Mat*_*ols 40 c# unit-testing code-generation
我正在寻找一个内存中的对象(或对象的JSON序列化)并发出C#代码来生成一个等效的对象.
这对于从存储库中提取已知良好的示例以用作单元测试中的起点非常有用.我们已经考虑过反序列化JSON,但C#代码在重构方面会有优势.
Eva*_*van 11
如果模型很简单,可以使用反射和字符串构建器直接输出C#.我这样做是为了填充单元测试数据,就像你讨论的那样.
下面的代码示例是在几分钟内编写的,并生成了一个需要进行一些手动调整的对象初始化程序.如果你计划做很多事情,可以编写一个更强大/更少的错误功能.
第二个函数是递归的,迭代对象中的任何列表并为这些列表生成代码.
免责声明:这适用于我的基本数据类型的简单模型.它生成了需要清理的代码,但允许我快速前进.它只是作为如何做到这一点的一个例子.希望它激励某人写自己的.
在我的例子中,我有一个从数据库加载的大型数据集(结果)的实例.为了从我的单元测试中删除数据库依赖项,我将对象交给了这个函数,该函数吐出了允许我在我的测试类中模拟对象的代码.
private void WriteInstanciationCodeFromObject(IList results)
{
//declare the object that will eventually house C# initialization code for this class
var testMockObject = new System.Text.StringBuilder();
//start building code for this object
ConstructAndFillProperties(testMockObject, results);
var codeOutput = testMockObject.ToString();
}
private void ConstructAndFillProperties(StringBuilder testMockObject, IList results)
{
testMockObject.AppendLine("var testMock = new " + results.GetType().ToString() + "();");
foreach (object obj in results)
{
//if this object is a list, write code for it's contents
if (obj.GetType().GetInterfaces().Contains(typeof(IList)))
{
ConstructAndFillProperties(testMockObject, (IList)obj);
}
testMockObject.AppendLine("testMock.Add(new " + obj.GetType().Name + "() {");
foreach (var property in obj.GetType().GetProperties())
{
//if this property is a list, write code for it's contents
if (property.PropertyType.GetInterfaces().Contains(typeof(IList)))
{
ConstructAndFillProperties(testMockObject, (IList)property.GetValue(obj, null));
}
testMockObject.AppendLine(property.Name + " = (" + property.PropertyType + ")\"" + property.GetValue(obj, null) + "\",");
}
testMockObject.AppendLine("});");
}
}
该对象可能具有支持转换为InstanceDescriptor的TypeConverter ，这是WinForms设计器在发出C＃代码生成对象时使用的格式。如果它不能转换为InstanceDescriptor，它将尝试使用无参数构造函数并简单地设置公共属性。InstanceDescriptor机制很方便，因为它允许您指定各种构造选项，例如带参数的构造函数，甚至是静态工厂方法调用。
我已经编写了一些实用程序代码，它们使用IL发出内存对象的加载，该对象基本上遵循上述模式（如果可能，请使用InstanceDescriptor；否则，只需编写公共属性。）请注意，这只会产生一个如果InstanceDescriptor正确实现或设置公共属性足以恢复对象状态，则等效对象。如果要发出IL，也可以直接作弊和读取/写入字段值（这是DataContractSerializer支持的功能），但是有很多令人讨厌的极端情况需要考虑。
可能有点晚了，但这是我对这个问题的 5 美分。
提到的 Visual Studio 扩展 (OmarElabd/ObjectExporter) 是一个好主意，但我需要在单元测试执行期间在运行时从内存中对象生成 C# 代码。这是从最初的问题演变而来的：https ://www.nuget.org/packages/ObjectDumper.NET/
ObjectDumper.Dump(obj, DumpStyle.CSharp);从变量返回 C# 初始值设定项代码。如果您发现问题，请告诉我，您可能想在 github 上报告它们。
我也是这方面的新手,但我还需要一个定义层次结构的C#对象并将其提取到对象初始化器以简化设置单元测试.我从上面借了很多钱,结果就是这样.我想改进它处理识别用户类的方式.
http://github.com/jefflomax/csharp-object-to-object-literal/blob/master/Program.cs
using System;
using System.Collections;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
namespace ObjectInitializer
{
public class Program
{
public enum Color { Red, Green, Blue, Yellow, Fidget } ;
public class Foo
{
public int FooId { get; set; }
public string FooName { get; set; }
}
public class Thing
{
public int ThingId { get; set; }
public string ThingName { get; set; }
public List<Foo> Foos { get; set; }
}
public class Widget
{
public long Sort { get; set; }
public char FirstLetter { get; set; }
}
public class TestMe
{
public Color Color { get; set; }
public long Key { get; set; }
public string Name { get; set; }
public DateTime Created { get; set; }
public DateTime? NCreated { get; set; }
public bool Deleted { get; set; }
public bool? NDeleted { get; set; }
public double Amount { get; set; }
public Thing MyThing { get; set; }
public List<Thing> Things { get; set; }
public List<Widget> Widgets { get; set; }
}
static void Main(string[] args)
{
var testMe = new TestMe
{
Color = Program.Color.Blue,
Key = 3,
Name = "SAK",
Created = new DateTime(2013,10,20,8,0,0),
NCreated = (DateTime?)null,
Deleted = false,
NDeleted = null,
Amount = 13.1313,
MyThing = new Thing(){ThingId=1,ThingName="Thing 1"},
Things = new List<Thing>
{
new Thing
{
ThingId=4,
ThingName="Thing 4",
Foos = new List<Foo>
{
new Foo{FooId=1, FooName="Foo 1"},
new Foo{FooId=2,FooName="Foo2"}
}
},
new Thing
{
ThingId=5,
ThingName="Thing 5",
Foos = new List<Foo>()
}
},
Widgets = new List<Widget>()
};
var objectInitializer = ToObjectInitializer(testMe);
Console.WriteLine(objectInitializer);
// This is the returned C# Object Initializer
var x = new TestMe { Color = Program.Color.Blue, Key = 3, Name = "SAK", Created = new DateTime(2013, 10, 20, 8, 0, 0), NCreated = null, Deleted = false, NDeleted = null, Amount = 13.1313, MyThing = new Thing { ThingId = 1, ThingName = "Thing 1", Foos = new List<Foo>() }, Things = new List<Thing> { new Thing { ThingId = 4, ThingName = "Thing 4", Foos = new List<Foo> { new Foo { FooId = 1, FooName = "Foo 1" }, new Foo { FooId = 2, FooName = "Foo2" } } }, new Thing { ThingId = 5, ThingName = "Thing 5", Foos = new List<Foo>() } }, Widgets = new List<Widget>() };
Console.WriteLine("");
}
public static string ToObjectInitializer(Object obj)
{
var sb = new StringBuilder(1024);
sb.Append("var x = ");
sb = WalkObject(obj, sb);
sb.Append(";");
return sb.ToString();
}
private static StringBuilder WalkObject(Object obj, StringBuilder sb)
{
var properties = obj.GetType().GetProperties();
var type = obj.GetType();
var typeName = type.Name;
sb.Append("new " + type.Name + " {");
bool appendComma = false;
DateTime workDt;
foreach (var property in properties)
{
if (appendComma) sb.Append(", ");
appendComma = true;
var pt = property.PropertyType;
var name = pt.Name;
var isList = property.PropertyType.GetInterfaces().Contains(typeof(IList));
var isClass = property.PropertyType.IsClass;
if (isList)
{
IList list = (IList)property.GetValue(obj, null);
var listTypeName = property.PropertyType.GetGenericArguments()[0].Name;
if (list != null && list.Count > 0)
{
sb.Append(property.Name + " = new List<" + listTypeName + ">{");
sb = WalkList( list, sb );
sb.Append("}");
}
else
{
sb.Append(property.Name + " = new List<" + listTypeName + ">()");
}
}
else if (property.PropertyType.IsEnum)
{
sb.AppendFormat("{0} = {1}", property.Name, property.GetValue(obj));
}
else
{
var value = property.GetValue(obj);
var isNullable = pt.IsGenericType && pt.GetGenericTypeDefinition() == typeof(Nullable<>);
if (isNullable)
{
name = pt.GetGenericArguments()[0].Name;
if (property.GetValue(obj) == null)
{
sb.AppendFormat("{0} = null", property.Name);
continue;
}
}
switch (name)
{
case "Int64":
case "Int32":
case "Int16":
case "Double":
case "Float":
sb.AppendFormat("{0} = {1}", property.Name, value);
break;
case "Boolean":
sb.AppendFormat("{0} = {1}", property.Name, Convert.ToBoolean(value) == true ? "true" : "false");
break;
case "DateTime":
workDt = Convert.ToDateTime(value);
sb.AppendFormat("{0} = new DateTime({1},{2},{3},{4},{5},{6})", property.Name, workDt.Year, workDt.Month, workDt.Day, workDt.Hour, workDt.Minute, workDt.Second);
break;
case "String":
sb.AppendFormat("{0} = \"{1}\"", property.Name, value);
break;
default:
// Handles all user classes, should likely have a better way
// to detect user class
sb.AppendFormat("{0} = ", property.Name);
WalkObject(property.GetValue(obj), sb);
break;
}
}
}
sb.Append("}");
return sb;
}
private static StringBuilder WalkList(IList list, StringBuilder sb)
{
bool appendComma = false;
foreach (object obj in list)
{
if (appendComma) sb.Append(", ");
appendComma = true;
WalkObject(obj, sb);
}
return sb;
}
}
}
小智 5
我在寻找马修所描述的同样方法时偶然发现了这一点,并且受到埃文回答编写我自己的扩展方法的启发.它生成可编译的C#代码作为可以复制/粘贴到Visual Studio中的字符串.我没有打扰任何特定的格式,只是在一行输出代码,并使用ReSharper很好地格式化它.我已经将它用于我们传递的一些大型DTO,到目前为止,它就像一个魅力.
这是扩展方法和几个辅助方法:
public static string ToCreationMethod(this object o)
{
return String.Format("var newObject = {0};", o.CreateObject());
}
private static StringBuilder CreateObject(this object o)
{
var builder = new StringBuilder();
builder.AppendFormat("new {0} {{ ", o.GetClassName());
foreach (var property in o.GetType().GetProperties())
{
var value = property.GetValue(o);
if (value != null)
{
builder.AppendFormat("{0} = {1}, ", property.Name, value.GetCSharpString());
}
}
builder.Append("}");
return builder;
}
private static string GetClassName(this object o)
{
var type = o.GetType();
if (type.IsGenericType)
{
var arg = type.GetGenericArguments().First().Name;
return type.Name.Replace("`1", string.Format("<{0}>", arg));
}
return type.Name;
}
GetCSharpString方法包含逻辑,并且对任何特定类型的扩展都是开放的.对我来说,它处理字符串,整数,小数,日期以及实现IEnumerable的任何东西都足够了:
private static string GetCSharpString(this object o)
{
if (o is String)
{
return string.Format("\"{0}\"", o);
}
if (o is Int32)
{
return string.Format("{0}", o);
}
if (o is Decimal)
{
return string.Format("{0}m", o);
}
if (o is DateTime)
{
return string.Format("DateTime.Parse(\"{0}\")", o);
}
if (o is IEnumerable)
{
return String.Format("new {0} {{ {1}}}", o.GetClassName(), ((IEnumerable)o).GetItems());
}
return string.Format("{0}", o.CreateObject());
}
private static string GetItems(this IEnumerable items)
{
return items.Cast<object>().Aggregate(string.Empty, (current, item) => current + String.Format("{0}, ", item.GetCSharpString()));
}
我希望有人觉得这很有用!
| 归档时间： |
|
| 查看次数： |
5077 次 |
| 最近记录： |