the missing [context] left out all the data

app/Domains/Prompts/EventPageAsArrayPrompt.php

@@ -0,0 +1,62 @@
+<?php
+
+namespace App\Domains\Prompts;
+
+use Illuminate\Support\Facades\Log;
+
+class EventPageAsArrayPrompt
+{
+    public static function prompt(string $context): string
+    {
+        Log::info('[LaraChain] - EventPageAsArrayPrompt');
+
+        return <<<PROMPT
+<ROLE>
+You are an AI assistant tasked with extracting event data from website content.
+
+<INSTRUCTIONS>
+1. Analyze the provided website HTML content below the <CONTENT> tag.
+2. Look for information about events within the content.
+3. If no event data is found, respond with a single word: false
+4. If event data is found, extract the following information for each event:
+   - Event Title
+   - Start Date
+   - End Date
+   - Location
+   - Description
+   - Any other relevant data
+5. Format the extracted data as a JSON array according to the specifications below.
+
+<OUTPUT_FORMAT>
+If events are found, return a JSON array with the following structure:
+
+[
+  {
+    "title": "Event Title",
+    "startDate": "Start Date",
+    "endDate": "End Date",
+    "location": "Location",
+    "description": "Description",
+    "additionalInfo": "Any other relevant data"
+  },
+  {
+    "title": "Event Title",
+    "startDate": "Start Date",
+    "endDate": "End Date",
+    "location": "Location",
+    "description": "Description",
+    "additionalInfo": "Any other relevant data"
+  }
+]
+
+If no events are found, return an empty JSON array: []
+
+<CONTENT>
+$context
+</CONTENT>
+
+Respond only with the JSON array or 'false' if no events are found. Do not include any explanations or additional text in your response.
+
+PROMPT;
+    }
+}

app/Domains/Prompts/EventPagePrompt.php

@@ -25,31 +25,20 @@ public static function prompt(string $context): string
    - Location
    - Description
    - Any other relevant data
-5. Format the extracted data as a JSON array according to the specifications below.
+5. Format the extracted data as a Markdown according to the specifications below.
 
 <OUTPUT_FORMAT>
-If events are found, return a JSON array with the following structure:
-
-[
-  {
-    "title": "Event Title",
-    "startDate": "Start Date",
-    "endDate": "End Date",
-    "location": "Location",
-    "description": "Description",
-    "additionalInfo": "Any other relevant data"
-  },
-  {
-    "title": "Event Title",
-    "startDate": "Start Date",
-    "endDate": "End Date",
-    "location": "Location",
-    "description": "Description",
-    "additionalInfo": "Any other relevant data"
-  }
-]
-
-If no events are found, return an empty JSON array: []
+If events are found, return a Markdown with the following structure for each event on the page:
+"title": "Event Title",
+"startDate": "Start Date",
+"endDate": "End Date",
+"location": "Location",
+"description": "Description",
+"additionalInfo": "Any other relevant data"
+
+
+If no events are found, return the word false
+
 
 <CONTENT>
 $context

app/Http/Controllers/Sources/WebPageSourceController.php

@@ -2,6 +2,7 @@
 
 namespace App\Http\Controllers\Sources;
 
+use App\Domains\Prompts\EventPageAsArrayPrompt;
 use App\Domains\Prompts\EventPagePrompt;
 use App\Domains\Prompts\WebPagePrompt;
 use App\Domains\Sources\SourceTypeEnum;
@@ -55,6 +56,7 @@ public function getPrompts(): array
         return [
             'web_page' => WebPagePrompt::prompt('[CONTEXT]'),
             'event_data' => EventPagePrompt::prompt('[CONTEXT]'),
+            'event_data_as_array' => EventPageAsArrayPrompt::prompt('[CONTEXT]'),
         ];
     }
 }

app/Jobs/GetWebContentJob.php

@@ -86,6 +86,8 @@ public function handle(): void
             Log::info('[LaraChain] - Web Source Skipping', [
                 'prompt' => $prompt,
             ]);
+
+            return;
         } else {
             $promptResultsOriginal = $results->content;
 
@@ -113,6 +115,7 @@ public function handle(): void
                         'source_id' => $this->source->id,
                         'type' => TypesEnum::HTML,
                         'subject' => to_utf8($title),
+                        'document_md5' => md5($promptResult),
                         'link' => $this->webResponseDto->url,
                         'collection_id' => $this->source->collection_id,
                     ],

app/Models/Message.php

@@ -126,6 +126,10 @@ public function getFilter(): ?Filter
 
     public function getPrompt(): string
     {
+        if (! str($this->body)->contains('[CONTEXT]')) {
+            $this->body = str($this->body)->append('[CONTEXT]')->toString();
+        }
+
         return $this->body;
     }
 

database/factories/DocumentFactory.php

@@ -31,6 +31,7 @@ public function definition(): array
             'link' => $this->faker->url(),
             'summary' => $this->faker->text(),
             'subject' => $this->faker->text(),
+            'document_md5' => $this->faker->uuid(),
             'original_content' => $this->faker->text(),
             'child_type' => StructuredTypeEnum::Narrative,
             'file_path' => $this->faker->url(),

database/migrations/2024_07_25_194618_add_md5_to_documents.php

@@ -0,0 +1,28 @@
+<?php
+
+use Illuminate\Database\Migrations\Migration;
+use Illuminate\Database\Schema\Blueprint;
+use Illuminate\Support\Facades\Schema;
+
+return new class extends Migration
+{
+    /**
+     * Run the migrations.
+     */
+    public function up(): void
+    {
+        Schema::table('documents', function (Blueprint $table) {
+            $table->string('document_md5')->nullable();
+        });
+    }
+
+    /**
+     * Reverse the migrations.
+     */
+    public function down(): void
+    {
+        Schema::table('documents', function (Blueprint $table) {
+            //
+        });
+    }
+};