Crispi/dogstats
Archived
2
0
Fork 0
Code Issues 1 Pull Requests Projects Wiki Activity

Merge branch 'master' of https://gitea.haschek.at/Crispi/dogstats

Browse Source
This commit is contained in:
piapassecker 2023-12-25 16:39:12 +01:00
commit 4de6df7175
3 changed files with 31 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitea/workflows/autoupdate.yml
View File

@ -14,6 +14,7 @@ jobs:
uses: actions/checkout@v4
- name: run crawler
run: |
sed -i "/^error_reporting/c\error_reporting = E_ALL & ~E_DEPRECATED & ~E_STRICT & ~E_WARNING & ~E_NOTICE" /etc/php/7.4/cli/php.ini
cd crawler
php crawler.php
git config --global user.email "gitea@haschek.at"

57
crawler/crawler.php
View File

@ -1,43 +1,41 @@
<?php
if(!file_exists('tmp/ergebnisse.html'))
{
$ergebnisse = file_get_contents('https://www.dognow.at/ergebnisse/');
file_put_contents('tmp/ergebnisse.html', $ergebnisse);
}
$GLOBALS['db'] = new SQLite3('data.db');
if(!$GLOBALS['db']) exit("Error loading database");
//389-7625.pdf.csv
// analyzeResultCSV('tmp/csv/389-7637.pdf.csv',7637,389);
// exit();
// if(!file_exists('tmp/ergebnisse.html'))
// {
// $ergebnisse = file_get_contents('https://www.dognow.at/ergebnisse/');
// file_put_contents('tmp/ergebnisse.html', $ergebnisse);
// }
$html = file_get_contents('tmp/ergebnisse.html');
// $html = file_get_contents('tmp/ergebnisse.html');
$dom = new DOMDocument;
$dom->loadHTML($html);
// $dom = new DOMDocument;
// $dom->loadHTML($html);
$xpath = new DOMXPath($dom);
$query = '//ul[@class="pagination"]/child::*';
$nodes = $xpath->query($query);
// $xpath = new DOMXPath($dom);
// $query = '//ul[@class="pagination"]/child::*';
// $nodes = $xpath->query($query);
$GLOBALS['pdfs'] = 0;
// $GLOBALS['pdfs'] = 0;
// Loop through the selected nodes
foreach ($nodes as $node) {
// Do something with each node, for example, echo its content
$url = $node->getElementsByTagName('a')[0]->getAttribute('href');
$number = intval($node->nodeValue);
if($number > $last_page){
$last_page = $number;
}
}
// // Loop through the selected nodes
// foreach ($nodes as $node) {
// // Do something with each node, for example, echo its content
// $url = $node->getElementsByTagName('a')[0]->getAttribute('href');
// $number = intval($node->nodeValue);
// if($number > $last_page){
// $last_page = $number;
// }
// }
echo "[i] Found $last_page pages\n";
// echo "[i] Found $last_page pages\n";
//create an array with all pages
$pages = range(1,65);
// //create an array with all pages
// $pages = range(1,65);
// foreach($pages as $page)
// {
@ -174,6 +172,11 @@ function getResults($run,$event)
$url = "https://www.dognow.at/ergebnisse/pdf.php?lauf=$run&event=$event";
if(!file_exists('tmp/results/' . $event . '-' . $run . '.pdf'))
file_put_contents('tmp/results/' . $event . '-' . $run . '.pdf',file_get_contents($url));
if($GLOBALS['db']->query("SELECT * FROM runs WHERE id = '$run' AND event = '$event'")->fetchArray() != false)
{
echo " [i] Skipping run $run in event $event\n";
return;
}
convertPDFtoCSV('tmp/results/' . $event . '-' . $run . '.pdf','tmp/csv/' . $event . '-' . $run . '.pdf.csv');
analyzeResultCSV('tmp/csv/' . $event . '-' . $run . '.pdf.csv',$run,$event);
}

BIN
crawler/data.db
View File

Binary file not shown.