从网页检索信息

Question

My application is meant to speed up the retrieval of phone call information from our telephone system. The best way to get this information is to create a new search on the telephone system's web interface and export the results to an Excel spreadsheet which my application then imports into a DataSet.

To get the export, from the login screen, the process goes as follows:

• Log in
• Navigate to Reports Page
• Click "Extension Detail" link
• Select "Extensions" CheckBox
• Select the extensions (typically all the ones currently being used) from the listbox
• Specify date range
• Click on Export button

It's not a big job to do it manually every day, but, for reliability, it would be great if I can make my application do this automatically the first time it starts every day. Since more than 1 person in the company is going to use this application, having a Windows Service do it would be even better.

I don't know if it'll help, but the system is Datatex Topaz Next Generation telephone management system: http://www.datatex.co.za/downloads/index.html#TNG

Can anyone give me a basic idea how to do this?

Also, can anyone post links (in comments if need be) to pages where I can learn more about how to do this?

Answer 1

I have done the something similar to fetch info from a website. I cannot give you a exact answer. But the idea is to send login info to the page with form values. If the site is relying on cookies, you can use this cookie aware WebClient:

public class CookieAwareWebClient : WebClient
{
    private CookieContainer cookieContainer = new CookieContainer();

    protected override WebRequest GetWebRequest(Uri address)
    {
        WebRequest request = base.GetWebRequest(address);
        if (request is HttpWebRequest)
        {
            (request as HttpWebRequest).CookieContainer = cookieContainer;
        }
        return request;
    }
}

You should be aware that some sites rely on a session id being passed so the first thing I did was to fetch the session id from the page:

var client = new CookieAwareWebClient();
client.Encoding = Encoding.UTF8;

var indexHtml = client.DownloadString(*index page url*);

string sessionID = fetchSessionID(indexHtml);

Then I had to log in to the page which you can do by uploading values to the page. You can see the specific form elements with "view source" but you have to know a little HTML to do so.

var values = new NameValueCollection();
values.Add("sessionid", sessionID); //Fetched session id
values.Add("brugerid", args[0]); //Username in my case
values.Add("adgangskode", args[1]); //Password in my case
values.Add("login", "Login");   //The login button

//Logging in
client.UploadValues(*url to login*, values); //If all goes perfect, I'm logged in now

And then I could download the page I needed. In your case you may use DownloadFile(...) if the file always have the same url (something like Export.aspx?From=2010-10-10&To=2010-11-11) or UploadValues(...) where you specify the values as before but saves the result.

string html = client.DownloadString(*url*);

It seems you have a lot more steps than I did. But the principle is the same. To see what values your send to the site to login etc. you can use programs such as Fiddler (windows) which can capture the activity going on. Essential you just do exactly the same thing but watch out for session id etc. which is temporary.

The best idea is really to use some native way to fetch data, but if don't got the code, database etc. you have to do it the ugly way. You may also need a HTML parser to fetch the data ( ups, you don't because you export to a file ). And last but not least, keep in mind that pages can change and there is great potential to fail to login, parse etc.

Please ask for if you are uncertain what is going on.

ADDITION

The CookieAwareWebClient is not my code:

• http://code.google.com/p/gardens/source/browse/Montrics/Physical.MyPyramid/CookieAwareWebClient.cs?r=26
• Using CookieContainer with WebClient class

I also found some relevant threads:

• What's a good tool to screen-scrape with Javascript support?
• http://forums.asp.net/t/1475637.aspx

Answer 2

With a HTTP client, you need to do the following:

• Log in, using cookies or HTTP authentication
• Request a page
• Submit form data

This means that you need some class or component in your program that can do HTTP, cookies, authentication and forms. With this, you do the same requests a user would do.