Viktar Karpach Web Developer Blog about ASP.NET,C#,T-SQL,Javascript,CSS,Silverlight

Save for android photoshop javascript

Sun, 17 Mar 2013 00:00:00 -0700

Due to different pixel density of different android devices developer / designer need to create different versions of graphic artifacts.

Recently I've been trying to create my first Android application and got tired of creating image artifacts for different pixel density. As a result I came up with a following photoshop javascript that automates with tedious task.

var docRef = activeDocument;    
function saveAs(path,filename,width,height,dpi) {
    savedState = docRef.activeHistoryState
    docRef.resizeImage (width, height, dpi);
    var folder = new Folder (path);
    if (!folder.exists)
    {
         folder.create();
    }
    pngFile = new File( path + filename)
    pngSaveOptions = new PNGSaveOptions()    
    app.activeDocument.saveAs(pngFile, pngSaveOptions, true, Extension.LOWERCASE)
    docRef.activeHistoryState = savedState    
}

function main(isIcon)  {
    if (documents.length == 0) {
        alert("There are no documents open.");
        return;
    }
    try {        
            Path = app.activeDocument.path;
     }
    catch (exception) {
         docRef.saveAs(File.saveDialog ("Save image",["*.psd","*.png","*.jpg"]));
    } 
    Path = app.activeDocument.path;
    var Name = app.activeDocument.name.replace(/\.[^\.]+$/, '');     
    startRulerUnits = app.preferences.rulerUnits;
    app.preferences.rulerUnits = Units.PIXELS;    
    globSavedState = docRef.activeHistoryState;
    try {
        docRef.mergeVisibleLayers();
    }
    catch (exception) {
        alert("Please select visible layer.");
        return;
    }
    if (isIcon) {
        saveAs(Path + "/drawable-xxhdpi/" , Name + ".png",144,144,480);
        saveAs(Path + "/drawable-xhdpi/"  , Name + ".png",96,96,320);
        saveAs(Path + "/drawable-hdpi/"   ,Name + ".png",72,72,240);
        saveAs(Path + "/drawable-mdpi/"  ,Name + ".png",48,48,160);
        saveAs(Path + "/drawable-ldpi/"   ,Name + ".png",36,36,120);
    }
    else {
        var dpi = 480;
        saveAs(Path + "/drawable-xxhdpi/" , Name + ".png", docRef.width * dpi / docRef.resolution, docRef.height * dpi / docRef.resolution, dpi);
        dpi = 320;
        saveAs(Path + "/drawable-xhdpi/" , Name + ".png", docRef.width * dpi / docRef.resolution,docRef.height * dpi / docRef.resolution, dpi);
        dpi = 240;
        saveAs(Path + "/drawable-hdpi/" , Name + ".png", docRef.width * dpi / docRef.resolution ,docRef.height * dpi / docRef.resolution, dpi);
        dpi = 160;
        saveAs(Path + "/drawable-mdpi/" , Name + ".png", docRef.width * dpi / docRef.resolution,docRef.height * dpi / docRef.resolution, dpi);
        dpi = 120;
        saveAs(Path + "/drawable-ldpi/" , Name + ".png", docRef.width * dpi / docRef.resolution, docRef.height * dpi / docRef.resolution, dpi);
    }
    docRef.activeHistoryState = globSavedState;   
    app.preferences.rulerUnits = startRulerUnits;
    docRef.save();            
}

main(true) generates following:

480dpi 144x144 - drawable-xxhdpi
320dpi 96x96 - drawable-xhdpi
240dpi 72x72 - drawable-hdpi
160dpi 48x48 - drawable-mdpi
120dpi 36x36 - drawable-ldpi

main(false) generates images for android in respective folders, where dimensions are proportional to selected resolution / dpi.

For example:

Resolution: 480dpi
Width:150
Height:80

Becomes:

480dpi 150x80 - drawable-xxhdpi
320dpi 100x53 - drawable-xhdpi
240dpi 75x40 - drawable-hdpi
160dpi 50x27 - drawable-mdpi
120dpi 37x20- drawable-ldpi

See Save4Android github repository for more details.

How to log a message from t-sql script?

Sat, 26 Jan 2013 00:00:00 -0800

Recently I've been working on a long running script that will be executed directly in PROD enviroment. I needed an easy way to see a progress of t-sql script execution.

I tried to use PRINT, but it doesn’t output anything until script execution is done.

You can use some log table and insert progress status messages there, but there is an easier way:

RAISERROR ('Message', 10, 1) WITH NOWAIT

RAISERROR with severity of 10 or less doesn't trigger try catch block and can be used for log purposes. Here is a little longer example:

DECLARE @i INT = 0

WHILE 1=1
BEGIN                                
      
    IF @i >= 10
                BREAK;    
    WAITFOR DELAY '00:00:01';                                    
    SET @i = @i + 1        
    RAISERROR ('%d sec running', 10, 1, @i) WITH NOWAIT
END

Results:

How to get both inserted source table id and target table id?

Sat, 05 Jan 2013 00:00:00 -0800

Lets say we have two tables source and target. We want to copy records from the source to the target table and save mappings between old and new ids in a third table.

DECLARE @source TABLE
(
   id INT IDENTITY(5,1) NOT NULL PRIMARY KEY,
   name VARCHAR(max)  
)

DECLARE @target TABLE
(
   id INT IDENTITY(1,1) NOT NULL PRIMARY KEY,
   name VARCHAR(max)
)

DECLARE @inserted TABLE
(
   id INT NOT NULL,
   clone_id INT NOT NULL  
)

Lets populate the source table:

INSERT INTO @subscription
(name)
VALUES
('Test 1'),
('Test 2'),
('Test 3')

Now we can populate the target table with the following code:

INSERT INTO @target
        (name )
SELECT name + ' - Cloned' FROM @source

You can try to populate inserted table using the following code:

INSERT INTO @target
        (name )
OUTPUT s.id, INSERTED.id INTO @inserted
SELECT name + ' - Cloned' FROM @source s

But you will get a syntax error:

The multi-part identifier "s.id" could not be bound.

This is due to a nature of the INSERT statement. The SELECT statement considered to be apart from the INSERT statement, so you can't reference its tables in the OUTPUT clause. However you can achieve desired outcome using MERGE sql statement, which is a part of SQL standard from 2003, but rarely used by anybody.

MERGE @target t
USING @source s
ON 0=1
WHEN NOT MATCHED THEN 
INSERT (name) VALUES (s.name + ' - Cloned')
OUTPUT s.id, INSERTED.id INTO @inserted;

SELECT * FROM @source
SELECT * FROM @target
SELECT * FROM @inserted

Results:

id	name
5	Test 1
6	Test 2
7	Test 3

id	name
1	Test 1 - Cloned
2	Test 2 - Cloned
3	Test 3 - Cloned

id	clone_id
5	1
6	2
7	3

Firebug vs Google Chrome Developer Tools

Mon, 06 Aug 2012 00:00:00 -0700

Recently I’ve been trying to use more and more Chrome Developer Tools instead of Firebug. Here are couple of things that I thought are show stoppers, but appeared to be a lack of knowledge on my side.

There are no AJAX requests in console window. There is a setting for this:

Secondly I can’t search all javascript files:

Actually you can search all javascript files and more. Just press Ctrl+Shift+F:

Do you miss any FireBug functionality in Chrome Developer Tools?

LINQ deferred loading using C# yield-return

Sat, 31 Mar 2012 00:00:00 -0700

Yield keyword exists in C# language since version 2 along with Visual Studio 2005, however it is rarely used by developers. Recently I read Demystifying the C# Yield-Return Mechanism article about yield-return keyword usage. This article gave three most common usage scenarios, but author (James McCaffrey) didn't mention LINQ. I think LINQ is where yield-return becomes most useful since it provides mechanism for deferred loading. Here is a quick example.

Lets say we have bunch of unit tests to run. It takes some time to run a single test, lets say 1 sec. For simplicity lets assume that even tests pass and odd fail:

public class Test

    public int Id { get; set; }

    public bool Assert()

        System.Threading.Thread.Sleep(1000);

        Console.WriteLine(string.Format("Test {0} was processed.", Id));

        return Id % 2 == 0;

Lets write couple LINQ like extension methods for Test collections. Process1 doesn't use yield return. Process2 uses yield return and PrintPass output "Test # is passed" message.

public static class Extensions

    public static IEnumerable<Test> Process1(this IEnumerable<Test> tests)

        var result = new List<Test>();

        foreach (var test in tests)

            if (test.Assert())

                result.Add(test);

        return result;

    public static IEnumerable<Test> Process2(this IEnumerable<Test> tests)

        foreach (var test in tests)

            if (test.Assert())

                yield return test;

    public static void PrintPass(this IEnumerable<Test> tests)

        foreach (var t in tests)

            Console.WriteLine(string.Format("Test {0} is passed.", t.Id));

Lets fill up test collection and compare output for Process1 and Process2 method:

class Program

    static void Main()

        var tests = new List<Test>(new []

                            new Test {Id = 1},

                            new Test {Id = 2},

                            new Test {Id = 3},

                            new Test {Id = 4},

                            new Test {Id = 5},

});

        Console.WriteLine("**** No yield return *******");

        tests.Process1().PrintPass();

        Console.WriteLine("**** Using yield return *******");

        tests.Process2().PrintPass();

        Console.ReadKey();

Output:

**** No yield return *******
Test 1 was processed.
Test 2 was processed.
Test 3 was processed.
Test 4 was processed.
Test 5 was processed.
Test 2 is passed.
Test 4 is passed.

**** Using yield return *******
Test 1 was processed.
Test 2 was processed.
Test 2 is passed.
Test 3 was processed.
Test 4 was processed.
Test 4 is passed.
Test 5 was processed.

As you can see, in first case whole collection was built before PrintPass method output anything. In second case you can see PrintPass prints "Test 2 is passed." right after Assert method for this test have done its processing.

How to use PIVOT in T-SQL queries?

Tue, 27 Mar 2012 00:00:00 -0700

Web has a lot of PIVOT examples, however all of them are based on fact tables. Regular relational database rarely has facts tables. Lets see how you can use PIVOT statement with regular database. First of all lets create sample Grocery database:

Here is a script that can create this database:

CREATE DATABASE Grocery

GO

USE Grocery

CREATE TABLE Products

  ProductId INT IDENTITY(1,1) PRIMARY KEY,

  ProductName VARCHAR(50),

  Price MONEY

CREATE TABLE Stores

    StoreId INT IDENTITY(1,1) PRIMARY KEY,

    StoreName VARCHAR(50)

CREATE TABLE Orders

    OrderId INT IDENTITY(1,1) PRIMARY KEY,

    StoreId INT,

    OrderDate DATE,

    CONSTRAINT FK_Orders_Store FOREIGN KEY (StoreId) REFERENCES Stores(StoreId)

CREATE TABLE OrderProducts

    OrderId INT,

    ProductId INT,

    Qty INT,

    CONSTRAINT PK_OrderId_ProductId PRIMARY KEY (OrderId,ProductId),

    CONSTRAINT FK_OrderProducts_Products FOREIGN KEY (ProductId) REFERENCES Products(ProductId),

    CONSTRAINT FK_OrderProducts_Orders FOREIGN KEY (OrderId) REFERENCES Orders(OrderId)

Lets populate those tables:

INSERT INTO Products

(ProductName, Price)

VALUES

('Milk', 2.99),

('Bread',1.99),

('Tomato',0.99),

('Grape',1.99)

INSERT INTO Stores

(StoreName )

VALUES

('Jewel'),

('Dominicks'),

('Walmart')

INSERT INTO Orders

(StoreId, OrderDate )

VALUES

(1,'1/11/2011'),

(1,'1/14/2011'),

(2,'2/05/2011'),

(3,'3/17/2011'),

(3,'3/29/2011'),

(3,'4/02/2011')

INSERT INTO OrderProducts

( OrderId, ProductId, Qty )

VALUES

(1,2,2),

(1,1,1),

(2,3,1),

(3,4,3),

(3,1,2),

(3,3,1),

(4,3,2),

(4,4,2),

(5,1,1),

(5,3,1),

(6,1,1),

(6,2,3),

(6,4,1)

Products:

ProductId	ProductName	Price
1	Milk	2.99
2	Bread	1.99
3	Tomato	0.99
4	Grape	1.99

Stores:

StoreId	StoreName
1	Jewel
2	Dominicks
3	Walmart

Orders:

OrderId	StoreId	OrderDate
1	1	2011-01-11
2	1	2011-01-14
3	2	2011-02-05
4	3	2011-03-17
5	3	2011-03-29
6	3	2011-04-02

OrderProducts:

OrderId	ProductId	Qty
1	1	1
1	2	2
2	3	1
3	1	2
3	3	1
3	4	3
4	3	2
4	4	2
5	1	1
5	3	1
6	1	1
6	2	3
6	4	1

Now our sample relation database is ready for some pivoting. Lets say we want to know what products are sold by what store. Lets pull needed data using regular query:

SELECT

    s.StoreName,p.ProductName, op.Qty * p.Price AS Totals

FROM

    Products p INNER JOIN

    OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

    Orders o ON op.OrderId = o.OrderId INNER JOIN

    Stores s ON o.StoreId = s.StoreId

Results:

StoreName	ProductName	Totals
Jewel	Milk	2.99
Jewel	Bread	3.98
Jewel	Tomato	0.99
Dominicks	Milk	5.98
Dominicks	Tomato	0.99
Dominicks	Grape	5.97
Walmart	Tomato	1.98
Walmart	Grape	3.98
Walmart	Milk	2.99
Walmart	Tomato	0.99
Walmart	Milk	2.99
Walmart	Bread	5.97
Walmart	Grape	1.99

Those result rows look like a fact table. Lets apply PIVOT to them:

SELECT * FROM

    SELECT

        s.StoreName,p.ProductName, op.Qty * p.Price AS Totals

    FROM

        Products p INNER JOIN

        OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

        Orders o ON op.OrderId = o.OrderId INNER JOIN

        Stores s ON o.StoreId = s.StoreId

) AS Source

PIVOT

    SUM(Totals)

    FOR ProductName IN ([Bread],[Milk],[Tomato],[Grape])

) AS p

PIVOT results:

StoreName	Bread	Milk	Tomato	Grape
Dominicks	NULL	5.98	0.99	5.97
Jewel	3.98	2.99	0.99	NULL
Walmart	5.97	5.98	2.97	5.97

Lets do another one: products sales by months.

SELECT

    p.ProductName, DATENAME(m,o.OrderDate) AS [Month], op.Qty * p.Price AS Totals

FROM

    Products p INNER JOIN

    OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

    Orders o ON op.OrderId = o.OrderId INNER JOIN

    Stores s ON o.StoreId = s.StoreId

Results:

ProductName	Month	Totals
Milk	January	2.99
Bread	January	3.98
Tomato	January	0.99
Milk	February	5.98
Tomato	February	0.99
Grape	February	5.97
Tomato	March	1.98
Grape	March	3.98
Milk	March	2.99
Tomato	March	0.99
Milk	April	2.99
Bread	April	5.97
Grape	April	1.99

PIVOT version:

SELECT * FROM

    SELECT

        p.ProductName, DATENAME(m,o.OrderDate) AS [Month], op.Qty * p.Price AS Totals

    FROM

        Products p INNER JOIN

        OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

        Orders o ON op.OrderId = o.OrderId INNER JOIN

        Stores s ON o.StoreId = s.StoreId

) AS Source

PIVOT

    SUM(Totals)

    FOR [Month] IN ([January],[February],[March],[April])

) AS p

PIVOT results:

ProductName	January	February	March	April
Bread	3.98	NULL	NULL	5.97
Grape	NULL	5.97	3.98	1.99
Milk	2.99	5.98	2.99	2.99
Tomato	0.99	0.99	2.97	NULL

And last one lets do stores sales by months:

SELECT

    s.StoreName,DATENAME(m,o.OrderDate) AS [Month], op.Qty * p.Price AS Totals

FROM

    Products p INNER JOIN

    OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

    Orders o ON op.OrderId = o.OrderId INNER JOIN

    Stores s ON o.StoreId = s.StoreId

Results:

StoreName	Month	Totals
Jewel	January	2.99
Jewel	January	3.98
Jewel	January	0.99
Dominicks	February	5.98
Dominicks	February	0.99
Dominicks	February	5.97
Walmart	March	1.98
Walmart	March	3.98
Walmart	March	2.99
Walmart	March	0.99
Walmart	April	2.99
Walmart	April	5.97
Walmart	April	1.99

PIVOT version:

SELECT * FROM

    SELECT

        s.StoreName,DATENAME(m,o.OrderDate) AS [Month], op.Qty * p.Price AS Totals

    FROM

        Products p INNER JOIN

        OrderProducts op ON p.ProductId = op.ProductId INNER JOIN

        Orders o ON op.OrderId = o.OrderId INNER JOIN

        Stores s ON o.StoreId = s.StoreId

) AS Source

PIVOT

    SUM(Totals)

    FOR [Month] IN ([January],[February],[March],[April])

) AS p

PIVOT results:

StoreName	January	February	March	April
Dominicks	NULL	12.94	NULL	NULL
Jewel	7.96	NULL	NULL	NULL
Walmart	NULL	NULL	9.94	10.95

How to list sql server table dependencies using T-SQL?

Sun, 19 Feb 2012 00:00:00 -0800

You can always use SQL Server Management Studio, just right click on a table in object explorer and select View Dependencies.

However this way doesn't give you ability to copy dependencies to clipboard or any other way to export dependencies list. Luckily you can find the same data using SysObjects and SysDepends tables. Here is a quick T-SQL snippet that you might want to use for this purpose.

SELECT DISTINCT dobj.name,dobj.type

FROM SysObjects obj INNER JOIN

SysDepends d ON obj.id = d.depid INNER JOIN

SysObjects dobj ON d.id = dobj.id

WHERE obj.name = 'WorkOrder' ORDER BY dobj.type,dobj.name

How to measure stored procedure execution time?

Wed, 15 Feb 2012 00:00:00 -0800

You can use SET STATISTICS TIME ON, but it would return you timing for all single queries inside of your stored procedure. If you have bunch of queries inside then it might be not really convenient. Here is a little code snippet alternative to STATISTICS TIME:

Declare @d datetime

Set @d = CURRENT_TIMESTAMP

-- Your stored procedure call goes here

SELECT DATEDIFF(ms,@d,CURRENT_TIMESTAMP)

Use DBCC DROPCLEANBUFFERS to clear sql server cache for cache independent time measurement.

Get and Set operators in Javascript

Sat, 04 Feb 2012 00:00:00 -0800

I am C# developer, hence I use C# properties a lot. However I just recently discovered that Javascript has properties as well. Get and Set operators supported by Firefox 2.0+, Safari 3.0+, Chrome, Opera 9.5+. However IE still doesn't support properties. May be Microsoft would introduce support in IE10, but so far no luck. Here is a little example of Get and Set declaration:

var obj = {

   get CustomProperty() {

     return 0;

},

   set CustomProperty(s) {

     alert ("It is readonly.");

};

Properties can be useful and scary at the same time. A simple variable assignment is not predictable anymore. For example with a little bit of work you can convert your AJAX calls to something like this:

ajax.url = "script.php";

alert(ajax.result);

YSlow and ASP.NET: 100 points "A" grade year 2012

Sun, 29 Jan 2012 00:00:00 -0800

Almost three years ago I wrote an original YSlow and ASP.NET: 100 points "A" grade is possible article. Since then I changed couple hosting providers, made numerous updates and converted my blog project to .NET 4.0. Today I decided to revise my blog to see if it still has 100 yslow points with final "A" grade.

First surprise, as of January 29, 2012, YSlow addon doesn't work in FireFox 9. However YSlow is now available for Google Chrome, Opera, Safari. It has bookmarket version and even command line version. I installed Google Chrome Yslow extension and began investigation. Right out of the box most of sections still had A grade. About a week ago I switched from Google App Engine "CDN" to true CDN Amazon CloudFront. Some files were not in CDN, so I had to take care of those. In general I highly recommend Amazon CloudFront. It is easy to setup and cost just a few bucks a month based on your traffic usage. Amazon CloudFront mimics your local file settings, so you need to make sure that you have compression, expire header and Etags taken care of. I use IIS7 compression instead of a compression module recommended by me in my previous article. Out of the box IIS7 compression doesn't work with AWS CloudFront, so you need to modify C:\Windows\System32\inetsrv\config\applicationHost.config file on your server to include following:

<serverRuntime enabled="true" frequentHitThreshold="1" frequentHitTimePeriod="00:00:20"/>

<httpCompression directory="%SystemDrive%\inetpub\temp\IIS Temporary Compressed Files" noCompressionForHttp10="False" noCompressionForProxies="False">

  <scheme name="gzip" dll="%Windir%\system32\inetsrv\gzip.dll" />

  <dynamicTypes>

    <add mimeType="text/*" enabled="true" />

    <add mimeType="message/*" enabled="true" />

    <add mimeType="application/x-javascript" enabled="true" />

    <add mimeType="*/*" enabled="false" />

  </dynamicTypes>

  <staticTypes>

    <add mimeType="text/*" enabled="true" />

    <add mimeType="message/*" enabled="true" />

    <add mimeType="application/javascript" enabled="true" />

    <add mimeType="*/*" enabled="false" />

  </staticTypes>

</httpCompression>

Also I modified my cache module to have ETags clean up:

public void Init(HttpApplication context)

...

    context.PostReleaseRequestState += new EventHandler(application_PostReleaseRequestState);

void application_PostReleaseRequestState(object sender, EventArgs e)

    HttpContext.Current.Response.Headers.Remove("Server");

    HttpContext.Current.Response.Headers.Remove("X-AspNet-Version");

    HttpContext.Current.Response.Headers.Remove("ETag");

YSlow added new section "Use Cookie-free Domains". My blog completely failed this one, since google analytics left cookies in top level karpach.com domain. I modified google analytics init script to have: _gaq.push(['_setDomainName', 'www.karpach.com']), so google cookies set for www.karpach.com instead of top level karpach.com. Then subdomain cdn.karpach.com, which points to Amazon CloudFront, became cookieless. Finally I am back to 100 YSlow points. Run Yslow on no externals version of karpach.com to see it yourself.

How to check if a javascript variable is undefined?

Thu, 12 Jan 2012 00:00:00 -0800

How to check if a javascript variable is undefined? This is actually a tricky question. There are three cases:

A variable is declared using var keyword, but were never assigned any value.
A variable is declared as function parameter, but when function were invoked parameter was not supplied.
A variable is never declared or assigned a value and you are trying to access it.

First two cases lead to undefined value of variable. Here is an example of first case:

var v; // Value not defined

if (v) {

    console.log("v=" + v);

else {

    console.log("v has undefined value");

v = "some value";

if (v) {

    console.log("v = " + v);

Output:
v has undefined value
v = some value

The second case is similar and you can use if statement as in above code or:

function Add(p1, p2) {

    p1 = p1 || 0; // set value to 0 if value is undefined

    p2 = p2 || 0; // set value to 0 if value is undefined

    return p1 + p2; // would always work

console.log(Add(1, 2));

console.log(Add(1));

console.log(Add());

Output:
3
1
0

The third case when a variable is never declared. If you try to use such variable you get exception. If statement or default value assignment won't help.

if(typeof neverDeclared == "undefined") {

    console.log("There is something really wrong!");

You would ask, what kind of sick people try to use a variable that was never declared or assigned a value? However I bet this is one of the most frequent javascript exceptions. Yes, simple variables are usually declared, but exception usually happens when you try to pass complex object with undeclared property. Javascript objects can be easily extended with additional properties and this comes at price of frequent undefined variable exceptions. Let me give you an example:

var obj = { FirstName: "Viktar", LastName: "Karpach" };

console.log(obj.FirstName);

console.log(obj.LastName);

console.log(obj.City); // Will give exception

var obj = { FirstName: "Viktar", LastName: "Karpach" };

obj.City = "Chicago";

console.log(obj.City); // Works!

if (typeof obj.State == "undefined") {

    console.log("State is undefined.");

Undefined variable exception very often happens when you try to consume objects received from third party API or from AJAX requests, always be careful what you consume from partially unreliable resource.

AutoHotkey DVDFab script for volume label conversion

Sun, 25 Dec 2011 00:00:00 -0800

From time to time I use free DVDFab HD Decrypter to rip rented movies, so I can watch them later using my Popcorn Hour A210 media streamer. DVDFab is kind enough to pull volume label from DVD disk. However this label has all uppercase letters and words are concatenated by underscores. For example:

THE_LONG_GOODBYE

I rename those labels to something like this: The Long Goodbye (1973), where year I lookup on IMDB web site. I don't like tedious repetitive tasks, so one day I decided to write a script for AutoHotkey to automate this process.

It is easy to replace underscores and correct letter case, but year should be retrieved from some kind of database. Google pointed me in a direction of IMDB API.

It might be a little bit challenging to do IMDB API access using just AutoHotkey script engine. Luckily AutoHotkey scripts can load dll libraries and with magic of COM.ahk and CLR.ahk you can load even .NET libraries. However those magic scripts can't properly execute .NET 3.5/4.0 libraries, so I had to use .NET 2.0. As you probably know .NET 2.0 doesn't have native support for JSON deserialization, so I had to use Newtonsoft.Json for this purpose. I also could use XML output from IMDB API, but I prefer to use JSON, since response is smaller in a size. Here is my code for .NET part of it:

public class Server

    public string GetMovieTitleWithYear(string volumeLablel)

        var request = WebRequest.Create(new Uri(string.Format("http://www.imdbapi.com/?t={0}",

            volumeLablel.Replace('_', '+'))));

        request.Method = "GET";

try

            var response = request.GetResponse();

            var responseStream = response.GetResponseStream();

            if (responseStream != null)

                var reader = new StreamReader(responseStream);

                var responseString = reader.ReadToEnd();

                var movie = JsonConvert.DeserializeObject<Movie>(responseString);

                if (movie != null && !string.IsNullOrEmpty(movie.Title))

                    return string.Format("{0} ({1})", movie.Title, movie.Year);

        catch

        return UppercaseWords(volumeLablel.Replace('_', ' ').ToLower());

    static string UppercaseWords(string value)

        char[] array = value.ToCharArray();

        // Handle the first letter in the string.

        if (array.Length >= 1)

            if (char.IsLower(array[0]))

                array[0] = char.ToUpper(array[0]);

        // Scan through the letters, checking for spaces.

        // ... Uppercase the lowercase letters following spaces.

        for (int i = 1; i < array.Length; i++)

            if (array[i - 1] == ' ')

                if (char.IsLower(array[i]))

                    array[i] = char.ToUpper(array[i]);

        return new string(array);

Then using JSON 2 C# Service you get following:

public class Movie

    public string Title { get; set; }

    public int Year { get; set; }

    public string Rated { get; set; }

    public string Released { get; set; }

    public string Genre { get; set; }

    public string Director { get; set; }

    public string Writer { get; set; }

    public string Actors { get; set; }

    public string Plot { get; set; }

    public string Poster { get; set; }

    public string Runtime { get; set; }

    public float Rating { get; set; }

    public int Votes { get; set; }

    public string ID { get; set; }

    public string Response { get; set; }

Library is ready. Lets plug it into AutoHotkey:

#include c:\Program Files (x86)\Autohotkey\Extras\Scripts\CLR.ahk
#include c:\Program Files (x86)\Autohotkey\Extras\Scripts\COM.ahk

RemoveToolTip:
SetTimer, RemoveToolTip, Off
ToolTip
return

#IfWinActive ahk_class QWidget
~Lbutton::
MouseGetPos, xpos,ypos
if (xpos>306 and ypos>474 and xpos<605 and ypos<495)
{
	ToolTip, Double click in text box below`nto get IMDB title and year,305,430
	SetTimer, RemoveToolTip, 5000
}
If (A_PriorHotKey = A_ThisHotKey and A_TimeSincePriorHotkey < 500) 
{
    CLR_Start()
	imdb := CLR_LoadLibrary("c:\Program Files (x86)\Autohotkey\Extras\Scripts\Karpach.IMDB.dll")
	;Type names must be fully qualified.
	server:= CLR_CreateObject(imdb,"Karpach.IMDB.Server")
	Send ^a
	Sleep 100
	clipboard =   
	Send ^c
	ClipWait
	title:=COM_Invoke(server,"GetMovieTitleWithYear",clipboard)
	COM_Release(imdb)
	COM_Release(server)
	clipboard = %title%
	Send ^v
	return  
}
return

#IfWinActive

Now whenever you double click in volume label text box content would be converted into proper movie title.

AutoHotkey scripts and DLLs

IMDB API .NET library source code

Update:

If you are using AutoHotkey_L then you don't need COM.ahk and you can call GetMovieTitleWithYear method directly:

title:= server.GetMovieTitleWithYear(clipboard)

AutoHotkey_L scripts and DLLs

Second Update:

It is better to use new url for imdb api: http://www.omdbapi.com/

Simple factory pattern based on interfaces

Mon, 05 Sep 2011 00:00:00 -0700

The main purpose of factory design patterns is to separate object creation process from current implementation context. During coding you use factory to create objects. Based on incoming parameters factory decides which object to create. You don't know exact type of the object. Usually you know either base abstract class type that created object was inherited from or interface that created object is implementing. Based on this knowledge you can use object (call abstract class methods or interface methods). I found a lot of factories examples with abstract class connection, but none with interface connection. However interface connection is most common in commercial factory design pattern implementations. Many developers use Unity Framework or Castle Project (example of connection based on interface). Below I'll show how to implement factory with interface connection like it is done in Unity Framework or Castle Project.

Lets build animal fabric, which would create dogs and cats objects. Let's define ICat and IDog interfaces.

public interface ICat

    string Meow();

public interface IDog

    string Bark();

    string Sit();

Now, lets define some Dog and Cat classes that implement ICat and IDog interfaces respectevly.

public class Cat:ICat

    public string Meow()

        return "Meow meow meow ...";

public class Dog: IDog

    public string Bark()

        return "Woof woof woof ...";

    public string Sit()

        return "I am sitting.";

Now lets create factory:

public class DefaultFactory

    public static T Create<T>()

        if (typeof(T) == typeof(IDog))

            return (T)(IDog)new Dog();

        else

            if (typeof(T) == typeof(ICat))

                return (T)(ICat)new Cat();

            else

                throw new NotImplementedException(String.Format("Creation of {0} interface is not supported yet.", typeof(T)));

And here how you can use it:

IDog dog = DefaultFactory.Create<IDog>();

ICat cat = DefaultFactory.Create<ICat>();

Console.WriteLine(dog.Bark());

Console.WriteLine(dog.Sit());

Console.WriteLine(cat.Meow());

Output:

Woof woof woof ...
I am sitting.
Meow meow meow ...

Source code for sample project above.

Video Links in a New Media Player Window

Thu, 12 May 2011 00:00:00 -0700

Google Chrome extension popups Youtube, IMDB, Vimeo and Twitvid links in new window with just a video player. Video player window can be resized to scale video. Works with online chats that post links dynamically, for example CampFire from 37 Signals.

Each video gets its own window, so you can open as many videos as you like. Cntrl-Click to open link regular way.

Please leave comments and suggestions.

Install Google Chrome Extension

After you install extension try to click on sample links below:

How to find column usage in stored procedure?

Sun, 08 May 2011 00:00:00 -0700

Sometimes you want to rename or delete column from a table in MS SQL Server. How to find column name dependency usage in MS SQL Server? There is no clean way of doing this. In management studio you can right click on object and select "View Dependency", but this would show only object dependency. So you need to write some T-SQL code in order to find dependency. Lets say we need to find dependecy of Title column from the Article table.

SELECT DISTINCT so.name

FROM sys.syscomments sc

INNER JOIN sys.objects so on sc.id=so.object_id

WHERE sc.text LIKE '%Title%'