start adding docx

app/Domains/Documents/Transformers/DocXTransformer.php

@@ -0,0 +1,82 @@
+<?php
+
+namespace App\Domains\Documents\Transformers;
+
+use App\Domains\Collections\CollectionStatusEnum;
+use App\Helpers\TextChunker;
+use App\Jobs\VectorlizeDataJob;
+use App\Models\Document;
+use App\Models\DocumentChunk;
+use Illuminate\Support\Facades\Log;
+use PhpOffice\PhpWord\IOFactory;
+
+class DocXTransformer
+{
+    protected Document $document;
+
+    public function handle(Document $document): array
+    {
+        $this->document = $document;
+
+        $filePath = $this->document->pathToFile();
+
+        $parser = IOFactory::createReader('Word2007');
+
+        if (! $parser->canRead($filePath)) {
+            throw new \Exception('Can not read the document '.$filePath);
+        }
+
+        $document = $parser->load($filePath);
+
+        $sections = $document->getSections();
+
+        $content = [];
+        $chunks = [];
+
+        foreach ($sections as $section) {
+            $elements = $section->getElements();
+            foreach ($elements as $element) {
+                /**
+                 * @TODO
+                 * what type of section
+                 * text is easy
+                 * what about images
+                 * what about tables
+                 * what about lists
+                 */
+                $content[] = $element->getText();
+
+            }
+        }
+
+        $content_flattened = implode(' ', $content);
+        $size = config('llmdriver.chunking.default_size');
+        $chunked_chunks = TextChunker::handle($content_flattened, $size);
+        $page = 1;
+
+        foreach ($chunked_chunks as $chunkSection => $chunkContent) {
+            $DocumentChunk = DocumentChunk::updateOrCreate(
+                [
+                    'document_id' => $this->document->id,
+                    'sort_order' => $page,
+                    'section_number' => $chunkSection,
+                ],
+                [
+                    'guid' => md5($chunkContent),
+                    'content' => $chunkContent,
+                    'meta_data' => [$chunkContent],
+                ]
+            );
+
+            $chunks[] = [
+                new VectorlizeDataJob($DocumentChunk),
+            ];
+        }
+
+        notify_collection_ui($document->collection, CollectionStatusEnum::PROCESSING, 'Processing Document');
+
+        Log::info('DocXTransformer:handle', ['chunks' => count($chunks)]);
+
+        return $chunks;
+    }
+}

app/Jobs/ParseDocxJob.php

@@ -0,0 +1,44 @@
+<?php
+
+namespace App\Jobs;
+
+use App\Models\Document;
+use Facades\App\Domains\Documents\Transformers\DocXTransformer;
+use Illuminate\Bus\Batchable;
+use Illuminate\Bus\Queueable;
+use Illuminate\Contracts\Queue\ShouldQueue;
+use Illuminate\Foundation\Bus\Dispatchable;
+use Illuminate\Queue\InteractsWithQueue;
+use Illuminate\Queue\SerializesModels;
+
+class ParseDocxJob implements ShouldQueue
+{
+    use Batchable, Dispatchable, InteractsWithQueue, Queueable, SerializesModels;
+
+    /**
+     * Create a new job instance.
+     */
+    public function __construct(public Document $document)
+    {
+        //
+    }
+
+    /**
+     * Execute the job.
+     */
+    public function handle(): void
+    {
+        if ($this->batch()->cancelled()) {
+            // Determine if the batch has been cancelled...
+
+            return;
+        }
+
+        $chunks = DocXTransformer::handle($this->document);
+
+        foreach ($chunks as $chunk) {
+            $this->batch()->add($chunk);
+        }
+
+    }
+}

app/Jobs/ProcessFileJob.php

@@ -54,6 +54,16 @@ public function handle(): void
                     DocumentProcessingCompleteJob::class,
                 ],
             ],
+            TypesEnum::Docx->value => [
+                'jobs' => [
+                    ParseDocxJob::class,
+                ],
+                'finally' => [
+                    SummarizeDocumentJob::class,
+                    TagDocumentJob::class,
+                    DocumentProcessingCompleteJob::class,
+                ],
+            ],
             TypesEnum::Txt->value => [
                 'jobs' => [
                     ProcessTextFilesJob::class,

composer.json

@@ -24,6 +24,7 @@
         "openai-php/laravel": "^0.8.1",
         "owenvoke/blade-fontawesome": "^2.3",
         "phpoffice/phppresentation": "dev-fix-pptx",
+        "phpoffice/phpword": "^1.2",
         "pusher/pusher-php-server": "^7.2",
         "roach-php/core": "^3.2",
         "roach-php/laravel": "^3.1",

database/factories/DocumentFactory.php

@@ -74,6 +74,15 @@ public function email(): Factory
         });
     }
 
+    public function docx(): Factory
+    {
+        return $this->state(function (array $attributes) {
+            return [
+                'type' => TypesEnum::Docx,
+            ];
+        });
+    }
+
     public function pptx(): Factory
     {
         return $this->state(function (array $attributes) {