Monday, 2 July 2012

Upload very large files - Part 2

Earlier in this blog I talked about the general idea behind one solution for uploading very large files (I consider very large when they are above 2GB), how the file API is becoming more popular and the possible client side approaches. This time I’ll describe the server side part of the solution. As a quick reminder, the server is responsible for:

  • To determine if the request is to be handled by the mechanism, if not, just let it go.
  • To determine if the request is an initial packet or a data packet.
  • To handle properly each type of packet giving the appropriate response so the client can continue.

This is part of a bit more complex solution, named FreshUplaod, which has the Server component and client components using either JavaScript or Silverlight. The core element is the UploadHelper class, as the name indicates is a helper, it encapsulates the dirty logic regarding the request data reading and processing all stuff.

The methods ProcessInitialPacket and ProcessDataPacket do the actual work, but it’s necessary to know if the request data corresponds to one or another, that’s the job for these other methods IsInitialPacket and IsDataPacket. The actual helper utility is another method, ProcessRequest that acts as a façade to the whole mechanism. An example on how to use it is as follows:

[HttpPost]
public ActionResult Upload(FormCollection collection)
{
    var packetSize = 4 * 1024 * 1024; // default to 4 MB
    var filePath = Server.MapPath("~/_temp_upload/");

    var result = UploadHelper.ProcessRequest(
        Request, filePath, packetSize);

    if (result != null)
        return Json(result);
    return Content("");
}
 

This is using a controller action on ASP.NET MVC 3, where the access to the request is quite easy and the conversion to Json is stress-free too. The key points here are the packet size and the folder where the file is to be stored, the packet size must be exactly the same configured on client side in order to expect a good behavior, I’ve used the 4 MB value because this is the default constraint imposed by ASP.NET and of course this can be tweaked as you want in web.config file.

Now let’s see how to use this element on ASP.NET WebForms, the idea I suggest here is to use a handler whether it was a Generic Handler or ASP.NET Handler, it can be also a WebForm but is not elegant enough solution because the WebForms are chained in a different pipeline and there can be a bit of noise, although it work as well. Here’s the code sample:

public void ProcessRequest(HttpContext context)
{
    var packetSize = 4 * 1024 * 1024; // default to 4 MB
    var filePath = context.Server.MapPath("~/_temp_upload/");

    var result = UploadHelper.ProcessRequest(
        context.Request.RequestContext.HttpContext.Request, 
        filePath, packetSize);

    if (result != null)
    {
        JsonResult(context.Response, result);
    }
    else
    {
        context.Response.Write("");
    }
}

private void JsonResult(HttpResponse response, object result)
{
    response.ContentType = "application/json";
    var serializer = new JavaScriptSerializer();
    var json = serializer.Serialize(result);
    response.Write(json);
}
 
 

Here we have a little more hand work but not so much, the json conversion and output is done completely manual (if some WebForms people knows how to achieve this with fewer code, please let me know cause I’m a little out of this part of ASP.NET, although I support Scott Hanselman with his slogan that “Only One ASP.NET”). The other key point is how to get the Request but with an instance of HttpRequestBase class, the default context.Response is an HttpRequest instance, so after some minutes of try-error-retry I found this way to get right value, again if some knows a better way to do it, you know.

In the next posts I’ll cover the other parts of this solution. The whole source code can be found at https://bitbucket.org/abelperezok/freshupload, I am still preparing the code samples so in a few weeks they will be ready.

Wednesday, 20 June 2012

Making easy asynchronous web requests

In web development it’s very frequent to communicate between applications via web requests, may be as client in Restful interfaces or APIs, or just simply send information to some server from any client technology. There are many ways to achieve this task in .NET Framework, the most popular probably is the HttpWebRequest class, which support either synchronous or asynchronous methods. The synchronous way is the simplest, but it has as consequence the thread-blocking problem and CPU time wastage and if we use it in some user interface it could get the so hated Not Responding state. If we don’t make a proper use of multi-threading aware, this is not a good approach.
The asynchronous tactic could be used instead, but it’s a little hard-to-try at glance, the same code I do in a few lines synchronous becomes at least three methods and a lot more lines of code. Just go to MSDN and search for HttpWebRequest.BeginGetRequestStream(...) and you’ll see the sample code like this.
var request = WebRequest.CreateHttp("http://www.example.com/data/to/request");
request.Method = "POST";
request.BeginGetRequestStream(AsyncCallbackRequest, request);
(...)

private void AsyncCallbackRequest(IAsyncResult ar) 
{
    HttpWebRequest request = (HttpWebRequest)ar.AsyncState;
    Stream stream = request.EndGetRequestStream(ar);
 (...) 
 // manipulate the stream: write data or wahtever
    request.BeginGetResponse(GetResponseCallback, request);
}

private void GetResponseCallback(IAsyncResult ar) 
{
    HttpWebRequest request = (HttpWebRequest)ar.AsyncState;
    HttpWebResponse response = (HttpWebResponse)request.EndGetResponse(ar);
    Stream streamResponse = response.GetResponseStream();
 (...) 
 // manipulate the response: read data and parse it 
}
In some scenarios for example in a Silverlight application, we have not available the whole .NET features and unless we use a third party library or another weird or abstract solution, this is the key for sending or receiving HTTP data across the network. If the application must do several types of request involving different formats or data types, then the amount of asynchronous-logic methods might be larger than the actual business-logic for our application and that’s not good for anyone, maintenance is compromised.
So I’ve done a small class that encapsulates all this garbage repetitive code, at the time I fire some custom events which will be listened outside of the class and the real logic to be performed in those event listeners, the code will be clear and focused on the real tasks to be done on it. I show the code here:
public class AsyncWebRequest
{
    private readonly string _uri;

    public AsyncWebRequest(string uri)
    {
        _uri = uri;
    }

    public void Start()
    {
        var request = WebRequest.CreateHttp(_uri);
        request.Method = "POST";

        var args = new PrepareDataEventArgs { Request = request };
        if (PrepareData != null)
        {
            PrepareData(this, args);
        }

        request.BeginGetRequestStream(AsyncCallbackRequest, request);
    }

    public event EventHandler<PrepareDataEventArgs> PrepareData;
    public event EventHandler<SendDataEventArgs> SendRequest;
    public event EventHandler<ReceiveDataEventArgs> ReceiveResponse; 

    private void AsyncCallbackRequest(IAsyncResult ar)
    {
        HttpWebRequest request = (HttpWebRequest)ar.AsyncState;

        var args = new SendDataEventArgs();

        // End the operation
        if (SendRequest != null)
        {
            SendRequest(this, args);
        }            

        // only if there's data or data is not empty, write the actual data
        if (args.Data != null && args.Data.Length > 0)
        {
            Stream stream = request.EndGetRequestStream(ar);

            // Write to the request stream.
            stream.Write(args.Data, 0, args.Length);

            // Close the stream
            stream.Close();
        }
        // Start the asynchronous operation to get the response
        request.BeginGetResponse(GetResponseCallback, request);
    }

    private void GetResponseCallback(IAsyncResult ar)
    {
        HttpWebRequest request = (HttpWebRequest)ar.AsyncState;

        // End the operation
        HttpWebResponse response = (HttpWebResponse)request.EndGetResponse(ar);

        Stream streamResponse = response.GetResponseStream();
        StreamReader streamRead = new StreamReader(streamResponse);
        string responseString = streamRead.ReadToEnd();

        if (ReceiveResponse != null)
        {
            ReceiveResponse(this, new ReceiveDataEventArgs{ Data = responseString });
        }

        // Close the stream object
        streamResponse.Close();
        streamRead.Close();

        // Release the HttpWebResponse
        response.Close();
        //Synchronization.Set();
    }
}
This class fires three events, first the PrepareData is fired before the request is modified for getting the stream, this is a constrain imposed by .NET because after this state the request cannot be modified, so this is the event used to set all the header information such as content/type, among others. After having got the request stream and before start getting the response the second event (SendRequest) is fired and here’s the place to set the actual data to send if any, the length must be set too because the buffer might be larger than the actual data to send. Finally, right after the data arrival, ReceiveResponse, the third event, allows us to set a listener in order to process the received data, for example to parse a JSON data. An example on how to use it is shown as follows:
var asweb = new AsyncWebRequest(_postDataurl);
asweb.PrepareData += (o, args) => 
{
    args.Request.ContentLength = bytes.Length;
    args.Request.ContentType = "application/x-www-form-urlencoded";
};

asweb.SendRequest += (o, args) =>
{
    args.Data = bytes;
    args.Length = bytes.Length;
};

asweb.ReceiveResponse += (o, args) =>
{
    var json = JsonValue.Parse(args.Data);
    (...)
};

asweb.Start();
I hope this tiny class will help you to deal with web request taking advantage of asynchrony.

Tuesday, 1 May 2012

Upload very large files - Part 1

In my last post I talked about settings we have to set server side on IIS in order to allow large uploads, and I also mentioned the security risks involved. The solution I’ll talk this time is becoming more common as HTML5 is being implemented more and more.

The main problem was related to the size of the post when uploading, but why not split file into several small pieces and upload piece by piece? Maybe some time ago was crazy to think about manipulating the file in client side from JavaScript, the only ways were using Flash or Silverlight, now the file API is more popular, almost all modern browsers implement it but IE which is always retarded implementing good stuff.

I’ve found some implementations such as plupload or http://hertzen.com/experiments/jsupload/ but I’m not 100 % satisfied with it, I took the ideas and hands on. In this first part I’ll explain how it works to achieve a bullet proof mechanism, for example: The connection might be interrupted or something can happen on the server and the worker thread was shot down. The mechanism must be protected against these events.

I talked to a friend of mine and after a session of brainstorming I decided to imitate some kind of TCP connection in regard to sending packet and receiving the ACK (Acknowledgement), this stuff. The idea is implemented as follows:

  • The browser opens the file dialog and a file is selected, the initial packet is prepared with file name, the type (MIME) and the size.
  • In the server a new GUID is generated for this file that will be the file id for the whole process.
  • A new file is created with this id that will be the metadata file, where the information from the client will be stored for further use.
  • A new blank file is created with this id and the extension of the original name that will be the data file.
  • The reply for this initial packet is a JSON with the GUID generated.

The following steps in client side are:

  • Read the file fragment according to the current packet number and send to the server the data packet with the id, the current packet number and the total packets.
  • Each data packet sent receives its ACK which indicates the number of the last packet received OK, after that the client will send the last received + 1 unless this is the last one.

The server process the data packet:

  • If the metadata file or actual file doesn’t exist, the result is an ACK with packet -1, which means the file must be uploaded from scratch.
  • If the file doesn’t have the required size: the packet number – 1 multiplied by packet size, then return the required value in order to the client knows which is the correct fragment to resend.
  • If everything goes well, a success ACK with the current packet number is sent encouraging to the client to send the next packet.

I’ve implemented this using HTML5 techniques but not all the browsers are capable for to do it, Modernizr could be a great helper on this, since there’s no HTML5 detection but feature detection on demand, for now it’s necessary the browser to be capable of use the file API and if you want to support retry after refresh the page, it should be capable of access to local Storage which allows to store data and retrieve it when necessary. The browser will able to use the XMLHttpRequest object for upload files.

I haven’t implemented yet any fallback mechanism, it will be nice to have a third party such as Silverlight which could substitute the missing browser HTML5 features, a good example of implementation using Silverlight can be found at HSS.Interlink. they cover all the aspects on read a fragment of file and send the request to the server.

My implementation I am planning to distribute it via nuget.org containing a js script and a dll with the classes to be called from any handler/controller without to have to deal with all the verifications mentioned above.

In next posts I’ll cover the implementation of this solution and I’ll provide a code sample for to understand it better.

Saturday, 14 April 2012

Magic settings for upload large files under IIS and .NET

I remember when I started this blog; I talked about upload large files using .NET and IIS. This time I’ll share with you some interesting points to be aware of when trying to upload large files. I consider large when the size is bigger than 2GB, because this is the actual limit is almost all web servers under Microsoft technology.
Based on my experience with this kind of situations, I can say that the web environment is not ready to transfer that much information. Remember the beginnings of WWW, when everything that roamed through the wires were just text and small graphics, I say more, remember what HTTP stands for (Hypertext Transfer Protocol) and what is HyperText?
However, the file transfer to and from the server is something very common nowadays, therefore, it’s necessary to know how to handle it. For those who doesn’t need to upload files greater than 500 MB, I recommend the solution I presented through the my firsts posts in this blog, even for upload a bit bigger files, but when the size start over the 1 GB it’s better to think about another solution.
The larger sizes you could upload the bigger security risk you have. I explain, solutions like NeatUpload and many others (open source or not) expect to make the upload in a big single request, although there are modules that make the copy block by block in order to be sure of not making big memory peaks, the web server must allow a very large request, exposing it to a DOS attack, just an example. Another possible problem may be if the client doesn’t have a fast connection (like me) there’s a (high) risk that a large file might be interrupted, and then? Start over again. On the server side the risk of execution timed out exists too (I’ve faced that), so the value of this setting must be increased.
If you wish to upload small/medium files there’s no problem using this kind of solution, but if you want go over the size, I suggest another kind of workaround.
Alongside my research for the best solution I’ve found the following links I hope to be useful to you, I let MSDN articles, Stack Overflow questions, among others very interesting sources written by the most involved people in this matter, where one can drink the honeys of knowledge.
As a quick conclusion, I can say the following in order to configure properly your application:
The default value for the request size in ASP.NET 2.0 and later is 4 MB. Under IIS6 the setting is configured by the httpRuntime node, example:
 <configuration> 
  <system.web>
   <httpRuntime maxRequestLength="102400" executionTimeout="3600" />
  </system.web>
 </configuration> 
Important to say here, the value of maxRequestLength is given in KB and the executionTimeout is given in seconds.
Under IIS7+ the setting is configured by the requestLimits node, example:
 <system.webServer>
  <security>
   <requestFiltering>
    <requestLimits maxAllowedContentLength="1024000000" />
   </requestFiltering>
  </security>
 </system.webServer>
In this case the value of maxAllowedContentLength is given in bytes. This must combined with the applicationHost.config file usually located at (C:\Windows\System32\inetsrv\Configs) whose section should look like:
 <section name="requestFiltering" overrideModeDefault="Allow" />

As I said before, I hope this compilation saves you time of google-research, in a further post I’ll talk about another solution that I consider that more fits better in web environment.

Friday, 30 March 2012

Language selector in ASP.NET MVC 3 - Part 3

In the last post I talked about implementing a custom route in order to improve the route handling work to support multiple languages in a web application. This custom route would remove the need of duplicate the route (with and without the language parameter) and in the first part of this series I showed an example where the view rendered the language selector using ActionLink for generating the urls, and the collection to iterate was expected to be as part of the view model.

But this is not always possible to do, in fact, the most common scenarios I’ve dealt with, involve putting this kind of component in a common place such as the Layout page or even a partial which is included in Layout page, but less frequent in a dedicated view. To achieve this separation, it’s common to use a Child Action for invoke a controller, which is going to execute an action and return usually a partial view.

If we take the solutions previously discussed and make a little refactoring: create an action (LanguageList) in the languages controller and move the code that renders the links to a partial view (_langSelector) and go to layout page and invoke it as:

@Html.Action("LanguageList", "Langs")

Sure you are going to see the results. The controller action should be like this:

[ChildActionOnly]
public ActionResult LanguageList()
{
    var langs = LanguageProviders.Current.GetList();
    return PartialView("_langSelector", langs);
}

The HTML rendered looks like this:

<a href="/Langs/LanguageList">English</a> 
<a href="/es-ES/Langs/LanguageList">Espa&#241;ol</a> 
<a href="/fr-FR/Langs/LanguageList">Fran&#231;ais</a> 

As you can see, the urls always point to controller Langs and action LanguageList instead of pointing to current controller and action, why does it happen?

When we invoke a child action, a new request is started to the target action, that means, when the route asks for current request values, it obtains those from the new request, which means, the “original” route values were lost at the moment of making the child request.

As these values were lost, my solution is simply to provide them via parameter at invocation time. Once in the controller, the values must be passed to the view, may be using the ViewBag or even a more sophisticated solution could involve the creation of a view model with these data ready to be read by the view.

Having said that, here is the quick solution:

The controller action

[ChildActionOnly]
public ActionResult LanguageList()
{
    ViewBag.CurrentValues = RouteData.Values["CurrentValues"];
    var langs = LanguageProviders.Current.GetList();
    return PartialView("_langSelector", langs);
}

At this point and only to remark, the attribute ChildActionOnly is strongly recommended in order to avoid this action to be called directly or even from an AJAX request if this is not intended to.

The invoke statement in the layout page.

@Html.Action("LanguageList", "Langs", 
 new { CurrentValues = ViewContext.RouteData.Values }
)

The goal of this new parameter is to pass the real current route values to the controller, so it will be able to grab them and pass to the view.

The partial view

@{
    var currentValues = (RouteValueDictionary)ViewBag.CurrentValues;
    var tempValues = new RouteValueDictionary();
    foreach (var item in currentValues)
    {
        tempValues.Add(item.Key, item.Value);
    }
}

@foreach (var lang in Model)
{
    tempValues["lang"] = lang.Code;
    
    @Html.ActionLink(lang.Name, null, tempValues)
    @Html.Raw(" ")
}

At this point, I consider important the lines that make a copy (clone) of the route data received from the controller, otherwise the route values should be modified inside the loop that follows it, and as a consequence, the dictionary will affect the rest of routes in the page, setting the last value in lang.Code forever in the route values if this value were not supplied explicitly when generating the links.

Well, I hope this little sand grain will help you to your day-to-day web programming task, any comment and ideas will be welcome.

Tuesday, 20 March 2012

Language selector in ASP.NET MVC 3 - Part 2

A few weeks ago I talked about giving support to multi lingual applications or i18n for brevity. There I mentioned the strategy I followed using two routes (one including the language parameter and another is just the default) and also I promised to talk more about routes, specially about routes whose goal is match with or without the language parameter. Well, actually I wasn’t too happy with that implementation, but if it isn’t broken, don’t fix it. However, something deep in my mind suspected there must be another solution, I mean, more elegant, while making google research I found an interesting article in codeproject part 1 and part 2. Nice try! But my question was: why to make a custom route class and finally to define the two different routes again at application startup?
I took it as starting point and worked on it, the result: a custom route which is able to deal with the dirty stuff about prepending properly the language fragment and matching from the request url.
The idea is to inherit from the Route class instead of RouteBase, I want to take advantage of the base implementation which is responsible of doing the really dirty stuff, I won’t reinvent the wheel. So the methods to override are GetRouteData and GetVirtualPath.
The custom route must know the default language, it has it as parameter, some comments on GetRouteData method. First, the goal for this method is to determine if the current request must be parsed by us or let it go. Manually split by / and remove the ~ element, after that if there are more than one segment and the first segment doesn’t match the regular expression ^[a-z]{2}(-[A-Z]{2})?$ (commented in the earlier post), then ask to base implementation and if it replies affirmative, set the route value indexed by language parameter to the default language. Otherwise, a little trick with the base implementation, save the current url and prepend {lang}/ to the current and ask to base implementation to check if is valid and finally restore the url value. As you can see, here is when I’ve done the work of two routes in one!
public override RouteData GetRouteData(HttpContextBase httpContext)
{
    var requestedURL = httpContext.Request.AppRelativeCurrentExecutionFilePath;
    var segments = requestedURL.Split('/').Where(x => x != "~").ToArray();

    // if request is not localized then call base
    if (segments.Length > 0 && !Regex.IsMatch(segments[0], @"^[a-z]{2}(-[A-Z]{2})?$"))
    {
        var result = base.GetRouteData(httpContext);
        if (result != null)
            result.Values[LangParam] = _defaultLang;
        return result;
    }

    // if request is localized then prepend the culture segment to the url
    var currentUrl = Url;
    Url = string.Format("{{" + LangParam + "}}/{0}", _url);
    var baseRoute = base.GetRouteData(httpContext);
    Url = currentUrl;
    // save and restore the current URL
    return baseRoute;
}
The next step to complete the route’s functionalities is GetVirtualPath which will generate the url according to route map data. Generate the url is a little harder than match the url, but not impossible. Let’s remember how the url generation occurs: if the dictionary received as parameter has all the necessary values for this route, the optional values may be taken from the current request (this will give us the idea of virtual directories commented in an earlier post). So here the important parameter is the language, find it either in current request or explicit values and remove it (the actual route doesn’t have this paramter). Call base implementation and restore each case if necessary (this is important, or the next routes generated since this will have altered values). Finally to prepend language segment if necessary and only if its value is different from the default language. I show here the code.
public override VirtualPathData GetVirtualPath(RequestContext requestContext, RouteValueDictionary values)
{
    string lang = "", lnRequest = "", lnValues = "";
    // look for the language param in current request context
    if (requestContext.RouteData.Values.ContainsKey(LangParam))
    {
        // if found then save it in lang variable
        lnRequest = lang = (string)requestContext.RouteData.Values[LangParam];
        // and remove it
        requestContext.RouteData.Values.Remove(LangParam);
    }

    // look for the language param in explicit values 
    if (values.ContainsKey(LangParam))
    {
        lnValues = lang = (string)values[LangParam];
        values.Remove(LangParam);
    }

    // call base method...
    var virtualPath = base.GetVirtualPath(requestContext, values);

    // restore from current request context if necessary
    if (!string.IsNullOrWhiteSpace(lnRequest))
        requestContext.RouteData.Values.Add(LangParam, lnRequest);
    // restore from explicit values if necessary
    if (!string.IsNullOrWhiteSpace(lnValues))
        values.Add(LangParam, lnRequest);

    if (virtualPath == null) return null;

    // prepend language segment if necessary and only if different from the default language
    if (!string.IsNullOrWhiteSpace(lang) && !string.Equals(lang, _defaultLang, StringComparison.OrdinalIgnoreCase))
    {
        virtualPath.VirtualPath = lang + "/" + virtualPath.VirtualPath;
    }
    return virtualPath;
}
Now, it’s time to invoke this brand-new route just implemented, I think it should be as similar as possible to use as the out of the box route, like this:
routes.AddLocalizedRoute(
    "es-ES",                                                                        //The default Language 
    "{controller}/{action}/{id}",                                                   // URL with parameters (Without any sign of i18n)
    new { controller = "SimpleUser", action = "Index", id = UrlParameter.Optional } // Parameter defaults
);
In order to achieve this I’ve implemented an extension for the class RouteCollection, like the original MapRoute which is an extension too.

Tuesday, 13 March 2012

Handling Excel spreadsheets XLS and XLSX with dot NET

More than once in my projects the format for input and/or output is Excel spreadsheet either using Excel 97-2003 (.xls) or Excel 2007 OOXML (.xlsx). The open source community contains several libraries, for legacy files the most famous is NPOI, which in fact is the .NET version of POI Java project at http://poi.apache.org/. The result about modern OOXML were more diverse, ExcelPackage was the first but this project was abandoned late 2010. However this code wasn’t to the trash, EPPlus is the evolved version based on it and strongly optimized by the new team.
If we want to use indiscriminately the two formats, it’s necessary to have updated the two libraries and since they are developed by two different teams, they have two completely different interfaces and ways to be used. Despite the fact it’s open source and the hard work is done, it’s a mess to deal with this difference in our code.
Start writing if-else statement alongside the code is not an option! This will get us to Spaghetti code anti-pattern and the consequent difficulties for maintaining the resulting software. So the design patterns to the rescue. The quick solution for the immediate needs, it’s necessary a common interface for basic operation with Excel files and we have two different implementations. The adapter pattern (aka wrapper) is the best match. The interfaces created are:
public interface IExcelHandler : IDisposable
{
    void LoadExelFile(string fileName);
    IExcelSheet GetSheet(int index);
    void Save();
    void Save(string filename);
}

public interface IExcelSheet
{
    string GetCellValue(int row, string column);
    void SetCellValue(int row, string column, object value);
}
 
The xml comments were removed for brevity but the methods’ names are self-explanatory (or at least I think so), the only thing might be confusing is the column parameter which is defined as string because it’s intended to be invoked with the actual Excel column name, such as A, B, etc. The operations defined above were enough in order to solve my needs; I know this might be extended for more complex scenarios but this is a first version and I don’t like to make over engineering or as we say in my town: killing mosquitos with cannons.
Having the common interface and all the implementations is not enough for a reusable component and abstract the client from the concrete implementation, how is to be created the instances of each implementation according to the file format? Design patterns again, the Factory Method may be one solution, in fact I made a simplified version using an interface that defines a Create method who is the responsible for create the concrete IExcelHandler instance.
public interface IExcelHandlerFactory
{
    IExcelHandler Create(string filename);
}
 
This interface might be injected via your favorite IoC Container and linked to the default implementation I provide in the class ExcelHandlerFactory or if you aren’t using dependency injection it can be used as a Singleton through the static property Instance.
As NPOI and EPPlus are already packages in the NuGet Gallery, I’ve done a package with these codes in order to contribute to the community and to be easier to bootstrap when handling Excel files. The package is available on NuGet feed at https://nuget.org/packages/ExcelHandler, just execute the following command at NuGet Package Manager Console:
Install-Package ExcelHandler
This will add automatically as dependencies the NPOI and EPPlus packages with all the corresponding references in the project. An example on how to use it is as follows:
Console.WriteLine("Opening and modifying a file with Excel 2007+ format...");
using (var excelHandler = ExcelHandlerFactory.Instance.Create(@"..\..\modern_file.xlsx"))
{
    var sheet = excelHandler.GetSheet(1);
    Console.WriteLine(sheet.GetCellValue(1, "A"));
    sheet.SetCellValue(1, "A", "Test value in A1 cell");
    excelHandler.Save();
}
Console.WriteLine("... done.");
Console.WriteLine("Opening and modifying a file with Excel 97-2003 format...");
using (var excelHandler = ExcelHandlerFactory.Instance.Create(@"..\..\legacy_file.xls"))
{
    var sheet = excelHandler.GetSheet(1);
    Console.WriteLine(sheet.GetCellValue(1, "A"));
    sheet.SetCellValue(1, "A", "Test value in A1 cell");
    excelHandler.Save();
}
Console.WriteLine("... done."); 
 
This example just opens two Excel files, read the content from the cell A1 and set some text in the same cell. Since the interface IExcelHandler inherit from IDisposable it can be used in a using statement in order to simplifying the resource disposing. The source code is available at https://bitbucket.org/abelperezok/excelhandler.