Add URL refiners

CHANGELOG.md

@@ -7,6 +7,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 ### Added
 * Static method `UserAgent::mozilla5CompatibleBrowser()` to get a `UserAgent` instance with the user agent string `Mozilla/5.0 (compatible)` and also the new method `withMozilla5CompatibleUserAgent` in the `AnonymousHttpCrawlerBuilder` that you can use like this: `HttpCrawler::make()->withMozilla5CompatibleUserAgent()`.
+* URL refiners: `UrlRefiner::withScheme()`, `UrlRefiner::withHost()`, `UrlRefiner::withPort()`, `UrlRefiner::withoutPort()`, `UrlRefiner::withPath()`, `UrlRefiner::withQuery()`, `UrlRefiner::withoutQuery()`, `UrlRefiner::withFragment()` and `UrlRefiner::withoutFragment()`.
 
 ## [1.9.5] - 2024-07-25
 ### Fixed

src/Steps/Refiners/String/StrAfterFirst.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $first) {}
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::afterFirst()', $value);
+            $this->logTypeWarning('StringRefiner::afterFirst()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrAfterLast.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $last) {}
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::afterLast()', $value);
+            $this->logTypeWarning('StringRefiner::afterLast()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrBeforeFirst.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $first) {}
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::beforeFirst()', $value);
+            $this->logTypeWarning('StringRefiner::beforeFirst()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrBeforeLast.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $last) {}
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::beforeLast()', $value);
+            $this->logTypeWarning('StringRefiner::beforeLast()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrBetweenFirst.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $start, protected readonly
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::betweenFirst()', $value);
+            $this->logTypeWarning('StringRefiner::betweenFirst()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrBetweenLast.php

@@ -11,7 +11,7 @@ public function __construct(protected readonly string $start, protected readonly
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::betweenLast()', $value);
+            $this->logTypeWarning('StringRefiner::betweenLast()', $value);
 
             return $value;
         }

src/Steps/Refiners/String/StrReplace.php

@@ -18,7 +18,7 @@ public function __construct(
     public function refine(mixed $value): mixed
     {
         if (!is_string($value)) {
-            $this->logTypeWarning('Str::replace()', $value);
+            $this->logTypeWarning('StringRefiner::replace()', $value);
 
             return $value;
         }

src/Steps/Refiners/Url/AbstractUrlRefiner.php

@@ -0,0 +1,34 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Crawler\Steps\Refiners\AbstractRefiner;
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+use Psr\Http\Message\UriInterface;
+
+abstract class AbstractUrlRefiner extends AbstractRefiner
+{
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    public function refine(mixed $value): mixed
+    {
+        if (!is_string($value) && !$value instanceof Url && !$value instanceof UriInterface) {
+            $this->logTypeWarning($this->staticRefinerMethod(), $value);
+
+            return $value;
+        }
+
+        if (!$value instanceof Url) {
+            $value = Url::parse($value);
+        }
+
+        return $this->refineUrl($value);
+    }
+
+    abstract protected function staticRefinerMethod(): string;
+
+    abstract protected function refineUrl(Url $url): string;
+}

src/Steps/Refiners/Url/WithFragment.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithFragment extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly string $fragment) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withFragment()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->fragment($this->fragment);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithHost.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithHost extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly string $host) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withHost()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->host($this->host);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithPath.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithPath extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly string $path) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withPath()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->path($this->path);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithPort.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithPort extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly int $port) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withPort()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->port($this->port);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithQuery.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithQuery extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly string $query) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withQuery()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->query($this->query);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithScheme.php

@@ -0,0 +1,27 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithScheme extends AbstractUrlRefiner
+{
+    public function __construct(protected readonly string $scheme) {}
+
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withScheme()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->scheme($this->scheme);
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/Url/WithoutPort.php

@@ -0,0 +1,25 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners\Url;
+
+use Crwlr\Url\Exceptions\InvalidUrlComponentException;
+use Crwlr\Url\Url;
+use Exception;
+
+class WithoutPort extends AbstractUrlRefiner
+{
+    protected function staticRefinerMethod(): string
+    {
+        return 'UrlRefiner::withoutPort()';
+    }
+
+    /**
+     * @throws InvalidUrlComponentException|Exception
+     */
+    protected function refineUrl(Url $url): string
+    {
+        $url->resetPort();
+
+        return (string) $url;
+    }
+}

src/Steps/Refiners/UrlRefiner.php

@@ -0,0 +1,59 @@
+<?php
+
+namespace Crwlr\Crawler\Steps\Refiners;
+
+use Crwlr\Crawler\Steps\Refiners\Url\WithFragment;
+use Crwlr\Crawler\Steps\Refiners\Url\WithHost;
+use Crwlr\Crawler\Steps\Refiners\Url\WithoutPort;
+use Crwlr\Crawler\Steps\Refiners\Url\WithPath;
+use Crwlr\Crawler\Steps\Refiners\Url\WithPort;
+use Crwlr\Crawler\Steps\Refiners\Url\WithQuery;
+use Crwlr\Crawler\Steps\Refiners\Url\WithScheme;
+
+class UrlRefiner
+{
+    public static function withScheme(string $scheme): WithScheme
+    {
+        return new WithScheme($scheme);
+    }
+
+    public static function withHost(string $host): WithHost
+    {
+        return new WithHost($host);
+    }
+
+    public static function withPort(int $port): WithPort
+    {
+        return new WithPort($port);
+    }
+
+    public static function withoutPort(): WithoutPort
+    {
+        return new WithoutPort();
+    }
+
+    public static function withPath(string $path): WithPath
+    {
+        return new WithPath($path);
+    }
+
+    public static function withQuery(string $query): WithQuery
+    {
+        return new WithQuery($query);
+    }
+
+    public static function withoutQuery(): WithQuery
+    {
+        return new WithQuery('');
+    }
+
+    public static function withFragment(string $fragment): WithFragment
+    {
+        return new WithFragment($fragment);
+    }
+
+    public static function withoutFragment(): WithFragment
+    {
+        return new WithFragment('');
+    }
+}

tests/Steps/Refiners/String/AfterFirstTest.php

@@ -15,7 +15,7 @@
 
     $logOutput = $this->getActualOutputForAssertion();
 
-    expect($logOutput)->toContain('Refiner Str::afterFirst() can\'t be applied to value of type ' . gettype($value));
+    expect($logOutput)->toContain('Refiner StringRefiner::afterFirst() can\'t be applied to value of type ' . gettype($value));
 
     expect($refinedValue)->toBe($value);
 })->with([

tests/Steps/Refiners/String/AfterLastTest.php

@@ -15,7 +15,7 @@
 
     $logOutput = $this->getActualOutputForAssertion();
 
-    expect($logOutput)->toContain('Refiner Str::afterLast() can\'t be applied to value of type ' . gettype($value));
+    expect($logOutput)->toContain('Refiner StringRefiner::afterLast() can\'t be applied to value of type ' . gettype($value));
 
     expect($refinedValue)->toBe($value);
 })->with([

tests/Steps/Refiners/String/BeforeFirstTest.php

@@ -15,7 +15,7 @@
 
     $logOutput = $this->getActualOutputForAssertion();
 
-    expect($logOutput)->toContain('Refiner Str::beforeFirst() can\'t be applied to value of type ' . gettype($value));
+    expect($logOutput)->toContain('Refiner StringRefiner::beforeFirst() can\'t be applied to value of type ' . gettype($value));
 
     expect($refinedValue)->toBe($value);
 })->with([

tests/Steps/Refiners/String/BeforeLastTest.php

@@ -15,7 +15,7 @@
 
     $logOutput = $this->getActualOutputForAssertion();
 
-    expect($logOutput)->toContain('Refiner Str::beforeLast() can\'t be applied to value of type ' . gettype($value));
+    expect($logOutput)->toContain('Refiner StringRefiner::beforeLast() can\'t be applied to value of type ' . gettype($value));
 
     expect($refinedValue)->toBe($value);
 })->with([