Remove mime type parameter, file discovered automatically

src/Connectors/ParseConnector.php

@@ -76,12 +76,11 @@ public function hasRequestFailed(Response $response): ?bool
      * @param  string  $mimeType  The mime type of the document. Default application/pdf
      * @param  \OneOffTech\Parse\Client\ParseOption  $options  Specifiy additional options for the specific parsing processor
      */
-    public function parse(string $url, string $mimeType = 'application/pdf', ?ParseOption $options = null): DocumentDto
+    public function parse(string $url, ?ParseOption $options = null): DocumentDto
     {
         return $this
             ->send((new ExtractTextRequest(
                 url: $url,
-                mimeType: $mimeType,
                 preferredDocumentProcessor: $options?->processor?->value ?? DocumentProcessor::PDFACT->value,
             ))->validate())
             ->dto();

src/Requests/ExtractTextRequest.php

@@ -18,7 +18,6 @@ class ExtractTextRequest extends Request implements HasBody
 
     public function __construct(
         protected readonly string $url,
-        protected readonly string $mimeType,
         protected readonly string $preferredDocumentProcessor = 'pdfact'
     ) {
         //
@@ -33,7 +32,6 @@ protected function defaultBody(): array
     {
         return [
             'url' => $this->url,
-            'mime_type' => $this->mimeType,
             'driver' => $this->preferredDocumentProcessor ?? 'pdfact',
         ];
     }
@@ -44,10 +42,6 @@ public function validate(): self
             throw new InvalidArgumentException('The [url] is required to be non-empty.');
         }
 
-        if (empty(trim($this->mimeType))) {
-            throw new InvalidArgumentException('The [mime type] is required to be non-empty.');
-        }
-
         return $this;
     }
 

tests/ParseProcessorSelectionTest.php

@@ -30,7 +30,6 @@
         $body = $request->body()->all();
 
         return $body['url'] === 'http://localhost/empty.pdf'
-            && $body['mime_type'] === 'application/pdf'
             && $body['driver'] === 'pdfact';
     });
 
@@ -62,7 +61,6 @@
         $body = $request->body()->all();
 
         return $body['url'] === 'http://localhost/empty.pdf'
-            && $body['mime_type'] === 'application/pdf'
             && $body['driver'] === 'pymupdf';
     });
 
@@ -93,7 +91,6 @@
         $body = $request->body()->all();
 
         return $body['url'] === 'http://localhost/empty.pdf'
-            && $body['mime_type'] === 'application/pdf'
             && $body['driver'] === 'llama';
     });
 
@@ -124,7 +121,6 @@
         $body = $request->body()->all();
 
         return $body['url'] === 'http://localhost/empty.pdf'
-            && $body['mime_type'] === 'application/pdf'
             && $body['driver'] === 'unstructured';
     });
 

tests/ParseTest.php

@@ -20,7 +20,7 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient($mockClient);
 
-    $document = $connector->parse('http://localhost/base.pdf', 'application/pdf');
+    $document = $connector->parse('http://localhost/base.pdf');
 
     expect($document)
         ->toBeInstanceOf(DocumentDto::class)
@@ -63,7 +63,7 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient($mockClient);
 
-    $document = $connector->parse('http://localhost/base.pdf', 'application/pdf', new ParseOption(DocumentProcessor::PYMUPDF));
+    $document = $connector->parse('http://localhost/base.pdf', new ParseOption(DocumentProcessor::PYMUPDF));
 
     expect($document)
         ->toBeInstanceOf(DocumentDto::class)
@@ -91,7 +91,7 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient(MockClient::getGlobal());
 
-    $document = $connector->parse('http://localhost/empty.pdf', 'application/pdf');
+    $document = $connector->parse('http://localhost/empty.pdf');
 
     expect($document)
         ->toBeInstanceOf(DocumentDto::class)
@@ -113,7 +113,7 @@
         /** @var array */
         $body = $request->body()->all();
 
-        return $body['url'] === 'http://localhost/empty.pdf' && $body['mime_type'] === 'application/pdf';
+        return $body['url'] === 'http://localhost/empty.pdf';
     });
 
     $mockClient->assertSentCount(1);
@@ -140,7 +140,7 @@
         /** @var array */
         $body = $request->body()->all();
 
-        return $body['url'] === 'http://localhost/base.docx' && $body['mime_type'] === 'application/pdf';
+        return $body['url'] === 'http://localhost/base.docx';
     });
 
     $mockClient->assertSentCount(1);
@@ -155,7 +155,7 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient(MockClient::getGlobal());
 
-    $connector->parse('http://localhost/test.pdf', 'application/pdf');
+    $connector->parse('http://localhost/test.pdf');
 
     $mockClient->assertSent(ExtractTextRequest::class);
 
@@ -167,7 +167,7 @@
         /** @var array */
         $body = $request->body()->all();
 
-        return $body['url'] === 'http://localhost/test.pdf' && $body['mime_type'] === 'application/pdf';
+        return $body['url'] === 'http://localhost/test.pdf';
     });
 
     $mockClient->assertSentCount(1);
@@ -182,6 +182,6 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient($mockClient);
 
-    $connector->parse('http://localhost/km-f.pdf', 'application/pdf');
+    $connector->parse('http://localhost/km-f.pdf');
 
 })->throws(ServiceUnavailableException::class, 'The pdfact service is not reachable.');

tests/ValidationTest.php

@@ -13,22 +13,8 @@
     $connector = new ParseConnector('fake', 'http://localhost:5002');
     $connector->withMockClient(MockClient::getGlobal());
 
-    $connector->parse('', 'application/pdf');
+    $connector->parse('');
 
     $mockClient->assertNothingSent();
 
 })->throws(InvalidArgumentException::class, 'The [url] is required to be non-empty.');
-
-test('mime type required to be non-null', function () {
-    $mockClient = MockClient::global([
-        ExtractTextRequest::class => MockResponse::fixture('extract-text-invalid-mime'),
-    ]);
-
-    $connector = new ParseConnector('fake', 'http://localhost:5002');
-    $connector->withMockClient(MockClient::getGlobal());
-
-    $connector->parse('http://localhost/test.pdf', '');
-
-    $mockClient->assertNothingSent();
-
-})->throws(InvalidArgumentException::class, 'The [mime type] is required to be non-empty.');