Решена проблема кодировок

This commit is contained in:
2024-08-12 16:58:12 +03:00
parent 6b7199a326
commit 9ae73e81cf
4 changed files with 7933 additions and 20 deletions

View File

@ -5,6 +5,7 @@ use GuzzleHttp\Exception\ClientException;
use GuzzleHttp\Exception\ConnectException;
use GuzzleHttp\Exception\RequestException;
use GuzzleHttp\Exception\ServerException;
use GuzzleHttp\RequestOptions;
use NilPortugues\Sql\QueryBuilder\Builder\GenericBuilder;
use App\Library\ContingentParser;
use App\Library\Database;
@ -30,7 +31,11 @@ $dbNiimko = new Database(
$dbconfig['user'],
$dbconfig['password']
);
$builder = new GenericBuilder();
// select kod as org_id, site from niimko.s_vuzes
// where ootype = 'vuz' and deleted = 'n' and fake = 'n'
$params = ['vuz', 'n', 'n'];
$query = $builder->select()
->setTable('s_vuzes')
@ -43,12 +48,33 @@ $query = $builder->select()
$sql = $builder->write($query);
$sites = $dbNiimko->executeQuery($sql, $params);
// select id, kod from niimko.s_specs where oopkodes = 'gos3p'
$params = ['gos3p'];
$query = $builder->select()
->setTable('s_specs')
->setColumns(['id', 'kod'])
->where()
->equals('oopkodes','gos3p')
->end();
$sql = $builder->write($query);
$specializations = $dbNiimko->executeQuery($sql, $params);
// print_r($sites);
// print_r($specializations);
// $sites = [ ['site' => "http://marsu.ru"], ['site' => "http://voenmeh.ru"], ['site' => "http://angtu.ru"] ];
$i = 0;
$succes = 0;
foreach ($sites as $site) {
try {
$client = new Client();
$client = new Client([
RequestOptions::ALLOW_REDIRECTS => [
'max' => 10, // allow at most 10 redirects.
'strict' => true, // use "strict" RFC compliant redirects.
'referer' => true, // add a Referer header
'track_redirects' => true,
],
]);
$route = "{$site['site']}/sveden/education/";
$route = str_replace("http","https", $route);
@ -61,7 +87,28 @@ foreach ($sites as $site) {
$html = $response->getBody()->getContents();
$parser = new ContingentParser($html, '//tr[@itemprop="eduChislen"]//');
$contingent = $parser->getDataTable();
// Добавляем поле spec_id по spec_code
foreach ($contingent as $key => $con) {
$needle = $con['spec_code'];
foreach ($specializations as $spec) {
if ($needle == $spec['kod']) {
$contingent[$key]['spec_id'] = $spec['id'];
continue;
}
}
$contingent[$key]['spec_id'] = null;
}
// Добавляем поле org_id
foreach ($contingent as $key => $spec) {
$contingent[$key]['org_id'] = $site['org_id'];
}
print_r($contingent);
if ($response->getStatusCode() == 200 && !empty($contingent)){
print("Succes: ".++$succes."\n");
}
} catch (ClientException $e) {
$response = $e->getCode();
} catch (RequestException $e) {
@ -73,17 +120,6 @@ foreach ($sites as $site) {
}
}
// $specializations = $dbNiimko->selectWhere('s_specs', ['id', 'kod'], ['oopkodes' => 'gos3p']);
// foreach ($contingent as $key => $con) {
// $needle = $con['spec_code'];
// foreach ($specializations as $spec) {
// if ($needle == $spec['kod']) {
// $contingent[$key] += ['spec_id' => $spec['id']];
// }
// }
// }
// Чтобы не дублировались в базе
// $dbOpendata->insert('sveden_education_contingent', $data);
// $dbOpendata->update('sveden_education_contingent', $specializations);

View File

@ -27,7 +27,7 @@ class ContingentParser
{
libxml_use_internal_errors(true);
$dom = new \DOMDocument();
$dom->loadHTML($html);
$dom->loadHTML(mb_convert_encoding($html,'HTML-ENTITIES','UTF-8'));
$this->xpath = new \DOMXPath($dom);
$this->template = $template;
}
@ -39,9 +39,9 @@ class ContingentParser
if (!is_array($tag)) {
$data[$field] = $this->xpath->query($this->template . $tag . "[@itemprop=\"$field\"]");
} else {
$x = $this->xpath->query($this->template . $tag[0] . "[@itemprop=\"$field\"]");
$y = $this->xpath->query($this->template . $tag[1] . "[@itemprop=\"$field\"]");
$data[$field] = $x > $y ? $x : $y;
$th = $this->xpath->query($this->template . $tag[0] . "[@itemprop=\"$field\"]");
$td = $this->xpath->query($this->template . $tag[1] . "[@itemprop=\"$field\"]");
$data[$field] = $th > $td ? $th : $td;
}
}

View File

@ -13,10 +13,10 @@ class ContingentRow
if ($contingent < 0) {
throw new \Exception("Недействительная численность обучающихся!");
}
$this->eduCode = $eduCode;
$this->eduName = $eduName;
$this->eduLevel = $eduLevel;
$this->eduForm = $eduForm;
$this->eduCode = trim($eduCode);
$this->eduName = trim($eduName);
$this->eduLevel = trim($eduLevel);
$this->eduForm = trim($eduForm);
$this->contingent = $contingent;
}