art.dambrine
/
robot-web-website
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
Browse Source

site-commit

master
art.dambrine 5 years ago
commit
8fd7dfcff4
14 changed files with 677 additions and 0 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 1
      .gitignore
  2. 11
      bdd.php.dist
  3. 4
      css/font-awesome.min.css
  4. 70
      css/style.css
  5. 39
      entete.php
  6. 31
      get_avancement.php
  7. 12
      head.php
  8. 45
      index.php
  9. 40
      pieddepage.php
  10. 45
      provenance.php
  11. 48
      recent.php
  12. 201
      robot.php
  13. 54
      site_info.php
  14. 76
      top.php

1
.gitignore
View File

@ -0,0 +1 @@
bdd.php

11
bdd.php.dist
View File

@ -0,0 +1,11 @@
<?php
// connexion a la base
try {
$bdd = new PDO('mysql:host=<ip>;dbname=<dbname>','<USER>','<PWD>', array(PDO::ATTR_ERRMODE => PDO::ERRMODE_EXCEPTION));
}
catch(PDOException $e)
{
echo " Erreur de connexion, contactez l'administrateur du site !";
}

4
css/font-awesome.min.css
View File

File diff suppressed because one or more lines are too long

70
css/style.css
View File

@ -0,0 +1,70 @@
form {
text-align: center;
/* display: inline-block; */
}
form .form-example {
margin-bottom: 20px;
}
form .form-example input {
text-align: center;
}
#number {
text-align: center;
width: 44px;
}
#site {
font-size: 18px;
line-height: 140%;
}
#sites-recents{
font-size: 18px;
}
#URL {
width: 434px;
/*width: 70vw;*/
}
#progressbar {
width:400px;
height: 40px;
}
#td-padding {
padding-left: 15px;
text-align: center;
}
.table-head-top-50{
font-weight: bold;
}
@media (max-width: 700px) {
#progressbar {
width: 70vw;
}
#URL {
width: 80vw;
}
#site {
font-size: 15px;
line-height: 120%;
}
.content-inside {
padding-left: 25px; /* -20 pour le tableau, faire une div + 20 pour le texte */
padding-bottom: 150px;
}
.text-padding{
padding-left: 20px;
}
}

39
entete.php
View File

@ -0,0 +1,39 @@
<!-- En tête de la page -->
<header id="header-desktop">
<div class="wrapper" id="entete">
<h1>Robot Index<span class="green">.</span></h1>
<nav>
<ul>
<li><a href="/">Robot</a></li>
<li><a href="/recent.php">Sites</a></li>
<li><a href="/top.php">Top 50</a></li>
</ul>
</nav>
</div>
</header>
<header id="header-smartphone">
<div id="entete">
<!-- Top Navigation Menu Smartphone -->
<div class="topnav">
<h1 id="nav-title">Robot Index<span class="green">.</span></h1>
<div id="myLinks">
<ul>
<li><a href="/">Robot</a></li>
<li><a href="/recent.php">Sites</a></li>
<li><a href="/top.php">Top 50</a></li>
</ul>
</div>
<a href="javascript:void(0);" class="ham-icon" onclick="myFunction()">
<img id="ham-ico" src="https://art-dambrine.ovh/learning/css/hamburger.png">
<img id="cross-ico" src="https://art-dambrine.ovh/learning/css/cross.png">
</a>
</div>
</div>
</header>

31
get_avancement.php
View File

@ -0,0 +1,31 @@
<?php
if (isset($_GET['port'])){
$port = $_GET['port'];
$port = (int) $port;
if ($port>8090) die();
$curl = curl_init();
curl_setopt_array($curl, array(
CURLOPT_URL => "http://art-dambrine.ovh:".$port,
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => "",
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => 0,
CURLOPT_FOLLOWLOCATION => true,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
CURLOPT_CUSTOMREQUEST => "GET",
CURLOPT_HTTPHEADER => array(
"Content-Type: application/json"
),
));
$response = curl_exec($curl);
curl_close($curl);
echo $response;
}

12
head.php
View File

@ -0,0 +1,12 @@
<!DOCTYPE html>
<head>
<meta charset="utf-8">
<link rel="stylesheet" href="https://art-dambrine.ovh/learning/css/font-awesome.min.css">
<link rel="stylesheet" href="https://art-dambrine.ovh/learning/css/bootstrap.min.css">
<link rel="stylesheet" href="https://art-dambrine.ovh/learning/css/styles.css">
<link rel="stylesheet" href="css/style.css">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Robot Index</title>
</head>

45
index.php
View File

@ -0,0 +1,45 @@
<?php /* Projet learning 2019 */
include("head.php");
include("entete.php");
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<h2>Robot Index</h2>
<p>
Notre robot teste et indexe les liens du site de votre choix <br> <br>
Commencez par copier l'adresse d'un site depuis la barre d'adresse de votre navigateur.
</p>
<br/>
<form action="robot.php" method="POST" class="form-example">
<div class="form-example">
<label for="name"><strong>URL</strong></label>
<input type="url" name="URL" id="URL" placeholder="Collez ici l'URL" required>
</div>
<div class="form-example">
<label for="number">Nombre de pages à parcourir </label>
<input type="number" name="number" id="number" value="30" min ="1" max="50" required>
</div>
<div class="form-example">
<input type="submit" value="Lancez le Robot !">
</div>
</form>
<br>
<h3><a style="color: #0C62A6;" href="top.php">Voir le Top 50 des liens indexés</a></h3>
<h3><a style="color: #0C62A6;" href="recent.php">Voir les résultats des sites testés récemment</a></h3>
</div>
</div>
<?php include("pieddepage.php");

40
pieddepage.php
View File

@ -0,0 +1,40 @@
<!-- Pied de la page -->
<footer id="footer">
<a href="#"><img class="gotopbtn" src="https://art-dambrine.ovh/learning/css/up-arrow.png" title="up-arrow" alt="up-arrow"></a>
<div class="container">
<div class="copyright">Copyright © 2019<span class="green">.</span> Arthur Dambrine & Seth Bio Yara<span class="green">.</span></div>
</div>
</footer>
<script>
function myFunction() {
var x = document.getElementById("myLinks");
if (x.style.display === "block") {
x.style.display = "none";
} else {
x.style.display = "block";
}
var y = document.getElementById("header-smartphone");
var title = document.getElementById("nav-title");
if (x.style.display === "block") {
y.style.margin = "0px 0px 165px 0px"; /* Aggrandissement du header à l'apparition du menu */
title.style.padding = "14px 16px 0px 16px";
} else {
y.style.margin = "0px 0px 0px 0px";
title.style.padding = "14px 16px";
}
var cross = document.getElementById("cross-ico");
var ham = document.getElementById("ham-ico");
if (x.style.display === "block") {
cross.style.display = "block";
ham.style.display = "none";
} else {
cross.style.display = "none";
ham.style.display = "block";
}
}
</script>

45
provenance.php
View File

@ -0,0 +1,45 @@
<?php /* Projet learning 2019 */
ini_set('display_errors', 1);
error_reporting(~0);
include("head.php");
include("entete.php");
include("bdd.php");
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<h1> <strong>Sites référençant le site</strong> <?php echo $_GET["link"]; ?><h1>
<h3>testés par <strong>Robot Index</strong></h3>
<table class="tab-result">
<tbody>
<?php
//phpinfo();
$requete = $bdd -> prepare('
SELECT site_id FROM robot.LINK WHERE link = ?;
');
$requete->execute(array($_GET["link"]));
while ($donnes = $requete->fetch()){
echo '<tr><th><a style="color: #0C62A6;" id="sites-recents" href="site_info.php?url='.$donnes['site_id'].' ">' . parse_url($donnes['site_id'], PHP_URL_HOST). '</a></th></tr>';
}
// site_info.php?url=http://www.arngren.net
?>
</tbody>
</table>
</div>
</div>
<?php include("pieddepage.php");

48
recent.php
View File

@ -0,0 +1,48 @@
<?php /* Projet learning 2019 */
ini_set('display_errors', 1);
error_reporting(~0);
include("head.php");
include("entete.php");
include("bdd.php");
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<h1>Sites testés il y a peu</h1>
<h3>par <strong>Robot Index</strong></h3>
<p> <strong>Cliquez sur les liens</strong> pour afficher le résultat des tests<p>
<table class="tab-result">
<tbody>
<?php
//phpinfo();
$requete = $bdd -> prepare('
SELECT url, creation_date as "most_recent"
FROM robot.SITE
WHERE length(url) > 2 AND url NOT like "%cdn%" AND url like "%.%" AND url NOT like "%:%:%"
ORDER BY creation_date DESC
LIMIT 20;
');
$requete->execute();
while ($donnes = $requete->fetch()){
echo '<tr><th><a style="color: #0C62A6;" id="sites-recents" href="site_info.php?url='.$donnes['url'].' ">' . parse_url($donnes['url'], PHP_URL_HOST). '</a></th></tr>';
}
// site_info.php?url=http://www.arngren.net
?>
</tbody>
</table>
</div>
</div>
<?php include("pieddepage.php");

201
robot.php
View File

@ -0,0 +1,201 @@
<?php
/* Defining a PHP Function */
function availablePort($min_port, $max_port) {
for($port = $min_port; $port <= $max_port; $port++){
// create curl resource
$ch = curl_init();
// set url ---> /!\ IMPORTANT Laisser en http (nécessaire)
curl_setopt($ch, CURLOPT_URL, "http://art-dambrine.ovh:".$port);
//return the transfer as a string
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
// $output contains the output string
$output = curl_exec($ch);
// close curl resource to free up system resources
curl_close($ch);
//echo $output;
if($output == null){
//echo $port.' is unused port </br>';
return $port;
} else {
echo 'Robot is busy.</br>';
if($port == $max_port)
return false;
}
}
}
//function isWorkingSite($url)
function isSiteAvailible($url){
// Check, if a valid url is provided
if(!filter_var($url, FILTER_VALIDATE_URL)){
return false;
}
// Initialize cURL
$curlInit = curl_init($url);
// Set options
curl_setopt($curlInit,CURLOPT_CONNECTTIMEOUT,10);
curl_setopt($curlInit,CURLOPT_HEADER,true);
curl_setopt($curlInit,CURLOPT_NOBODY,true);
curl_setopt($curlInit,CURLOPT_RETURNTRANSFER,true);
// Get response
$response = curl_exec($curlInit);
// Close a cURL session
curl_close($curlInit);
return $response?true:false;
}
/* Projet learning 2019 */
include("head.php");
include("entete.php");
$regex_url = "%^(?:(?:https?|ftp)://)(?:\S+(?::\S*)?@|\d{1,3}(?:\.\d{1,3}){3}|(?:(?:[a-z\d\x{00a1}-\x{ffff}]+-?)*[a-z\d\x{00a1}-\x{ffff}]+)(?:\.(?:[a-z\d\x{00a1}-\x{ffff}]+-?)*[a-z\d\x{00a1}-\x{ffff}]+)*(?:\.[a-z\x{00a1}-\x{ffff}]{2,6}))(?::\d+)?(?:[^\s]*)?$%iu" ;
$url = $_POST['URL'];
$urlIsVerified = false;
$clean_url = $url;
if (preg_match($regex_url, $url))
{
$scheme = parse_url($url, PHP_URL_SCHEME);
$domain = parse_url($url, PHP_URL_HOST);
$clean_url = $scheme."://".$domain;
$urlIsVerified = true;
}
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<h2>Robot Index</h2> <br>
<?php
$iterations = $_POST['number'];
if ($iterations >= 50){
$iterations = 50;
}
echo'<div id="site"><p>';
if ($launchJava = isSiteAvailible($_POST['URL'])){
echo $clean_url;
} else {
echo '<p>Désolé ' . $url .' n\'est pas une url valide pour notre robot <br>';
echo 'Pouvez vous utiliser une url copiée directement
<br> depuis la barre d\'adresse du navigateur ? :) </p>';
}
echo'</p></div>';
echo '<p>';
echo '</br>';
echo 'Pages à visiter : '. $iterations;
echo '</br>';
// Tester si $_POST['URL'] est une URL valide
// Checkin what port to use for the JAVA Robot WebServer
$available_port = availablePort(8085, 8088);
echo 'Notification port : <strong id="port">' . $available_port .'</strong><br>';
$java_shell_exec = "cd /var/www/java &&
nohup java -cp mysql-connector-java.jar:RobotIndex.jar robotindex.RobotIndex " .$clean_url. " " . $iterations . " " .$available_port." &";
// code JAVA exec
if($launchJava && $urlIsVerified && $available_port!=false){
//echo 'Launch JAVA cmd :' . $java_shell_exec;
echo '</br>';
// to execute a command in background <3
pclose(popen($java_shell_exec, 'r'));
} else {
echo 'Essayer une url differente.';
}
echo '</p>';
?>
<br/>
<h4>Dialog-box: </h4>
<p id="dialogbox"></p>
<div class="text-center" style="margin-left: auto; margin-right:auto;">
<progress id="progressbar" value="0" max="100"></progress>
</div>
</div>
</div>
<script>
// Boîte de dialogue pour afficher l'avancement
function loopNotif()
{
//Continuez le traitement après la pause dans cette fonction
var myHeaders = new Headers();
myHeaders.append("Content-Type", "application/json");
var requestOptions = {
method: 'GET',
headers: myHeaders,
redirect: 'follow'
};
fetch("../get_avancement.php?port=" + document.getElementById("port").textContent, requestOptions)
.then(function(response) {
return response.text();
})
.then(function(myText) {
//console.log("TEST: result :" + myText);
var dialogbox = document.getElementById("dialogbox");
previous = dialogbox.textContent;
console.log("previous: " + previous);
if(myText==""){
myText = "rien a afficher";
if(previous!="" && previous != "rien a afficher"){
console.log("changer de page!!!");
// Simulate an HTTP redirect:
var url = document.getElementById("site");
window.location.replace("site_info.php?url=" + url.textContent);
}
}
// Parse JSON here and get progression :
//myText
const json = myText;
const obj = JSON.parse(json);
// obj.progression; pourcentage de progression
document.getElementById("progressbar").value = obj.progression;
dialogbox.textContent = obj.message;
})
.then(result => console.log(result))
.catch(error => console.log('error', error));
}
setInterval(loopNotif, 1000);
</script>
<?php include("pieddepage.php");

54
site_info.php
View File

@ -0,0 +1,54 @@
<?php /* Projet learning 2019 */
ini_set('display_errors', 1);
error_reporting(~0);
if (isset($_GET['url'])){
include("head.php");
include("entete.php");
include("bdd.php");
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<h2>Robot Index</h2>
<?php echo '<h4>' . $_GET['url'] . '</h4></br>';?>
<table class="tab-result">
<tbody>
<?php
//phpinfo();
$requete = $bdd -> prepare('
SELECT url, creation_date, link, score
FROM SITE, LINK
WHERE url = site_id
AND url = ? AND length(link) > 2 AND link NOT like "%cdn%" AND link like "%.%"
ORDER BY score DESC;
');
$requete->execute(array($_GET['url']));
// Declaration du tableau associatif
while ($donnes = $requete->fetch()){
echo '<tr><th>' . $donnes['link'] . '</th>';
echo '<td id="td-padding">' . $donnes['score'] . '</td></tr>';
}
?>
</tbody>
</table>
<h4 style="margin-top:30px;"><a style="color: #0C62A6;" href="index.php">Relancer une recherche</a></h4>
</div>
</div>
<?php include("pieddepage.php");
} else {
echo 'Manque un parametre url';
} // fin du isset

76
top.php
View File

@ -0,0 +1,76 @@
<?php /* Projet learning 2019 */
ini_set('display_errors', 1);
error_reporting(~0);
include("head.php");
include("entete.php");
include("bdd.php");
// SELECT count(*) FROM robot.SITE WHERE url NOT like "%:%:%";
$requete = $bdd -> prepare('
SELECT count(*)
FROM robot.SITE
WHERE url NOT like "%:%:%";');
$requete->execute();
$nbsites = $requete->fetch();
?>
<!-- Corps de la page -->
<div class="content">
<div class="content-inside">
<div class="text-padding">
<h1>Top 50 des liens</h1>
<h3>trouvés par <strong>Robot Index</strong></h3>
<p> <strong>Liens externes apparus</strong> le plus souvent parmi les <strong> <?php echo $nbsites['count(*)'] ?> sites </strong> testés avec notre robot<p>
<p> <strong>O : occurences</strong> somme des apparitions de ce lien - <strong>P : pertinence</strong> nombre de sites visant ce lien.<p>
</div>
<table class="tab-result">
<tbody>
<?php
//phpinfo();
$requete = $bdd -> prepare('
SELECT link, sum(score) as "sum" , count(link) as "from_nb_sites"
FROM LINK
WHERE length(link) > 2 AND link NOT like "%cdn%" AND link like "%.%"
GROUP BY link
HAVING count(link) > 1
ORDER BY count(link) DESC
LIMIT 50;
');
$requete->execute();
echo '<div class="tab-top-50">
<tr class="table-head-top-50">
<td>Liens</td>
<td id ="td-padding">O</td>
<td id ="td-padding">P</td>
</tr>';
while ($donnes = $requete->fetch()){
echo '<tr><th><a style="color: #0C62A6;" href="//'.$donnes['link'].'">' . $donnes['link'] . '</a></th>';
echo '<td id ="td-padding">' . $donnes['sum'] . '</td>';
//echo '<td id ="td-padding">' . $donnes['from_nb_sites'] . '</td></tr> </div">';
echo '<td id ="td-padding"><a style="color: #0C62A6;" href="provenance.php?link='.$donnes['link'].' ">' . $donnes['from_nb_sites']. '</a></td></tr> </div">';
// <a style="color: #0C62A6;" href="provenance.php?url='.$donnes['link'].' ">' . $donnes['from_nb_sites']. '</a>
}
?>
</tbody>
</table>
</div>
</div>
<?php include("pieddepage.php");
Write Preview
Loading…
Cancel
Save