Reflector on the job

 

You have a list full of custom class items and want to find according to one property.Ok the best way is lambda functions(Or Not?).Other than lambda functions you can use reflection and use some code like:

 

public NeuronList FindByProperty(string PropertyName, object Value)

 {

     try

     {

         NeuronList foundItems = new NeuronList();

         foreach (NeuronBase item in List)

         {

             System.Type itemType = item.GetType();

             PropertyInfo[] itemProperties = itemType.GetProperties();

             foreach (PropertyInfo info in itemProperties)

             {

                 if (info.Name == PropertyName)

                 {

                     object itemValue = info.GetValue(item, null);

                     if (Convert.ChangeType(itemValue, info.PropertyType).Equals(Value))

                     {

                         foundItems.Add(item);

                     }

                 }

             }

         }

         return foundItems;

     }

     catch (Exception ex)

     {

         throw new Exception("Can not find property", ex);

     }

 }

Neural Networks

Have written a basic ann app which can do anything known about neural networks and has a simple ui, of course it needs lots of improvements but to get the basic idea you can take a look.

I am assuming you have some idea about neural networks

BaseNeuron is my neuron class which derives from user control. Input Neuron Middle Neuron and Output Neurons are derived from Base Neuron .Input Neuron is the neurons where you apply input output neurons are as name implies output :).Every neuron can be connected to another by synaps.A synaps becomes more powerful as it transmit signals

NeuralNetwork is our network and contains two collections inputneurons and neurons(middle and output).NeuralNetwork stores the fired neurons you can find all fired neurons in our network in nth turn.

NeuralPort is the ports where you can make connections(input and output) to ann.

NeuronList is our List that hold neurons and enables us to find desired neuron by neuronNo

NetworkVisualiser is the gui that shows our neural network connections and fired neurons.

Network builder is obsolete now network build her own.

I am uploading a copy in google code under hyacinth.

http://code.google.com/p/hyacinth/downloads/list

N-Depend

I have been busy with artificial neural networks for some time so couldn’t post much.

NDepend is a static code analyzer it inspects your code and tells dependincies metrics code analysis. I found metrics very useful.It tell if you used boxing if your namings are wrong, if you used wrong access modifier, you have too many fields.You can also write your own queries it uses cql code query language very similar to sql, You can select methods which returns int.

http://www.ndepend.com/ is the web page.

Take a look

Hyacinth v1.1

I added some missing things to hyacinth my jabber client.She can now encrypt messages .

Pollard’s Rho factoring algorithm in C#

Pollard’s rho algorithm is used to factor composite integers. Suppose you have 8633, 89 *97 are the factors.

Here is the code:

static void Main(string[] args)

        {

            Int64 a = p.PollardRho(12187823);

        }

        public Int64 customFunct(Int64 param, Int64 mod)//Custom number generator

        {

            return ((param * param + 1 )% mod);

        }

        public Int64 PollardRho(Int64 number)

        {

            Int64 a = 2, b = 2, tmp;//initail starting values

            while(true)

            {

            a = customFunct(a,number);//get first number

            b = customFunct(customFunct(b,number),number);//get second double run custom func

            tmp = GCD(Math.Abs(b-a), number);// if a =  b mod d(divisor one of the divisors of our number) then a-b is multiple of d and number is multiple of d 

            if (tmp > 1)

                break;

            }

            return tmp;

        }

        public Int64 GCD(Int64 x, Int64 y)//Greatest Common Divisor

        {

            if (x == 0)//Find it? return

                return y;

            if (y == 0)//Find it? return

                return x;

            x = x % y;// take the mod of the large val to small one if the order is wrong in the next recursive loop it will be correct

            return GCD(y, x);//Parameters changed place

        }

I put comments on the code and included usage.The code includes a custom function you may experiment with and a greatest common divisor calculator.

Reflection

This code is used to change all off  one kind of controls one property to any value in a form.Hard to understand ? Ok let me give you an ex: you want to disable all text box in a form property is enabled value is false very easy.

private static ArrayList controls = new ArrayList();

    public static Control[] LookControls(Form f,Type ctrlType,string property,object value)

    {

        foreach (Control c in f.Controls)

        {

            LookControls(c, ctrlType,property,value);

        }

        return (Control[])controls.ToArray(typeof(Control));

    }

    private static void LookControls(Control ctrl, Type ctrlType,string property,object value)

    {

        if (ctrl.GetType() == ctrlType)

        {

            controls.Add(ctrl);

            PropertyInfo p = ctrl.GetType().GetProperty(property);

            p.SetValue(ctrl, value, null);

        }

        if (ctrl.Controls != null)

        {

            foreach (Control ctrl1 in ctrl.Controls)

            {

                LookControls(ctrl1,ctrlType,property,value);

            }

        }

    }

LU Decomposition and Linear Equation Solving in c#

LU decomposition is used to write a matrix in Upper and Lower triangle forms which is used to solve linear equations.Here is my code for LU decomposition and Linear Equation Solving, it is based on some very old fortran code, i think lapack may have the original sources but they are hard to understand.I added a few comments but to really understand what is going on i can recommend using  pen and paper and doing with hand.

static void Main(string[] args)

{

    double[,] lu = new double[3, 3] { { 2, 1, -1 }, { -3, -1, 2 }, { -2, 1, 2 } };

    double[] b = new double[] { 8, -11, -3 };

    double[] x = new double[3];

    int[] indx = new int[b.Length];

    Program p = new Program();

    p.LUDecompose(ref lu, ref indx);

    p.Solve(ref b, ref x, indx, lu);

}

private void LUDecompose(ref double[,] lu, ref int[] indx)

{

    int i, imax = 0, j, k, n = lu.GetLength(0);

    double big, temp;

    double[] vv = new double[n];

    //preChecks

    if (lu.GetLength(0) != lu.GetLength(1) || lu.GetLength(0) != indx.Length)

        throw new Exception("matrix dimension problem only use square matrices");

    //for each row find the absolute value of the greatest cell and store in vv

    for (i = 0; i < n; i++)

    {

        big = 0.0;

        for (j = 0; j < n; j++)

            if ((temp = Math.Abs(lu[i, j])) > big) big = temp;

        if (big == 0.0)

        throw new Exception("singular matrix");

        vv[i] = 1.0 / big;//calculate scaling and save

    }

    //k is for colums start with the left look for the columns under the diagonal for the biggest value want to move the largest over diagonal

    for (k = 0; k < n; k++)//find the largest pivot element 

    {

        big = 0.0;

        for (i = k; i < n; i++)

        {

            temp = vv[i] * Math.Abs(lu[i, k]);

            if (temp > big)

            {

                big = temp;

                imax = i;

            }

        }

        if (k != imax)//do we need a row change 

        {

            for (j = 0; j < n; j++)// counter for the colums

            {

                temp = lu[imax, j];// change the rows

                lu[imax, j] = lu[k, j];

                lu[k, j] = temp;

            }

            vv[imax] = vv[k];

        }

        indx[k] = imax;

        for (i = k + 1; i < n; i++)

        {

            temp = lu[i, k] /= lu[k, k];//divide pilot element

            for (j = k + 1; j < n; j++)

                lu[i, j] -= temp * lu[k, j];

        }

    }

}

private void Solve(ref double[] b, ref double[] x, int[] indx, double[,] lu)

{

    if (b.Length != lu.GetLength(0) || x.Length != lu.GetLength(0))

        throw new Exception("vector dimension problem");

    int n = lu.GetLength(0);

    int i, ii = 0, ip, j;

    double sum = 0;

    for (i = 0; i < n; i++) x[i] = b[i];

    for (i = 0; i < n; i++)

    {

        ip = indx[i];

        sum = x[ip];

        x[ip] = x[i];

        if (ii != 0)

            for (j = ii - 1; j < i; j++) sum -= lu[i, j] * x[j];

        else if (sum != 0.0)

            ii = i + 1;

        x[i] = sum;

    }

    for (i = n - 1; i >= 0; i--)

    {

        sum = x[i];

        for (j = i + 1; j < n; j++) sum -= lu[i, j] * x[j];

        x[i] = sum / lu[i, i];

    }

}

The code also has an example how to use b is the vector right hand side of the equation x is the solution vector, lu is the matrix.

tricky or idiomatic

Ok  have written all this page and windows live writer stopped working when  i was posting, i hate programs when they do that (dont   get backup) i will tell about backing up in another post.I will not write all the stuff back i have things to do, please tell your best wishes to live writer group .

int x = 2;

int y = x << 1 | 1;

int z = x << 1 + 1;

ok these are not same y = 5 and z is 8 because of order of evaluation.

//x   y

//0   2   

//1   0

//2   1

int x, y = 1; ;

x = (2 - y) * (1 + 3 * y) / 2;

x = y + 1;

if (x == 3) x = 0;

switch (y)

{

    case 2:

        x = 0;

        break;

    case 0:

        x = 1;

        break;

    case 1:

        x = 2;

        break;

    default:

        {

           throw new Exception("incorrect parameter");

        }

}

ok here is 3 different codes that do the same thing : return values associated in table above code like 0 if 2 is given as parameter.

the first one is tricky and bad the second is good idiomatic and the last one is suboptimal and slow only may be used if algorith should be broken with future numbers.

Ok i know i didn’t told anything but i am sure codes will be explanatory enough.

Mixin ins in c# or code extensions

Technorati Tags: mixin,mixin in,code extension,extension,c#,interface

    public interface IExtension

    {

    }

    public static class Extensions

    {

        public static string GetTypeInfo(this IExtension ext)

        {

            return String.Format("{0} {1}", ext.GetType().FullName, ext.GetHashCode().ToString());

        }

    }

    public static class Test

    {

        public static string GetTest(this IExtension debug)

        {

            return "Test string";

        }

    }

    class Program

    {

        static void Main()

        {

            var myObj = new MyClass();

            Console.WriteLine(myObj.GetTypeInfo());

            Console.WriteLine(myObj.GetTest());

            Console.Read();

        }

    }

    class MyClass : IExtension

    {

    }

C# does not support mixins ; mixins are like prewritten codes functions to be implemented by subclasses, it is different from interfaces interfaces are only signature does not contain functions mixins does. We can use code extensions to add extra functions to a class.It is pretty easy to use code extensions like above.

Lock

Lock actually is the short for Monitor.Enter and Monitor.Exit

Here is the code

class Program

    {

        static void Main(string[] args)

        {

            for (int i = 0; i < 100; i++)

            {

                new Thread(Go).Start();

            }

            Thread.Sleep(1000);

            Console.Read();

        }

        static int val1 = 90, val2 = 10;

        static object locker = new object();

        static void Go()

        {

            lock (locker)

            {

                if (val2 != 0)

                {

                    Thread.Sleep(20);

                    Console.WriteLine(val1 / val2);

                    val2 = 0;

                }

                else

                {

                    Console.WriteLine("attempted to divide by zero");

                }

            }

            }

    }

lets look what happens when you start 100 threads of Go function starts and if you don’t lock gives a attempted to divide by zero exception.When you lock problem is solved, lets look at the il code:

.method private hidebysig static void  Go() cil managed

{

  // Code size       88 (0x58)

  .maxstack  2

  .locals init ([0] object CS$2$0000,

           [1] bool CS$4$0001)

  IL_0000:  nop

  IL_0001:  ldsfld     object ConsoleApplication12.Program::locker

  IL_0006:  dup

  IL_0007:  stloc.0

  IL_0008:  call       void [mscorlib]System.Threading.Monitor::Enter(object)

  IL_000d:  nop

  .try

  {

    IL_000e:  nop

    IL_000f:  ldsfld     int32 ConsoleApplication12.Program::val2

    IL_0014:  ldc.i4.0

    IL_0015:  ceq

    IL_0017:  stloc.1

    IL_0018:  ldloc.1

    IL_0019:  brtrue.s   IL_003e

    IL_001b:  nop

    IL_001c:  ldc.i4.s   20

    IL_001e:  call       void [mscorlib]System.Threading.Thread::Sleep(int32)

    IL_0023:  nop

    IL_0024:  ldsfld     int32 ConsoleApplication12.Program::val1

    IL_0029:  ldsfld     int32 ConsoleApplication12.Program::val2

    IL_002e:  div

    IL_002f:  call       void [mscorlib]System.Console::WriteLine(int32)

    IL_0034:  nop

    IL_0035:  ldc.i4.0

    IL_0036:  stsfld     int32 ConsoleApplication12.Program::val2

    IL_003b:  nop

    IL_003c:  br.s       IL_004b

    IL_003e:  nop

    IL_003f:  ldstr      "attempted to divide by zero"

    IL_0044:  call       void [mscorlib]System.Console::WriteLine(string)

    IL_0049:  nop

    IL_004a:  nop

    IL_004b:  nop

    IL_004c:  leave.s    IL_0056

  }  // end .try

  finally

  {

    IL_004e:  ldloc.0

    IL_004f:  call       void [mscorlib]System.Threading.Monitor::Exit(object)

    IL_0054:  nop

    IL_0055:  endfinally

  }  // end handler

  IL_0056:  nop

  IL_0057:  ret

} // end of method Program::Go

As you can see it is a try finally block and monitor.enter and monitor.exit methods.If  you replace lock with monitor codes given below il code will be same

    class Program

    {

        static void Main(string[] args)

        {

            for (int i = 0; i < 100; i++)

            {

                new Thread(Go).Start();

            }

            Thread.Sleep(1000);

            Console.Read();

        }

        static int val1 = 90, val2 = 10;

        static object locker = new object();

        static void Go()

        {

            Monitor.Enter(locker);

            try

            {

                if (val2 != 0)

                {

                    Thread.Sleep(20);

                    Console.WriteLine(val1 / val2);

                    val2 = 0;

                }

                else

                {

                    Console.WriteLine("attempted to divide by zero");

                }

            }

            finally

            {

                Monitor.Exit(locker);

            }

        }

    }

lock(this) is not a good programming example i will tell more about it another time but for now don’t use it. You must lock a reference variable (please remember structs are value and classes are reference types),