Update picofeed
This commit is contained in:
Frederic Guillot
parent
c40640e582
commit
cd80440ae2
|
|
@ -6,7 +6,7 @@
|
|||
"fguillot/simple-validator": "v1.0.0",
|
||||
"fguillot/json-rpc": "v1.0.2",
|
||||
"fguillot/picodb": "v1.0.2",
|
||||
"fguillot/picofeed": "v0.1.10"
|
||||
"fguillot/picofeed": "v0.1.11"
|
||||
},
|
||||
"require-dev": {
|
||||
"phpunit/phpunit": "4.8.3",
|
||||
|
|
|
|||
|
|
@ -4,4 +4,4 @@
|
|||
|
||||
require_once __DIR__ . '/composer' . '/autoload_real.php';
|
||||
|
||||
return ComposerAutoloaderInit26454215d679bdaf4be648601e93de89::getLoader();
|
||||
return ComposerAutoloaderInit6f38ac58c5736de0153f570459656edc::getLoader();
|
||||
|
|
|
|||
|
|
@ -2,7 +2,7 @@
|
|||
|
||||
// autoload_real.php @generated by Composer
|
||||
|
||||
class ComposerAutoloaderInit26454215d679bdaf4be648601e93de89
|
||||
class ComposerAutoloaderInit6f38ac58c5736de0153f570459656edc
|
||||
{
|
||||
private static $loader;
|
||||
|
||||
|
|
@ -19,9 +19,9 @@ class ComposerAutoloaderInit26454215d679bdaf4be648601e93de89
|
|||
return self::$loader;
|
||||
}
|
||||
|
||||
spl_autoload_register(array('ComposerAutoloaderInit26454215d679bdaf4be648601e93de89', 'loadClassLoader'), true, true);
|
||||
spl_autoload_register(array('ComposerAutoloaderInit6f38ac58c5736de0153f570459656edc', 'loadClassLoader'), true, true);
|
||||
self::$loader = $loader = new \Composer\Autoload\ClassLoader();
|
||||
spl_autoload_unregister(array('ComposerAutoloaderInit26454215d679bdaf4be648601e93de89', 'loadClassLoader'));
|
||||
spl_autoload_unregister(array('ComposerAutoloaderInit6f38ac58c5736de0153f570459656edc', 'loadClassLoader'));
|
||||
|
||||
$map = require __DIR__ . '/autoload_namespaces.php';
|
||||
foreach ($map as $namespace => $path) {
|
||||
|
|
@ -42,14 +42,14 @@ class ComposerAutoloaderInit26454215d679bdaf4be648601e93de89
|
|||
|
||||
$includeFiles = require __DIR__ . '/autoload_files.php';
|
||||
foreach ($includeFiles as $file) {
|
||||
composerRequire26454215d679bdaf4be648601e93de89($file);
|
||||
composerRequire6f38ac58c5736de0153f570459656edc($file);
|
||||
}
|
||||
|
||||
return $loader;
|
||||
}
|
||||
}
|
||||
|
||||
function composerRequire26454215d679bdaf4be648601e93de89($file)
|
||||
function composerRequire6f38ac58c5736de0153f570459656edc($file)
|
||||
{
|
||||
require $file;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -37,55 +37,6 @@
|
|||
"description": "Simple validator library",
|
||||
"homepage": "https://github.com/fguillot/simpleValidator"
|
||||
},
|
||||
{
|
||||
"name": "fguillot/picofeed",
|
||||
"version": "v0.1.10",
|
||||
"version_normalized": "0.1.10.0",
|
||||
"source": {
|
||||
"type": "git",
|
||||
"url": "https://github.com/fguillot/picoFeed.git",
|
||||
"reference": "a2cc36244278afbfb9578037b1700ca3cd5a87d7"
|
||||
},
|
||||
"dist": {
|
||||
"type": "zip",
|
||||
"url": "https://api.github.com/repos/fguillot/picoFeed/zipball/a2cc36244278afbfb9578037b1700ca3cd5a87d7",
|
||||
"reference": "a2cc36244278afbfb9578037b1700ca3cd5a87d7",
|
||||
"shasum": ""
|
||||
},
|
||||
"require": {
|
||||
"ext-dom": "*",
|
||||
"ext-iconv": "*",
|
||||
"ext-libxml": "*",
|
||||
"ext-simplexml": "*",
|
||||
"ext-xml": "*",
|
||||
"php": ">=5.3.0"
|
||||
},
|
||||
"suggest": {
|
||||
"ext-curl": "PicoFeed will use cURL if present"
|
||||
},
|
||||
"time": "2015-09-12 11:05:28",
|
||||
"bin": [
|
||||
"picofeed"
|
||||
],
|
||||
"type": "library",
|
||||
"installation-source": "dist",
|
||||
"autoload": {
|
||||
"psr-0": {
|
||||
"PicoFeed": "lib/"
|
||||
}
|
||||
},
|
||||
"notification-url": "https://packagist.org/downloads/",
|
||||
"license": [
|
||||
"MIT"
|
||||
],
|
||||
"authors": [
|
||||
{
|
||||
"name": "Frédéric Guillot"
|
||||
}
|
||||
],
|
||||
"description": "Modern library to handle RSS/Atom feeds",
|
||||
"homepage": "https://github.com/fguillot/picoFeed"
|
||||
},
|
||||
{
|
||||
"name": "fguillot/json-rpc",
|
||||
"version": "v1.0.2",
|
||||
|
|
@ -162,5 +113,54 @@
|
|||
],
|
||||
"description": "Minimalist database query builder",
|
||||
"homepage": "https://github.com/fguillot/picoDb"
|
||||
},
|
||||
{
|
||||
"name": "fguillot/picofeed",
|
||||
"version": "v0.1.11",
|
||||
"version_normalized": "0.1.11.0",
|
||||
"source": {
|
||||
"type": "git",
|
||||
"url": "https://github.com/fguillot/picoFeed.git",
|
||||
"reference": "14080ad7eef9f88744629d67a62da7c7b43d4e2c"
|
||||
},
|
||||
"dist": {
|
||||
"type": "zip",
|
||||
"url": "https://api.github.com/repos/fguillot/picoFeed/zipball/14080ad7eef9f88744629d67a62da7c7b43d4e2c",
|
||||
"reference": "14080ad7eef9f88744629d67a62da7c7b43d4e2c",
|
||||
"shasum": ""
|
||||
},
|
||||
"require": {
|
||||
"ext-dom": "*",
|
||||
"ext-iconv": "*",
|
||||
"ext-libxml": "*",
|
||||
"ext-simplexml": "*",
|
||||
"ext-xml": "*",
|
||||
"php": ">=5.3.0"
|
||||
},
|
||||
"suggest": {
|
||||
"ext-curl": "PicoFeed will use cURL if present"
|
||||
},
|
||||
"time": "2015-10-12 18:20:52",
|
||||
"bin": [
|
||||
"picofeed"
|
||||
],
|
||||
"type": "library",
|
||||
"installation-source": "dist",
|
||||
"autoload": {
|
||||
"psr-0": {
|
||||
"PicoFeed": "lib/"
|
||||
}
|
||||
},
|
||||
"notification-url": "https://packagist.org/downloads/",
|
||||
"license": [
|
||||
"MIT"
|
||||
],
|
||||
"authors": [
|
||||
{
|
||||
"name": "Frédéric Guillot"
|
||||
}
|
||||
],
|
||||
"description": "Modern library to handle RSS/Atom feeds",
|
||||
"homepage": "https://github.com/fguillot/picoFeed"
|
||||
}
|
||||
]
|
||||
|
|
|
|||
|
|
@ -88,9 +88,10 @@ class Html
|
|||
*/
|
||||
public function __construct($html, $website)
|
||||
{
|
||||
$this->config = new Config;
|
||||
$this->input = XmlParser::HtmlToXml($html);
|
||||
$this->output = '';
|
||||
$this->tag = new Tag;
|
||||
$this->tag = new Tag($this->config);
|
||||
$this->website = $website;
|
||||
$this->attribute = new Attribute(new Url($website));
|
||||
}
|
||||
|
|
|
|||
|
|
@ -3,7 +3,9 @@
|
|||
namespace PicoFeed\Filter;
|
||||
|
||||
use DOMXpath;
|
||||
|
||||
use PicoFeed\Parser\XmlParser;
|
||||
use PicoFeed\Config\Config;
|
||||
|
||||
/**
|
||||
* Tag Filter class
|
||||
|
|
@ -13,6 +15,14 @@ use PicoFeed\Parser\XmlParser;
|
|||
*/
|
||||
class Tag
|
||||
{
|
||||
/**
|
||||
* Config object
|
||||
*
|
||||
* @access private
|
||||
* @var \PicoFeed\Config\Config
|
||||
*/
|
||||
private $config;
|
||||
|
||||
/**
|
||||
* Tags blacklist (Xpath expressions)
|
||||
*
|
||||
|
|
@ -71,6 +81,11 @@ class Tag
|
|||
'q',
|
||||
);
|
||||
|
||||
public function __construct(Config $config)
|
||||
{
|
||||
$this->config = $config;
|
||||
}
|
||||
|
||||
/**
|
||||
* Check if the tag is allowed and is not a pixel tracker
|
||||
*
|
||||
|
|
@ -130,7 +145,10 @@ class Tag
|
|||
*/
|
||||
public function isAllowedTag($tag)
|
||||
{
|
||||
return in_array($tag, $this->tag_whitelist);
|
||||
return in_array($tag, array_merge(
|
||||
$this->tag_whitelist,
|
||||
array_keys($this->config->getFilterWhitelistedTags(array()))
|
||||
));
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
|
|||
|
|
@ -2,14 +2,14 @@
|
|||
return array(
|
||||
'grabber' => array(
|
||||
'%.*%' => array(
|
||||
'test_url' => 'http://www.numerama.com/magazine/26857-bientot-des-robots-dans-les-cuisines-de-mcdo.html',
|
||||
'test_url' => 'http://www.numerama.com/sciences/125959-recherches-ladn-recompensees-nobel-de-chimie.html',
|
||||
'body' => array(
|
||||
'//div[@class="col_left"]//div[@class="content"]',
|
||||
'//article',
|
||||
),
|
||||
'strip' => array(
|
||||
'//div[@class="news_social"]',
|
||||
'//div[@id="newssuiv"]',
|
||||
'//footer',
|
||||
'//section[@class="related-article"]',
|
||||
)
|
||||
)
|
||||
)
|
||||
);
|
||||
);
|
||||
|
|
|
|||
Loading…
Reference in New Issue