 3, 4); + +// 获取子形状 +var (subShape, offset) = arr.Shape.GetSubshape(1); // 获取第二"页"的形状和偏移 +Console.WriteLine($"子形状: {subShape}, 偏移: {offset}"); + +// 获取坐标 +int linearIndex = 10; +int[] coords = arr.Shape.GetCoordinates(linearIndex); +Console.WriteLine($"线性索引{linearIndex}对应的坐标: [{string.Join(", ", coords)}]"); + +// 获取偏移 +int[] testCoords = {1, 2, 3}; +int offsetFromCoords = arr.Shape.GetOffset(testCoords); +Console.WriteLine($"坐标[{string.Join(", ", testCoords)}]对应的偏移: {offsetFromCoords}"); +``` + +## 高级操作示例 + +### 数学运算 +```csharp +// 基本数学运算 +NDArray a = np.array(new double[] {1, 2, 3, 4}); +NDArray b = np.array(new double[] {5, 6, 7, 8}); + +NDArray sum = a + b; // [6, 8, 10, 12] +NDArray product = a * b; // [5, 12, 21, 32] +NDArray subtract = b - a; // [4, 4, 4, 4] +NDArray divide = b / a; // [5, 3, 2.33, 2] + +Console.WriteLine($"a: {a}"); +Console.WriteLine($"b: {b}"); +Console.WriteLine($"a + b: {sum}"); +Console.WriteLine($"a * b: {product}"); + +// 一元运算 +NDArray sqrtA = np.sqrt(a); // [1, 1.41, 1.73, 2] +NDArray squareA = np.square(a); // [1, 4, 9, 16] +NDArray absA = np.abs(a - 2.5); // [1.5, 0.5, 0.5, 1.5] + +Console.WriteLine($"sqrt(a): {sqrtA}"); +Console.WriteLine($"square(a): {squareA}"); +Console.WriteLine($"abs(a - 2.5): {absA}"); +``` + +### 统计操作 +```csharp +NDArray arr = np.array(new double[,] {{1, 2, 3}, {4, 5, 6}}); + +// 基本统计 +double sumAll = np.sum(arr).GetDouble(); // 21 +double meanAll = np.mean(arr).GetDouble(); // 3.5 +double maxAll = np.max(arr).GetDouble(); // 6 +double minAll = np.min(arr).GetDouble(); // 1 + +Console.WriteLine($"总和: {sumAll}, 平均值: {meanAll}"); +Console.WriteLine($"最大值: {maxAll}, 最小值: {minAll}"); + +// 按轴统计 +NDArray sumAxis0 = np.sum(arr, axis: 0); // [5, 7, 9] - 按列求和 +NDArray sumAxis1 = np.sum(arr, axis: 1); // [6, 15] - 按行求和 + +Console.WriteLine($"按列求和: {sumAxis0}"); +Console.WriteLine($"按行求和: {sumAxis1}"); + +// 获取最值索引 +int argMax = np.argmax(arr).GetInt32(); // 5 (最大值6的索引) +int argMin = np.argmin(arr).GetInt32(); // 0 (最小值1的索引) + +Console.WriteLine($"最大值索引: {argMax}, 最小值索引: {argMin}"); +``` + +### 线性代数操作 +```csharp +// 矩阵乘法 +NDArray matA = np.array(new double[,] {{1, 2}, {3, 4}}); +NDArray matB = np.array(new double[,] {{5, 6}, {7, 8}}); + +NDArray matMul = np.dot(matA, matB); +Console.WriteLine($"矩阵A:\n{matA}"); +Console.WriteLine($"矩阵B:\n{matB}"); +Console.WriteLine($"A * B:\n{matMul}"); + +// 转置 +NDArray transposed = matA.T; // 或者使用 np.transpose(matA) +Console.WriteLine($"A的转置:\n{transposed}"); + +// 其他操作 +NDArray reshaped = np.arange(12).reshape(3, 4); +NDArray transposed2D = np.transpose(reshaped); +Console.WriteLine($"原矩阵:\n{reshaped}"); +Console.WriteLine($"转置后:\n{transposed2D}"); +``` + +### 数组操作 +```csharp +// 数组合并 +NDArray arr1 = np.array(new int[] {1, 2, 3}); +NDArray arr2 = np.array(new int[] {4, 5, 6}); + +NDArray hStack = np.hstack(new NDArray[] {arr1, arr2}); // 水平拼接 +NDArray vStack = np.vstack(new NDArray[] {arr1, arr2}); // 垂直拼接 + +Console.WriteLine($"原数组1: {arr1}"); +Console.WriteLine($"原数组2: {arr2}"); +Console.WriteLine($"水平拼接: {hStack}"); +Console.WriteLine($"垂直拼接:\n{vStack}"); + +// 数组复制 +NDArray original = np.array(new int[] {1, 2, 3}); +NDArray cloned = original.Clone(); // 深拷贝 +NDArray view = original.view(); // 视图（共享数据） + +Console.WriteLine($"原数组: {original}"); +Console.WriteLine($"拷贝: {cloned}"); +Console.WriteLine($"视图: {view}"); + +// 类型转换 +NDArray floatArr = original.astype(typeof(float)); +Console.WriteLine($"转换为float: {floatArr}, 类型: {floatArr.dtype}"); +``` + +### 条件操作 +```csharp +NDArray arr = np.array(new int[] {-2, -1, 0, 1, 2}); + +// 条件选择 +NDArray positive = np.where(arr > 0, arr, 0); // 大于0的保持，否则设为0 +NDArray absValues = np.abs(arr); // 绝对值 + +Console.WriteLine($"原数组: {arr}"); +Console.WriteLine($"正数保持: {positive}"); +Console.WriteLine($"绝对值: {absValues}"); + +// 布尔索引（通过条件获取索引） +NDArray condition = arr > 0; +NDArray positiveOnly = arr[condition]; // 只获取正数 +Console.WriteLine($"正数元素: {positiveOnly}"); +``` + +### 排序和搜索 +```csharp +NDArray arr = np.array(new int[] {3, 1, 4, 1, 5, 9, 2, 6, 5}); + +// 排序 +NDArray sorted = np.sort(arr); +NDArray sortedIndices = np.argsort(arr); // 返回排序后的索引 + +Console.WriteLine($"原数组: {arr}"); +Console.WriteLine($"排序后: {sorted}"); +Console.WriteLine($"排序索引: {sortedIndices}"); + +// 唯一值 +NDArray unique = np.unique(arr); +Console.WriteLine($"唯一值: {unique}"); +``` + +### 广播操作 +```csharp +// NumSharp支持广播操作 +NDArray mat = np.array(new double[,] {{1, 2, 3}, {4, 5, 6}}); +NDArray vec = np.array(new double[] {10, 20, 30}); + +// 向量与矩阵的广播加法 +NDArray result = mat + vec; // 每行都加上向量vec +Console.WriteLine($"矩阵:\n{mat}"); +Console.WriteLine($"向量: {vec}"); +Console.WriteLine($"广播加法结果:\n{result}"); + +// 标量操作（自动广播到所有元素） +NDArray scaled = mat * 2.0; +Console.WriteLine($"矩阵乘以2:\n{scaled}"); +``` \ No newline at end of file diff --git a/docfx_project/articles/intro.md b/docfx_project/articles/intro.md index 6459c451..3773d894 100644 --- a/docfx_project/articles/intro.md +++ b/docfx_project/articles/intro.md @@ -1,4 +1,4 @@ -# Introduction +# Introduction The following pages are for the users who want to use NumSharp. @@ -61,11 +61,8 @@ Column wise also starts with the element [0,0] but! it stays in the 1st column a The image below (taken from https://en.wikipedia.org/wiki/File:Row_and_column_major_order.svg) shows again the 'algorithm' for storing data from matrix to vector. - ![Row Wise Column Wise](../images/rowWise_ColumnWise.png) - - **N dim tensor** Now we come to the most tricky question - how to store a general n dimensional tensor inside a 1D array. @@ -100,17 +97,3 @@ So fill first dimension, increase next, fill again, etc. also in n dimensional t And this you can imagine as **forward filling layout**. That's it. Now you have enough knowledge about NDArray, NDStorage and Shape. Check the other chapters for a how to use. :) - - - - - - - - - - - - - - diff --git a/examples/NeuralNetwork.NumSharp/Cost/BinaryCrossEntropy.cs b/examples/NeuralNetwork.NumSharp/Cost/BinaryCrossEntropy.cs index dbc68638..a42efc35 100644 --- a/examples/NeuralNetwork.NumSharp/Cost/BinaryCrossEntropy.cs +++ b/examples/NeuralNetwork.NumSharp/Cost/BinaryCrossEntropy.cs @@ -15,8 +15,7 @@ public BinaryCrossEntropy() : base("binary_crossentropy") public override NDArray Forward(NDArray preds, NDArray labels) { //ToDo: np.clip - //var output = Clip(preds, Epsilon, 1 - Epsilon); - var output = preds; + var output = np.clip(preds, Epsilon, 1 - Epsilon); output = np.mean(-(labels * np.log(output) + (1 - labels) * np.log(1 - output))); return output; } @@ -24,8 +23,8 @@ public override NDArray Forward(NDArray preds, NDArray labels) public override NDArray Backward(NDArray preds, NDArray labels) { //ToDo: np.clip - //var output = Clip(preds, Epsilon, 1 - Epsilon); - var output = preds; + var output = np.clip(preds, Epsilon, 1 - Epsilon); + output = preds; return (output - labels) / (output * (1 - output)); } } diff --git a/src/NumSharp.Core/APIs/np.where.cs b/src/NumSharp.Core/APIs/np.where.cs new file mode 100644 index 00000000..4b477457 --- /dev/null +++ b/src/NumSharp.Core/APIs/np.where.cs @@ -0,0 +1,71 @@ +using System; +using System.Linq; +using System.Collections; +using System.Collections.Generic; + +namespace NumSharp +{ + public static partial class np + { + /// + /// NumPy-like where: If only condition is given, return tuple of indices where condition is True. + /// If x and y are given, return elements chosen from x or y depending on condition. + /// Supports NDArray, bool[], IEnumerable<bool>, Array, IEnumerable, etc. + /// + public static NDArray where(NDArray condition, NDArray x, NDArray y) + { + var broadcasted = np.broadcast_arrays(condition, x, y); + var broadcasted_condition = broadcasted[0]; + var broadcasted_x = broadcasted[1]; + var broadcasted_y = broadcasted[2]; + + var cond = broadcasted_condition.MakeGeneric(); + var result = new NDArray(broadcasted_x.dtype, broadcasted_x.shape); + var condSpan = cond.GetData().AsSpan(); + var xArr = broadcasted_x.GetData(); + var yArr = broadcasted_y.GetData(); + var resArr = result.GetData(); + + for (int i = 0; i < condSpan.Length; i++) + { + resArr.SetIndex(i, condSpan[i] ? xArr.GetIndex(i) : yArr.GetIndex(i)); + } + return result; + } + + public static NDArray where(NDArray condition, NDArray x, NDArray y, Type outType) + { + return where(condition, x.astype(outType), y.astype(outType)); + } + + public static NDArray where(NDArray condition, NDArray x, NDArray y, NPTypeCode typeCode) + { + return where(condition, x.astype(typeCode), y.astype(typeCode)); + } + + public static NDArray[] where(NDArray condition) + { + return condition.TensorEngine.NonZero(condition); + } + + public static NDArray[] where(bool[] condition) + { + return where(np.array(condition)); + } + + public static NDArray[] where(IEnumerable condition) + { + return where(np.array(condition.ToArray())); + } + + public static NDArray where(bool[] condition, Array x, Array y) + { + return where(np.array(condition), np.array(x), np.array(y)); + } + + public static NDArray where(IEnumerable condition, IEnumerable x, IEnumerable y) + { + return where(np.array(condition.ToArray()), np.array(x.Cast().ToArray()), np.array(y.Cast().ToArray())); + } + } +} diff --git a/test/NumSharp.UnitTest/APIs/np.where.Test.cs b/test/NumSharp.UnitTest/APIs/np.where.Test.cs new file mode 100644 index 00000000..17608a1d --- /dev/null +++ b/test/NumSharp.UnitTest/APIs/np.where.Test.cs @@ -0,0 +1,78 @@ +using Microsoft.VisualStudio.TestTools.UnitTesting; +using NumSharp; + +namespace NumSharp.UnitTest.APIs +{ + [TestClass] + public class NpWhereTest + { + [TestMethod] + public void WhereWithConditionOnly() + { + var condition = np.array(new bool[] { true, false, true, false }); + var result = np.where(condition); + + Assert.AreEqual(1, result.Length); // Should return one array of indices + Assert.AreEqual(2, result[0].size); // Two true values + Assert.IsTrue(NDArray.Equals(result[0], np.array(new int[] { 0, 2 }))); // Indices should be [0, 2] + } + + [TestMethod] + public void WhereWith3Args1D() + { + var condition = np.array(new bool[] { true, false, true, false }); + var x = np.array(new int[] { 1, 2, 3, 4 }); + var y = np.array(new int[] { 10, 20, 30, 40 }); + + var result = np.where(condition, x, y); + + Assert.IsTrue(NDArray.Equals(result, np.array(new int[] { 1, 20, 3, 40 }))); + } + + [TestMethod] + public void WhereWith3Args2D() + { + var condition = np.array(new bool[,] { { true, false }, { true, true } }); + var x = np.array(new int[,] { { 1, 2 }, { 3, 4 } }); + var y = np.array(new int[,] { { 9, 8 }, { 7, 6 } }); + + var result = np.where(condition, x, y); + + Assert.IsTrue(NDArray.Equals(result, np.array(new int[,] { { 1, 8 }, { 3, 4 } }))); + } + + [TestMethod] + public void WhereWithBroadcasting() + { + var x = np.arange(3).reshape(3, 1); + var y = np.arange(4).reshape(1, 4); + var condition = x < y; + + var result = np.where(condition, x, 10 + y); + + // Expected: when x < y, take from x, otherwise from 10 + y + var expected = np.array(new int[,] { + { 10, 0, 0, 0 }, + { 10, 11, 1, 1 }, + { 10, 11, 12, 2 } + }); + + Assert.IsTrue(NDArray.Equals(result, expected)); + } + + [TestMethod] + public void WhereWithScalarBroadcasting() + { + var a = np.array(new int[,] { { 0, 1, 2 }, { 0, 2, 4 }, { 0, 3, 6 } }); + var result = np.where(a < 4, a, -1); + + var expected = np.array(new int[,] { + { 0, 1, 2 }, + { 0, 2, -1 }, + { 0, 3, -1 } + }); + + Assert.IsTrue(NDArray.Equals(result, expected)); + } + } +} diff --git a/test/NumSharp.UnitTest/Logic/np.any.Test.cs b/test/NumSharp.UnitTest/Logic/np.any.Test.cs index db11390b..5d1d2fe6 100644 --- a/test/NumSharp.UnitTest/Logic/np.any.Test.cs +++ b/test/NumSharp.UnitTest/Logic/np.any.Test.cs @@ -1,3 +1,6 @@ +using System; +using Microsoft.VisualStudio.TestTools.UnitTesting; +using System.Linq; using System; using Microsoft.VisualStudio.TestTools.UnitTesting; using NumSharp; @@ -13,6 +16,8 @@ public void Any1DArrayTest() // 测试1维数组 var arr = np.array(new int[] { 0, 1, 2 }); var result = np.any(arr, axis: 0, keepdims: false); + //TODO：TEST + Assert.AreEqual(true, result[0]); Assert.AreEqual(true, result.GetItem(0)); } @@ -105,4 +110,5 @@ public void AnyNullArrayTest() np.any(arr, axis: 0, keepdims: false); } } -} \ No newline at end of file +} +}