/// <summary>
/// Creates a new instance of the <see cref="Controller"/> class.
/// </summary>
public Controller()
{
globals = Globals.Instance();
log = new QueueEventLogger(100);
crawler = null;
stats = new long[10];
proxy = CrawlWaveServerProxy.Instance(globals);
}
private Globals globals; //Provides access to the global variables and application settings
#endregion
#region Constructor and Singleton Instance Members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private HostRequestFilter()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the storage for the HostRequestFilterEntry objects
hostTable = new Dictionary <string, HostRequestFilterEntry>(128);
//Get a reference to the global variables and application settings
globals = Globals.Instance();
}
private Globals globals; //Provides access to the global variables and application settings
#endregion
#region Constructor and Singleton Instance Members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private HostBanFilter()
{
//Initialize the storage for the banned host entries
hostTable = new Hashtable();
//Get a reference to the global variables and application settings
globals = Globals.Instance();
//Initialize the list of banned hosts
//proxy = WebServiceProxy.Instance();
proxy = CrawlWaveServerProxy.Instance(globals);
InitializeBannedHosts();
}
private const string supportedContentType = "application/x-shockwave-flash"; //The Content Type supported by the parser
#endregion
#region Constructor and Singleton Instance members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private SwfParser()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the Encoding
encoding = Encoding.UTF8; //GetEncoding("ISO-8859-7");
//Initialize the converters and parsers
converter = new CSwf2HtmlConverterClass();
parser = HtmlParser.Instance();
//Get a reference to the global variables and application settings
globals = Globals.Instance();
}
//The alternative content type description (PDF has 2 IANA reserved content-types)
#endregion
#region Constructor and Singleton Instance members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private PdfParser()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the encoding
encoding = Encoding.GetEncoding("ISO-8859-7");
//Initialize the converters and parsers
converter = new XpdfTextClass();
parser = TextParser.Instance();
//Get a reference to the global variables and application settings
globals = Globals.Instance();
}
private Encoding encoding = System.Text.Encoding.UTF8; //Needed to parse the robots.txt files
#endregion
#region Constructor and Singleton Instance Members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private RobotsFilter()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the storage for the RobotsTxtEntry objects
robotsTable = new Dictionary <string, RobotsTxtEntry>(1024);
//Get a reference to the global variables and application settings
globals = Globals.Instance();
//Initialize the various strings. Interning them saves us a little memory.
userAgent = new string [] { String.Intern("User-agent: "), String.Intern("User-agent: *"), String.Intern("User-agent: CrawlWave") };
disallow = String.Intern("Disallow: ");
FileName = String.Intern(globals.AppDataPath + "RobotsCache.xml");
}
//The Content Type supported by the parser
#endregion
#region Constructor and Singleton Instance members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private TextParser()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the Regular Expressions
hrefRegex = new Regex(@"(http|https)://[\w]+(\.[\w]+)+([\w\-\.,@?^=%&:/~\+#]*[\w\-\@?^=%&/~\+#])?", RegexOptions.CultureInvariant | RegexOptions.Multiline | RegexOptions.IgnoreCase | RegexOptions.Compiled);
//use "(http|ftp|https)://[\w]+(\.[\w]+)+([\w\-\.,@?^=%&:/~\+#]*[\w\-\@?^=%&/~\+#])?" to enable ftp urls
sessionIDRegex = new Regex(@"([0-9a-fA-F]{40,64})|([\{|\(]?[0-9a-fA-F]{8}[-]?([0-9a-fA-F]{4}[-]?){3}[0-9a-fA-F]{12}[\)|\}]?)$", RegexOptions.CultureInvariant | RegexOptions.Multiline | RegexOptions.IgnoreCase | RegexOptions.Compiled); //@"^([0-9a-f]{32})|(\{?[0-9a-f]{8}-([0-9a-f]{4}-){3}-[0-9a-f]{12}\}?)$"
spacesRegex = new Regex(@"\s+", RegexOptions.CultureInvariant | RegexOptions.Multiline | RegexOptions.IgnoreCase | RegexOptions.Compiled);
//Initialize the filters
robotsFilter = RobotsFilter.Instance();
domainFilter = DomainFilter.Instance();
//Get a reference to the global variables and application settings
globals = Globals.Instance();
}
private Globals globals; //Provides access to the global variables and application settings
#endregion
#region Constructor and Singleton Instance Members
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private DomainFilter()
{
//Initialize the synchronization mechanism
mutex = new Mutex();
//Initialize the storage for the IP Addresses
ipTable = new IPCountryTable(16); //keyLength of 16 will create 65536 root nodes
//Initialize the various strings. Interning them saves us a little memory.
FileNames = new string [] { String.Intern("apnic.latest"), String.Intern("arin.latest"), String.Intern("lacnic.latest"), String.Intern("ripencc.latest"), };
//initialize the regular expression
ipAddressRegex = new Regex(@"^(?:(?:25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)(?(\.?\d)\.)){4}$", RegexOptions.CultureInvariant | RegexOptions.Multiline | RegexOptions.IgnoreCase | RegexOptions.Compiled);
//For IPv6 addresses the following pattern can be used:
// ^(([\dA-Fa-f]{1,4}:){7}[\dA-Fa-f]{1,4})(:([\d]{1,3}.){3}[\d]{1,3})?$
// and the input length must be between 16 and 39 characters
//Get a reference to the global variables and application settings
globals = Globals.Instance();
//Load the IP Address tables into the storage
LoadIPAddresses();
}
/// <summary>
/// The constructor is private so that only the class itself can create an instance.
/// </summary>
private Client()
{
globals = Globals.Instance();
}