From 70e78bcf42d85630afb47feb0406dc4ee0054bbb Mon Sep 17 00:00:00 2001
From: Lewis Dale <lewis.dale@fortacode.co.uk>
Date: Fri, 20 Sep 2024 13:22:05 +0100
Subject: [PATCH] Setup feeds, request, and filter them

---
 bin/doctrine                       |  9 ++-
 composer.json                      |  5 +-
 src/Controllers/FeedController.php |  2 +-
 src/Models/Data/Feed.php           | 61 ++++++++++++++++++++
 src/Models/Data/FeedFilter.php     | 43 ++++++++++++++-
 src/Models/Types/XMLElement.php    | 40 ++++++++++++++
 src/Requests/Robots.php            | 89 ++++++++++++++++++++++++++++++
 src/Tools/Console/TestFeed.php     | 79 ++++++++++++++++++++++++++
 src/app.php                        |  2 +
 src/dependencies.php               |  4 ++
 10 files changed, 330 insertions(+), 4 deletions(-)
 create mode 100644 src/Models/Types/XMLElement.php
 create mode 100644 src/Requests/Robots.php
 create mode 100644 src/Tools/Console/TestFeed.php

diff --git a/bin/doctrine b/bin/doctrine
index 20ceb0e..600436b 100644
--- a/bin/doctrine
+++ b/bin/doctrine
@@ -4,6 +4,8 @@ require_once __DIR__ . '/../vendor/autoload.php';
 use Doctrine\ORM\EntityManager;
 use Doctrine\ORM\Tools\Console\ConsoleRunner;
 use Doctrine\ORM\Tools\Console\EntityManagerProvider\SingleManagerProvider;
+use Lewisdale\App\Models\Repositories\FeedRepository;
+use Lewisdale\App\Tools\Console\TestFeed;
 
 $dotenv = Dotenv\Dotenv::createImmutable([__DIR__, __DIR__ . "/.."]);
 $dotenv->load();
@@ -12,6 +14,11 @@ $dotenv->load();
 require_once __DIR__ . '/../src/dependencies.php';
 global $container;
 
+$commands = [
+   $container->get(TestFeed::class),
+];
+
 ConsoleRunner::run(
-    new SingleManagerProvider($container->get(EntityManager::class))
+    new SingleManagerProvider($container->get(EntityManager::class)),
+    $commands
 );
\ No newline at end of file
diff --git a/composer.json b/composer.json
index 39f699f..d77a07d 100644
--- a/composer.json
+++ b/composer.json
@@ -20,7 +20,10 @@
         "symfony/dom-crawler": "^6.3",
         "symfony/css-selector": "^6.3",
         "symfony/http-client": "^6.3",
-        "league/uri": "^6.8"
+        "league/uri": "^6.8",
+        "ext-simplexml": "*",
+        "ext-curl": "*",
+        "ext-dom": "*"
     },
     "require-dev": {
         "phpunit/phpunit": "^10.0"
diff --git a/src/Controllers/FeedController.php b/src/Controllers/FeedController.php
index 628d6e1..b711aab 100644
--- a/src/Controllers/FeedController.php
+++ b/src/Controllers/FeedController.php
@@ -26,7 +26,7 @@ class FeedController
         return $this->view->render($response, 'index.twig.html', ['feeds' => $feeds]);
     }
 
-    public function create(ServerRequestInterface $request, ResponseInterface $response)
+    public function create(ServerRequestInterface $request, ResponseInterface $response): ResponseInterface
     {
         $this->logger->info("FeedController::create() called");
         $body = $request->getParsedBody();
diff --git a/src/Models/Data/Feed.php b/src/Models/Data/Feed.php
index 356d1a4..11c5e72 100644
--- a/src/Models/Data/Feed.php
+++ b/src/Models/Data/Feed.php
@@ -4,9 +4,16 @@ declare(strict_types=1);
 
 namespace Lewisdale\App\Models\Data;
 
+use Doctrine\Common\Collections\Collection;
+use Doctrine\DBAL\Types\Types;
 use Doctrine\ORM\Mapping as ORM;
+use DOMXPath;
+use Lewisdale\App\Models\Types\XMLElement;
+use Lewisdale\App\Requests\Robots;
 use Ramsey\Uuid\Doctrine\UuidGenerator;
 use Ramsey\Uuid\UuidInterface;
+use SimpleXMLElement;
+use function Symfony\Component\String\s;
 
 #[ORM\Entity]
 #[ORM\Table(name: 'feeds')]
@@ -24,4 +31,58 @@ class Feed
 
     #[ORM\Column(type: 'string')]
     public string $title;
+
+
+    /**
+     * @var Collection<FeedFilter>
+     */
+    #[ORM\OneToMany(mappedBy: 'feed', targetEntity: FeedFilter::class, cascade: ['persist', 'remove'])]
+    public Collection $feedFilters;
+
+    #[ORM\Column(type: XMLElement::NAME, nullable: true)]
+    public SimpleXMLElement|null $remoteFeed;
+
+    #[ORM\Column(type: XMLElement::NAME, nullable: true)]
+    public SimpleXMLElement|null $filteredFeed;
+
+    public function fetch(): void
+    {
+        if (Robots::allowed($this->url)) {
+            $this->remoteFeed = simplexml_load_file($this->url);
+        } else {
+            throw new \Exception("Robots.txt disallows fetching this feed");
+        }
+    }
+
+    private function test_item(\DOMNode $item): bool {
+        foreach ($this->feedFilters as $filter) {
+            if (!$filter->execute(simplexml_import_dom($item))) {
+                return false;
+            }
+        }
+        return true;
+    }
+
+    public function filter() : void
+    {
+        if ($this->remoteFeed === null) {
+            $this->fetch();
+        }
+
+        $dom = dom_import_simplexml($this->remoteFeed);
+        $doc = $dom->ownerDocument;
+
+        $xpath = new DOMXPath($doc);
+        $filter_queries = implode("|", array_map(fn($filter) => $filter->to_xpath(), $this->feedFilters->toArray()));
+
+        $xpath_query = '//item[' . $filter_queries . ']';
+
+        $to_remove = $xpath->query($xpath_query);
+        echo "Removing " . $to_remove->length . " items\n";
+        foreach ($to_remove as $item) {
+            $item->parentNode->removeChild($item);
+        }
+
+        $this->filteredFeed = simplexml_import_dom($dom);
+    }
 }
\ No newline at end of file
diff --git a/src/Models/Data/FeedFilter.php b/src/Models/Data/FeedFilter.php
index 0d1f15a..13cdf0f 100644
--- a/src/Models/Data/FeedFilter.php
+++ b/src/Models/Data/FeedFilter.php
@@ -13,6 +13,24 @@ use Ramsey\Uuid\UuidInterface;
 #[ORM\HasLifecycleCallbacks]
 class FeedFilter
 {
+    public function __construct(
+        FilterTarget $target,
+        FilterType $filter,
+        string $value,
+        Feed $feed,
+        UuidInterface $id = null,
+    )
+    {
+        if ($id !== null) {
+            $this->id = $id;
+        }
+
+        $this->target = $target;
+        $this->filter = $filter;
+        $this->value = $value;
+        $this->feed = $feed;
+    }
+
     #[ORM\Id]
     #[ORM\Column(type: "uuid", unique: true)]
     #[ORM\GeneratedValue(strategy: "CUSTOM")]
@@ -33,7 +51,30 @@ class FeedFilter
     #[ORM\Column(type: 'string')]
     public string $value;
 
-    #[ORM\ManyToOne(targetEntity: Feed::class)]
+    #[ORM\ManyToOne(targetEntity: Feed::class, cascade: ['persist', 'remove'], inversedBy: 'feedFilters')]
     #[ORM\JoinColumn(name: 'feed_id', referencedColumnName: 'id')]
     public Feed $feed;
+
+    public function to_xpath(): string {
+        return match ($this->filter) {
+            FilterType::EXACT => "{$this->target->value}='{$this->value}'",
+            FilterType::INCLUDE => "{$this->target->value}[not(contains(text(), \"{$this->value}\"))]",
+            FilterType::EXCLUDE => "{$this->target->value}[contains(text(), \"{$this->value}\")]",
+        };
+    }
+
+    public function execute(\SimpleXMLElement $item): bool {
+        $value = (string) match ($this->target) {
+            FilterTarget::TITLE => $item->title,
+            FilterTarget::DESCRIPTION => $item->description,
+            FilterTarget::LINK => $item->link,
+        };
+
+        return match ($this->filter) {
+            FilterType::INCLUDE => str_contains(strtolower($value), strtolower($this->value)),
+            FilterType::EXACT => $value === $this->value,
+            FilterType::REGEX => preg_match($this->value, $value),
+            FilterType::EXCLUDE => !str_contains($value, $this->value),
+        };
+    }
 }
\ No newline at end of file
diff --git a/src/Models/Types/XMLElement.php b/src/Models/Types/XMLElement.php
new file mode 100644
index 0000000..84a035b
--- /dev/null
+++ b/src/Models/Types/XMLElement.php
@@ -0,0 +1,40 @@
+<?php
+declare(strict_types=1);
+
+
+namespace Lewisdale\App\Models\Types;
+
+use Doctrine\DBAL\Platforms\AbstractPlatform;
+use Doctrine\DBAL\Types\Type;
+
+class XMLElement extends Type
+{
+    const NAME = 'xml_element';
+
+    public function getSQLDeclaration(array $column, AbstractPlatform $platform): string
+    {
+        return 'TEXT';
+    }
+
+    public function convertToPHPValue($value, AbstractPlatform $platform)
+    {
+        if ($value === null) {
+            return null;
+        }
+        return simplexml_load_string($value);
+    }
+
+    public function convertToDatabaseValue($value, AbstractPlatform $platform)
+    {
+        if ($value instanceof \SimpleXMLElement) {
+            return $value->asXML();
+        } else {
+            return $value;
+        }
+    }
+
+    public function getName(): string
+    {
+        return self::NAME;
+    }
+}
\ No newline at end of file
diff --git a/src/Requests/Robots.php b/src/Requests/Robots.php
new file mode 100644
index 0000000..1fe020c
--- /dev/null
+++ b/src/Requests/Robots.php
@@ -0,0 +1,89 @@
+<?php
+declare(strict_types=1);
+
+
+namespace Lewisdale\App\Requests;
+
+class Robots
+{
+    // Original PHP code by Chirp Internet: www.chirpinternet.eu
+    // Adapted to include 404 and Allow directive checking by Eric at LinkUp.com
+    // Please acknowledge use of this code by including this header.
+    public static function allowed(string $url, string | null $useragent = "Baleen"): bool
+    {
+        // parse url to retrieve host and path
+        $parsed = parse_url($url);
+
+        $agents = [preg_quote('*')];
+        if ($useragent) {
+            $agents[] = preg_quote($useragent, '/');
+        }
+        $agents = implode('|', $agents);
+
+        // location of robots.txt file, only pay attention to it if the server says it exists
+        if (function_exists('curl_init')) {
+            $handle = curl_init("http://{$parsed['host']}/robots.txt");
+            curl_setopt($handle, CURLOPT_RETURNTRANSFER, TRUE);
+            $response = curl_exec($handle);
+            $httpCode = curl_getinfo($handle, CURLINFO_HTTP_CODE);
+            if (200 == $httpCode) {
+                $robots_txt = explode("\n", $response);
+            } else {
+                $robots_txt = FALSE;
+            }
+            curl_close($handle);
+        } else {
+            $robots_txt = @file("http://{$parsed['host']}/robots.txt");
+        }
+
+        // if there isn't a robots, then we're allowed in
+        if (empty($robots_txt)) {
+            return true;
+        }
+
+        $rules = [];
+        $rule_applies = FALSE;
+
+        foreach ($robots_txt as $line) {
+            // skip blank lines
+            if (!$line = trim($line)) continue;
+
+            // following rules only apply if User-agent matches $useragent or '*'
+            if (preg_match('/^\s*User-agent: (.*)/i', $line, $match)) {
+                $rule_applies = preg_match("/($agents)/i", $match[1]);
+                continue;
+            }
+
+            if ($rule_applies) {
+                list($type, $rule) = explode(':', $line, 2);
+                $type = trim(strtolower($type));
+                // add rules that apply to array for testing
+                $rules[] = [
+                    'type' => $type,
+                    'match' => preg_quote(trim($rule), '/'),
+                ];
+            }
+
+        }
+
+        $isAllowed = TRUE;
+        $currentStrength = 0;
+
+        foreach ($rules as $rule) {
+            // check if page hits on a rule
+            if (preg_match("/^{$rule['match']}/", $parsed['path'])) {
+                // prefer longer (more specific) rules and Allow trumps Disallow if rules same length
+                $strength = strlen($rule['match']);
+                if ($currentStrength < $strength) {
+                    $currentStrength = $strength;
+                    $isAllowed = ("allow" == $rule['type']);
+                } elseif ($currentStrength == $strength && ("allow" == $rule['type'])) {
+                    $currentStrength = $strength;
+                    $isAllowed = TRUE;
+                }
+            }
+        }
+
+        return $isAllowed;
+    }
+}
\ No newline at end of file
diff --git a/src/Tools/Console/TestFeed.php b/src/Tools/Console/TestFeed.php
new file mode 100644
index 0000000..717882f
--- /dev/null
+++ b/src/Tools/Console/TestFeed.php
@@ -0,0 +1,79 @@
+<?php
+declare(strict_types=1);
+
+
+namespace Lewisdale\App\Tools\Console;
+
+use Symfony\Component\Console\{Attribute\AsCommand,
+    Command\Command,
+    Input\InputArgument,
+    Input\InputInterface,
+    Output\OutputInterface};
+use Doctrine\Common\Collections\ArrayCollection;
+use Doctrine\ORM\EntityManager;
+use Lewisdale\App\Models\Data\Feed;
+use Lewisdale\App\Models\Data\FeedFilter;
+use Lewisdale\App\Models\Data\FilterTarget;
+use Lewisdale\App\Models\Data\FilterType;
+use Lewisdale\App\Models\Repositories\FeedRepository;
+use function DI\add;
+
+#[AsCommand(name: "test:feed")]
+class TestFeed extends Command
+{
+    public function __construct(private readonly EntityManager $em,
+                                private readonly FeedRepository $feedRepository)
+    {
+        parent::__construct();
+    }
+
+    protected function configure(): void
+    {
+        $this
+            ->addArgument("url", InputArgument::REQUIRED, "The URL of the feed to test")
+            ->addArgument("title", InputArgument::REQUIRED, "The title of the feed to test");
+    }
+
+    protected function execute(InputInterface $input, OutputInterface $output): int
+    {
+        $output->writeln([
+            "Testing Feed",
+            "============",
+        ]);
+
+        $url = $input->getArgument("url");
+        $title = $input->getArgument("title");
+
+//        $output->writeln("Setting up a new feed for: $url");
+//
+//        $feed = new Feed();
+//        $feed->url = $url;
+//        $feed->title = $title;
+//        $feed->fetch();
+//        $feed->feedFilters = new ArrayCollection();
+//        $feed->feedFilters->add( new FeedFilter(FilterTarget::TITLE, FilterType::INCLUDE, "[No Ads]", $feed));
+//
+//        $this->em->persist($feed);
+//        $this->em->flush();
+//
+//        $cnt = $this->feedRepository->count([]);
+//        $output->writeln("Feed count: $cnt");
+
+        $saved_feed = $this->feedRepository->findOneBy(['url' => $url, 'title' => $title]);
+        $output->writeln("Feed title: " . $saved_feed->title);
+
+        $s2_filter = new FeedFilter(FilterTarget::TITLE, FilterType::EXCLUDE, "S2", $saved_feed);
+        $this->em->persist($s2_filter);
+        $this->em->flush();
+
+        $saved_feed = $this->feedRepository->findOneBy(['url' => $url, 'title' => $title]);
+        $saved_feed->filter();
+
+        $this->em->persist($saved_feed);
+        $this->em->flush();
+
+        $output->writeln("Filtered feed: " . $saved_feed->filteredFeed->asXML('feed_filtered.xml'));
+
+        return Command::SUCCESS;
+    }
+}
\ No newline at end of file
diff --git a/src/app.php b/src/app.php
index c4cde4b..e4efe59 100644
--- a/src/app.php
+++ b/src/app.php
@@ -4,6 +4,8 @@ use Lewisdale\App\Controllers\FeedController;
 use Lewisdale\App\Controllers\SampleController;
 use Slim\Views\TwigMiddleware;
 
+ini_set('user_agent', 'Baleen/1.0 (https://lewisdale.dev)');
+
 require_once __DIR__ . "/dependencies.php";
 
 global $container;
diff --git a/src/dependencies.php b/src/dependencies.php
index bc28fe6..4950c53 100644
--- a/src/dependencies.php
+++ b/src/dependencies.php
@@ -6,6 +6,7 @@ use Doctrine\DBAL\Types\Type;
 use Doctrine\ORM\EntityManager;
 use Doctrine\ORM\ORMSetup;
 use Lewisdale\App\Logging\FileLogger;
+use Lewisdale\App\Models\Types\XMLElement;
 use Lewisdale\App\TwigExtensions\CsrfExtension;
 use Psr\Log\LoggerInterface;
 use Ramsey\Uuid\Doctrine\UuidType;
@@ -40,6 +41,7 @@ $container->set(LoggerInterface::class, $container->get(FileLogger::class));
 
 $container->set(EntityManager::class, static function() {
     Type::addType('uuid', UuidType::class);
+    Type::addType(XMLElement::NAME, XMLElement::class);
 
     $config = ORMSetup::createAttributeMetadataConfiguration(
         paths: array(__DIR__."/Models/Data"),
@@ -51,6 +53,8 @@ $container->set(EntityManager::class, static function() {
         'path' => __DIR__ . '/../' . getenv("SQLITE_DB_NAME"),
     ], $config);
 
+    $connection->executeQuery("PRAGMA foreign_keys = ON");
+
     return new EntityManager($connection, $config);
 });