[BUGFIX] Extract request rejected because of missing multipart boundary (#614)

When a ExtractQuery is executed Apache Solr throws the following error:

* SEVERE: null:org.apache.commons.fileupload.FileUploadException: the request was rejected because no multipart boundary was found

The patch:

* Add's the boundary to the header and to the body

Fixes: #301

src/Core/Client/Adapter/AdapterHelper.php

+<?php
+
+namespace Solarium\Core\Client\Adapter;
+
+use Solarium\Core\Client\Request;
+
+/**
+ * Helper class for shared adapter functionality.
+ */
+class AdapterHelper
+{
+    /**
+     * This method is used to build the upload body for a file upload with the boundary markers.
+     *
+     * @param Request $request
+     *
+     * @return string
+     */
+    public function buildUploadBodyFromRequest(Request $request)
+    {
+        $baseName = basename($request->getFileUpload());
+        $body = "--{$request->getHash()}\r\n";
+        $body .= 'Content-Disposition: form-data; name="file"; filename="'.$baseName.'"';
+        $body .= "\r\nContent-Type: application/octet-stream\r\n\r\n";
+        $body .= file_get_contents($request->getFileUpload(), 'r');
+        $body .= "\r\n--{$request->getHash()}--\r\n";
+
+        return $body;
+    }
+}

src/Core/Client/Adapter/Curl.php

@@ -120,12 +120,9 @@ class Curl extends Configurable implements AdapterInterface
             curl_setopt($handler, CURLOPT_POST, true);
 
             if ($request->getFileUpload()) {
-                if (version_compare(PHP_VERSION, '5.5.0') >= 0) {
-                    $curlFile = curl_file_create($request->getFileUpload());
-                    curl_setopt($handler, CURLOPT_POSTFIELDS, ['content' => $curlFile]);
-                } else {
-                    curl_setopt($handler, CURLOPT_POSTFIELDS, ['content' => '@'.$request->getFileUpload()]);
-                }
+                $helper = new AdapterHelper();
+                $data = $helper->buildUploadBodyFromRequest($request);
+                curl_setopt($handler, CURLOPT_POSTFIELDS, $data);
             } else {
                 curl_setopt($handler, CURLOPT_POSTFIELDS, $request->getRawData());
             }

src/Core/Client/Adapter/Guzzle.php

@@ -106,7 +106,8 @@ class Guzzle extends Configurable implements AdapterInterface
         }
 
         if ($request->getFileUpload()) {
-            return fopen($request->getFileUpload(), 'r');
+            $helper = new AdapterHelper();
+            return $helper->buildUploadBodyFromRequest($request);
         }
 
         return $request->getRawData();

src/Core/Client/Adapter/Guzzle3.php

@@ -101,7 +101,9 @@ class Guzzle3 extends Configurable implements AdapterInterface
         }
 
         if ($request->getFileUpload()) {
-            return fopen($request->getFileUpload(), 'r');
+            $helper = new AdapterHelper();
+            $body = $helper->buildUploadBodyFromRequest($request);
+            return $body;
         }
 
         return $request->getRawData();

src/Core/Client/Adapter/Http.php

@@ -87,22 +87,9 @@ class Http extends Configurable implements AdapterInterface
 
         if (Request::METHOD_POST == $method) {
             if ($request->getFileUpload()) {
-                $boundary = '----------'.md5(time());
-                $CRLF = "\r\n";
-                $file = $request->getFileUpload();
-                // Add the proper boundary to the Content-Type header
-                $headers = $request->getHeaders();
-                // Remove the Content-Type header, because we will replace it with something else.
-                if (false !== ($key = array_search('Content-Type: multipart/form-data', $headers, true))) {
-                    unset($headers[$key]);
-                }
-                $request->setHeaders($headers);
-                $request->addHeader("Content-Type: multipart/form-data; boundary={$boundary}");
-                $data = "--{$boundary}".$CRLF;
-                $data .= 'Content-Disposition: form-data; name="upload"; filename='.$file.$CRLF;
-                $data .= 'Content-Type: application/octet-stream'.$CRLF.$CRLF;
-                $data .= file_get_contents($file).$CRLF;
-                $data .= '--'.$boundary.'--';
+                $helper = new AdapterHelper();
+                $data = $helper->buildUploadBodyFromRequest($request);
+
                 $content_length = strlen($data);
                 $request->addHeader("Content-Length: $content_length\r\n");
                 stream_context_set_option(

src/Core/Client/Request.php

@@ -306,4 +306,12 @@ class Request extends Configurable implements RequestParamsInterface
             }
         }
     }
+
+    /**
+     * @return string
+     */
+    public function getHash()
+    {
+        return spl_object_hash($this);
+    }
 }

src/QueryType/Extract/RequestBuilder.php

@@ -68,7 +68,7 @@ class RequestBuilder extends BaseRequestBuilder
         } elseif (is_readable($file)) {
             $request->setFileUpload($file);
             $request->addParam('resource.name', basename($query->getFile()));
-            $request->addHeader('Content-Type: multipart/form-data');
+            $request->addHeader('Content-Type: multipart/form-data; boundary='.$request->getHash());
         } else {
             throw new RuntimeException('Extract query file path/url invalid or not available');
         }

tests/Core/Client/Adapter/Guzzle3Test.php

@@ -115,7 +115,7 @@ final class Guzzle3Test extends TestCase
         $this->assertCount(1, $receivedRequests);
 
         $this->assertSame('POST', $receivedRequests[0]->getMethod());
-        $this->assertStringEqualsFile(__FILE__, (string) $receivedRequests[0]->getBody());
+        $this->assertContains(file_get_contents(__FILE__), (string) $receivedRequests[0]->getBody());
         $this->assertSame(
             'request value',
             (string) $receivedRequests[0]->getHeader('X-PHPUnit')

tests/Core/Client/Adapter/GuzzleTest.php

@@ -115,7 +115,7 @@ final class GuzzleTest extends TestCase
         $this->assertCount(1, $container);
         $this->assertSame('POST', $container[0]['request']->getMethod());
         $this->assertSame('request value', $container[0]['request']->getHeaderline('X-PHPUnit'));
-        $this->assertStringEqualsFile(__FILE__, (string) $container[0]['request']->getBody());
+        $this->assertContains(file_get_contents(__FILE__), (string) $container[0]['request']->getBody());
     }
 
     /**

tests/Integration/AbstractTechproductsTest.php

@@ -8,6 +8,7 @@ use Solarium\Component\QueryTraits\TermsTrait;
 use Solarium\Component\Result\Terms\Result;
 use Solarium\Core\Client\ClientInterface;
 use Solarium\QueryType\Select\Query\Query as SelectQuery;
+use Solarium\QueryType\Select\Result\Document;
 
 abstract class AbstractTechproductsTest extends TestCase
 {
@@ -443,6 +444,54 @@ abstract class AbstractTechproductsTest extends TestCase
 
         $this->assertSame($without, $with);
     }
+
+    public function testExtractIntoDocument()
+    {
+        $extract = $this->client->createExtract();
+        $extract->setUprefix('attr_');
+        $extract->setFile(__DIR__.DIRECTORY_SEPARATOR.'Fixtures'.DIRECTORY_SEPARATOR.'testpdf.pdf');
+        $extract->setCommit(true);
+        $extract->setCommitWithin(0);
+        $extract->setOmitHeader(false);
+
+        // add document
+        $doc = $extract->createDocument();
+        $doc->id = 'extract-test';
+        $extract->setDocument($doc);
+
+        $this->client->extract($extract);
+
+        // now get the document and check the content
+        $select = $this->client->createSelect();
+        $select->setQuery('id:extract-test');
+        $selectResult = $this->client->select($select);
+        $iterator = $selectResult->getIterator();
+
+        /** @var Document $document */
+        $document = $iterator->current();
+        $this->assertSame('PDF Test', trim($document['content'][0]), 'Written document does not contain extracted result');
+
+        // now cleanup the document the have the initial index state
+        $update = $this->client->createUpdate();
+        $update->addDeleteById('extract-test');
+        $update->addCommit(true, true);
+        $this->client->extract($update);
+    }
+
+    public function testExtractTextOnly()
+    {
+        $query = $this->client->createExtract();
+        $fileName = 'testpdf.pdf';
+        $query->setFile(__DIR__.DIRECTORY_SEPARATOR.'Fixtures'.DIRECTORY_SEPARATOR.$fileName);
+        $query->setExtractOnly(true);
+        $query->addParam('extractFormat', 'text');
+
+        $response = $this->client->extract($query);
+        $json = json_decode($response->getResponse()->getBody());
+
+        $content = $json->{$fileName};
+        $this->assertSame('PDF Test', trim($content), 'Can not extract the plain content from the file');
+    }
 }
 
 class TestQuery extends SelectQuery

tests/QueryType/Extract/RequestBuilderTest.php

@@ -106,10 +106,11 @@ class RequestBuilderTest extends TestCase
 
     public function testContentTypeHeader()
     {
+        $request = $this->builder->build($this->query);
         $headers = [
-            'Content-Type: multipart/form-data',
+            'Content-Type: multipart/form-data; boundary='.$request->getHash(),
         ];
-        $request = $this->builder->build($this->query);
+
         $this->assertSame($headers, $request->getHeaders());
     }
 }