#!/usr/bin/perl -w

# Author   : Benoit PAPILLAULT
# Creation : 16/01/2001

use Socket;
use IO::Handle;

sub petites_annonces_fr
{
	print localtime() . ": interrogation de www.petites-annonces.fr\n";
	socket (SOCKET, PF_INET, SOCK_STREAM, 0);
	$ipaddr = gethostbyname('www.petites-annonces.fr');
	$sin = sockaddr_in(80,$ipaddr);
	connect (SOCKET, $sin);

    $oldfh = select(SOCKET);
    $| = 1;
    select($oldfh);

	$msg = "GET /index.cfm?RequestTimeout=120&typepage=recherche&ChoixRub=504&acces_rech=1&show_result=1&sousrub1=5040103&prix_min=2000&prix_max=4000&cp=06&commune=&origine=&tri=c_date HTTP/1.0\r\n"
	. "Referer: http://www.petites-annonces.fr/index.cfm?ChoixRub=504&acces_rech=1\r\n"
    . "User-Agent: Mozilla/4.72 [en] (X11; I; Linux 2.4.0 i686)\r\n"
	. "Host: www.petites-annonces.fr\r\n"
	. "Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg , image/png, */*\r\n"
	. "Accept-Encoding: gzip\r\n"
	. "Accept-Language: fr, en\r\n"
	. "Accept-Charset: iso-8859-1,*,utf-8\r\n"
	. "Cookie: VALIDUSER=\r\n"
	. "\r\n" ;

	print SOCKET $msg;

	open FIC, ">logement-petites-annonces.fr";

    while (<SOCKET>) {
		print FIC $_;
	if (/NAME=\"an_date\" VALUE=\"(.*)\"/) {
		$publication = $1;
	}
	if (/NAME=\"accroche\" VALUE=\"(.*)\"/) {
		$accroche = $1;
	}
	if (/NAME=\"texte\" VALUE=\"(.*)\"/) {
		$texte = $1;
	}
	if (/NAME=\"cX\" VALUE=\"(.*)\"/) {
		$loyer = $1;
	}
	if (/NAME=\"cP\" VALUE=\"(.*)\"/) {
		$cP = $1;
	}
	if (/NAME=\"cC\" VALUE=\"(.*)\"/) {
		$cC = $1;
	}
	if (/NAME=\"cY\" VALUE=\"(.*)\"/) {
		$cY = $1;
	}
	if (defined($publication) && defined ($accroche) && defined($texte)
		&& defined($cP) && defined($cC) && defined ($cY) && defined($loyer)) {
		$annonce = "date: $publication, $accroche, $texte, adresse : $cP $cC $cY, loyer: $loyer, source: www.petites-annonces.fr";
		push (@annonces_courantes,$annonce);

		undef $publication;
		undef $accroche;
		undef $texte;
		undef $cP;
		undef $cC;
		undef $cY;
		undef $loyer;

#		print "++ $annonce\n";
	}
#	print $_;
    }

    close SOCKET;
	close FIC;
}

sub nicematin_fr
{
	print localtime() . ": interrogation de www.nicematin.fr\n";
	socket (SOCKET, PF_INET, SOCK_STREAM, 0);
	$ipaddr = gethostbyname('www.nicematin.fr');
	$sin = sockaddr_in(80,$ipaddr);
	connect (SOCKET, $sin);
	
	$oldfh = select(SOCKET);
	$| = 1;
	select($oldfh);
	
	$msg = "GET http://www.nicematin.fr/journal/journalnm/webPA/PA.0000121.0000129.0000294.html HTTP/1.0\r\n"
		. "User-Agent: Mozilla/4.72 [en] (X11; I; Linux 2.4.0 i686)\r\n"
		. "Pragma: no-cache\r\n"
		. "Host: www.nicematin.fr\r\n"
		. "Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, image/png, */*\r\n"
		. "Accept-Encoding: gzip\r\n"
		. "Accept-Language: fr, en\r\n"
		. "Accept-Charset: iso-8859-1,*,utf-8\r\n"
		. "\r\n";

    print SOCKET $msg;
	open FIC, ">logement-nicematin.fr";
    while (<SOCKET>) {
		print FIC $_;
		if (/puce.gif.*>(.*)<br>/) {
			$date = $1;
		}
		if (/<br>(.*)<br>/) {
			$texte = $1;
		}
		if (defined ($date) && defined($texte)) {
			$annonce = "date: $date, $texte, source: www.nicematin.fr";
			push (@annonces_courantes, $annonce);

			undef $texte;
		}
    }
    close SOCKET;
	close FIC;
}

sub bonjour_fr
{
	print localtime() . ": interrogation de www.bonjour.fr\n";
	socket (SOCKET, PF_INET, SOCK_STREAM, 0);
	$ipaddr = gethostbyname('web3.bonjour.fr');
	$sin = sockaddr_in(80,$ipaddr);
	connect (SOCKET, $sin);


	$oldfh = select(SOCKET);
	$| = 1;
	select($oldfh);

	$msg = "POST http://web3.bonjour.fr/bonjour_fichiers/AffichePa2.cfm?noeud=ILHAP000 HTTP/1.0\r\n"
		. "Referer: http://web3.bonjour.fr/bonjour_fichiers/SelectDep.cfm?pa=5&Idnoeud=ILHAP000\r\n"
		. "User-Agent: Mozilla/4.72 [en] (X11; I; Linux 2.4.0 i686)\r\n"
		. "Host: web3.bonjour.fr\r\n"
		. "Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, image/png, */*\r\n"
		. "Accept-Encoding: gzip\r\n"
		. "Accept-Language: fr, en\r\n"
		. "Accept-Charset: iso-8859-1,*,utf-8\r\n"
		. "Cookie: CFGLOBALS=HITCOUNT%3D114%23LASTVISIT%3D%7Bts+%272001%2D02%2D01+14%3A54%3A39%27%7D%23TIMECREATED%3D%7Bts+%272001%2D01%2D17+16%3A02%3A06%27%7D%23; CFID=7543430; CFMAGIC=7543430%3A80630942; CFTOKEN=80630942; RMID=c365d2d13a65b590; REFERER=bonjour; CFID=7543430; CFTOKEN=80630942\r\n"
		. "Content-type: application/x-www-form-urlencoded\r\n"
		. "Content-length: 160\r\n"
		. "\r\n"
			. "setCritereL=TypeLoc&Dept=06&Ville=&Arrondissement=&rechsur=ville&OL5CMin=2000&OL5CMax=4000&ONBPMin=2&ONBPMax=3&OSURMin=&OSURMax=&Rechercher.x=55&Rechercher.y=10";

	print SOCKET $msg;
	open FIC, ">logement-bonjour.fr";

	$resultat = "";

	while (<SOCKET>) {
#		print FIC $_;
		chomp ;
		if (/([^
]+)[
]*/) {
			$resultat .= $1;
		}
	}

	print FIC $resultat;

	while ( $resultat =~ /<BR>[ \t]*([^<]*)<BR>/g) {
		$line = $1;
		$line =~ s/[ \t]+/ /;
		$line =~ s/[ \t]*$//;
		push (@texte, $line);
	}
	while ($resultat =~ /<B>[ \t]*([^<]*)<\/B>/g) {
		$line = $1;
		$line =~ s/[ \t]+/ /;
		$line =~ s/[ \t]*$//;
		push (@lieu, $line);
	}

	for ($i=0; $i<$#lieu; $i++) {
		$annonce = "$lieu[$i] $texte[$i] source: www.bonjour.fr";
		push (@annonces_courantes, $annonce);
	}

	close SOCKET;
	close FIC;
}

sub load_base
{
	if (open FIC, "logement.txt" ) {
		print localtime() . ": chargement des annonces precedentes\n";
		while (<FIC>) {
			chomp $_;
			push (@annonces_precedentes, $_ );
		}
		close FIC;
	}
}

sub save_base
{
	rename "logement.txt","logement.old";
	open FIC, ">logement.txt";
	foreach $annonce (@annonces_courantes) {
		print FIC "$annonce\n";
	}
	close FIC;
}

sub check_new
{
	$nb_courantes = @annonces_courantes;
	$nb_precedentes = @annonces_precedentes;
	print "annonces trouvees : $nb_courantes\n";
	print "annonces precedentes : $nb_precedentes\n";
	foreach $annonce (@annonces_courantes) {
		$new = 1;
		foreach $old (@annonces_precedentes) {
			if ($old eq $annonce) {
				$new = 0;
			}
		}

		if ($new) {
			print "$new " . localtime() . ": $annonce\n";
		}
	}
}

sub main
{
	load_base();

	petites_annonces_fr();
	nicematin_fr();
	bonjour_fr();

	save_base();
	check_new();
}

main;