Merge pull request #44 from LlmLaraHub/add_firecrawl

Add FireCrawl Scraper

Modules/LlmDriver/app/ClaudeClient.php

@@ -54,6 +54,8 @@ public function chat(array $messages): CompletionResponse
 
         $payload = $this->modifyPayload($payload);
 
+        put_fixture('claude_payload_chat.json', $payload);
+
         $results = $this->getClient()->post('/messages', $payload);
 
         if (! $results->ok()) {
@@ -303,7 +305,6 @@ public function functionPromptChat(array $messages, array $only = []): array
 
         $results = $this->getClient()->post('/messages', [
             'model' => $model,
-            'system' => 'Return a markdown response.',
             'max_tokens' => $maxTokens,
             'messages' => $messages,
             'tools' => $this->getFunctions(),

app/Domains/Messages/RoleEnum.php

@@ -7,4 +7,5 @@ enum RoleEnum: string
     case User = 'user';
     case System = 'system';
     case Assistant = 'assistant';
+    case Tool = 'tool';
 }

app/Domains/Prompts/EventPagePrompt.php

@@ -16,8 +16,8 @@ public static function prompt(string $context): string
 
 <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
+2. Look for information about sporting events within the content.
+3. If no event data is found summarize what is on the page
 4. If event data is found, extract the following information for each event:
    - Event Title
    - Start Date
@@ -37,15 +37,12 @@ public static function prompt(string $context): string
 "additionalInfo": "Any other relevant data"
 
 
-If no events are found, return the word false
+If no events are found, return the words "No Content Found" and summarize what was on the page
 
 
 <CONTENT>
 $context
 </CONTENT>
-
-Respond only with Markdown or 'false' if no events are found. Do not include any explanations or additional text in your response.
-
 PROMPT;
     }
 }

app/Domains/Prompts/SummarizePrompt.php

@@ -18,7 +18,6 @@ public static function prompt(string $originalPrompt, string $context): string
 **Format**
 Deliver the response in a concise, clear Markdown format (Text). Use quotes as needed from the context.
 
-[DO NOT INCLUDE THE ABOVE IN THE RESPONSE]
 
 **The User's Query**:
 ```$originalPrompt```

app/Domains/Sources/WebSearch/GetPage.php

@@ -2,15 +2,19 @@
 
 namespace App\Domains\Sources\WebSearch;
 
+use App\Domains\WebParser\WebContentResultsDto;
 use App\Models\Collection;
+use App\Models\Setting;
+use Facades\App\Domains\WebParser\DefaultClient;
+use Facades\App\Domains\WebParser\FireCrawlClient;
+use Illuminate\Support\Facades\Log;
 use Illuminate\Support\Facades\Storage;
 use League\HTMLToMarkdown\Converter\CodeConverter;
 use League\HTMLToMarkdown\Converter\PreformattedConverter;
 use League\HTMLToMarkdown\Converter\TableConverter;
 use League\HTMLToMarkdown\Converter\TextConverter;
 use League\HTMLToMarkdown\Environment;
 use League\HTMLToMarkdown\HtmlConverter;
-use Spatie\Browsershot\Browsershot;
 
 class GetPage
 {
@@ -24,27 +28,24 @@ public static function make(Collection $collection): self
         return new static($collection);
     }
 
-    public function handle(string $url, bool $parseHtml = true): string
+    public function handle(string $url, bool $parseHtml = true): WebContentResultsDto
     {
-        $results = Browsershot::url($url)
-            ->userAgent('DailyAI Studio Browser 1.0, helping users automate workflows')
-            ->dismissDialogs()
-            ->fullPage();
-
+        $name = md5($url).'.pdf';
         /**
-         * @TODO this can repeat
+         * @TODO
+         * Make this a driver like the rest of the system
          */
-        $name = md5($url).'.pdf';
-
-        Storage::disk('collections')->put($this->collection->id.'/'.$name, $results->pdf());
-
-        $body = $results->bodyHtml();
-
-        if ($parseHtml) {
-            $body = $this->parseHtml($body);
+        if (Setting::getSecret('fire_crawl', 'api_key')) {
+            Log::info('Using FireCrawl');
+            $results = FireCrawlClient::scrape($url);
+        } else {
+            Log::info('Using Default Browsershot');
+            /** @var WebContentResultsDto $results */
+            $results = DefaultClient::scrape($url);
+            Storage::disk('collections')->put($this->collection->id.'/'.$name, $results->browserShot->pdf());
         }
 
-        return $body;
+        return $results;
     }
 
     public function parseHtml(string $html): string

app/Domains/WebParser/BaseWebParserClient.php

@@ -0,0 +1,7 @@
+<?php
+
+namespace App\Domains\WebParser;
+
+abstract class BaseWebParserClient
+{
+}

app/Domains/WebParser/DefaultClient.php

@@ -0,0 +1,56 @@
+<?php
+
+namespace App\Domains\WebParser;
+
+use League\HTMLToMarkdown\Converter\CodeConverter;
+use League\HTMLToMarkdown\Converter\PreformattedConverter;
+use League\HTMLToMarkdown\Converter\TableConverter;
+use League\HTMLToMarkdown\Converter\TextConverter;
+use League\HTMLToMarkdown\Environment;
+use League\HTMLToMarkdown\HtmlConverter;
+use LlmLaraHub\LlmDriver\BaseClient;
+use Spatie\Browsershot\Browsershot;
+
+class DefaultClient extends BaseClient
+{
+    public function scrape(string $url): WebContentResultsDto
+    {
+        $results = Browsershot::url($url)
+            ->userAgent('DailyAI Studio Browser 1.0, helping users automate workflows')
+            ->dismissDialogs()
+            ->fullPage();
+
+        $plainResults = $this->parseHtml($results->bodyHtml());
+
+        return WebContentResultsDto::from([
+            'title' => str($plainResults)->limit(128)->title()->toString(),
+            'description' => str($plainResults)->limit(256)->title()->toString(),
+            'content' => $plainResults,
+            'content_raw' => $results->bodyHtml(),
+            'url' => $url,
+            'browserShot' => $results,
+        ]);
+    }
+
+    public function parseHtml(string $html): string
+    {
+        $environment = new Environment([
+            'strip_tags' => true,
+            'suppress_errors' => true,
+            'hard_break' => true,
+            'strip_placeholder_links' => true,
+            'remove_nodes' => 'nav footer header script style meta',
+        ]);
+        $environment->addConverter(new TableConverter());
+        $environment->addConverter(new CodeConverter());
+        $environment->addConverter(new PreformattedConverter());
+        $environment->addConverter(new TextConverter());
+
+        $converter = new HtmlConverter($environment);
+
+        $markdown = $converter->convert($html);
+
+        return str($markdown)->trim()->toString();
+
+    }
+}

app/Domains/WebParser/FireCrawlClient.php

@@ -0,0 +1,36 @@
+<?php
+
+namespace App\Domains\WebParser;
+
+use App\Domains\WebParser\Results\FireCrawResultsDto;
+use App\Models\Setting;
+use Illuminate\Http\Client\PendingRequest;
+use Illuminate\Support\Facades\Http;
+
+class FireCrawlClient extends BaseWebParserClient
+{
+    public function scrape(string $url): WebContentResultsDto
+    {
+        $results = $this->getClient()->post('/scrape', [
+            'url' => $url,
+        ]);
+
+        if ($results->failed()) {
+            throw new \Exception('FireCrawl API Error '.$results->json());
+        }
+
+        $data = $results->json();
+
+        return FireCrawResultsDto::from($data);
+    }
+
+    protected function getClient(): PendingRequest
+    {
+        $url = Setting::getSecret('fire_crawl', 'api_url');
+        $token = Setting::getSecret('fire_crawl', 'api_key');
+
+        return Http::baseUrl($url)->withHeaders([
+            'Authorization' => 'Bearer '.$token,
+        ]);
+    }
+}

app/Domains/WebParser/Results/FireCrawResultsDto.php

@@ -0,0 +1,24 @@
+<?php
+
+namespace App\Domains\WebParser\Results;
+
+use App\Domains\WebParser\WebContentResultsDto;
+use Spatie\LaravelData\Attributes\MapInputName;
+
+class FireCrawResultsDto extends WebContentResultsDto
+{
+    public function __construct(
+        #[MapInputName('data.metadata.title')]
+        public string $title,
+        #[MapInputName('data.markdown')]
+        public string $content,
+        #[MapInputName('data.content')]
+        public string $content_raw,
+        #[MapInputName('data.metadata.sourceURL')]
+        public string $url,
+        #[MapInputName('data.metadata.description')]
+        public string $description = '',
+    ) {
+
+    }
+}

app/Domains/WebParser/WebContentResultsDto.php

@@ -0,0 +1,19 @@
+<?php
+
+namespace App\Domains\WebParser;
+
+use Spatie\Browsershot\Browsershot;
+use Spatie\LaravelData\Data;
+
+class WebContentResultsDto extends Data
+{
+    public function __construct(
+        public string $title,
+        public string $content,
+        public string $url,
+        public string $description = '',
+        public ?Browsershot $browserShot = null,
+    ) {
+
+    }
+}

app/Http/Controllers/SettingController.php

@@ -52,6 +52,23 @@ public function updateOllama(Request $request, Setting $setting)
         return back();
     }
 
+    public function updateFireCrawl(Request $request, Setting $setting)
+    {
+        $validated = $request->validate([
+            'api_key' => 'string|required',
+            'api_url' => 'string|required',
+        ]);
+
+        $secrets = $setting->secrets;
+        $secrets['fire_crawl'] = $validated;
+        $setting->secrets = $secrets;
+        $setting->save();
+        $setting->updateStep($setting);
+        $this->clearCache();
+
+        return back();
+    }
+
     public function updateGroq(Request $request, Setting $setting)
     {
         $validated = $request->validate([

app/Jobs/GetWebContentJob.php

@@ -73,7 +73,9 @@ public function handle(): void
             ->handle($this->webResponseDto->url, true);
 
         $prompt = Templatizer::appendContext(true)
-            ->handle($this->source->getPrompt(), $htmlResults);
+            ->handle($this->source->getPrompt(), $htmlResults->content);
+
+        put_fixture('web_page_prompt.txt', $prompt, false);
 
         $results = LlmDriverFacade::driver(
             $this->source->getDriver()
@@ -112,7 +114,7 @@ public function handle(): void
                         'source_id' => $this->source->id,
                         'type' => TypesEnum::HTML,
                         'subject' => to_utf8($title),
-                        'document_md5' => md5($htmlResults),
+                        'document_md5' => md5($htmlResults->content),
                         'link' => $this->webResponseDto->url,
                         'collection_id' => $this->source->collection_id,
                     ],
@@ -121,7 +123,7 @@ public function handle(): void
                         'file_path' => $this->webResponseDto->url,
                         'status_summary' => StatusEnum::Pending,
                         'meta_data' => $this->webResponseDto->toArray(),
-                        'original_content' => $htmlResults,
+                        'original_content' => $htmlResults->content,
                     ]
                 );
 

app/Models/Message.php

@@ -223,11 +223,12 @@ public function run(): void
         $meta_data->driver = $chat->getDriver();
         $message->updateQuietly(['meta_data' => $meta_data]);
 
-        if ($message->meta_data?->tool === 'completion') {
+        if ($message->meta_data?->tool === 'chat') {
             Log::info('[LaraChain] Running Simple Completion');
 
             $messages = $chat->getChatResponse();
-            $response = LlmDriverFacade::driver($chat->getDriver())->chat($messages);
+            $response = LlmDriverFacade::driver($chat->getDriver())
+                ->chat($messages);
             $response = $response->content;
 
             $chat->addInput(

database/migrations/2024_08_06_002122_add_fields_to_settings.php

@@ -0,0 +1,30 @@
+<?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('settings', function (Blueprint $table) {
+            $table->longText('main_prompt')->nullable();
+            $table->longText('source_prompt')->nullable();
+            $table->longText('output_prompt')->nullable();
+        });
+    }
+
+    /**
+     * Reverse the migrations.
+     */
+    public function down(): void
+    {
+        Schema::table('settings', function (Blueprint $table) {
+            //
+        });
+    }
+};

resources/js/Components/ScrollButton.vue

@@ -0,0 +1,22 @@
+<template>
+    <button @click="scrollToBottom"
+
+            class="fixed top-4 right-4 btn btn-secondary btn-circle text-white">
+        <svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke-width="3" stroke="currentColor" class="size-6">
+            <path stroke-linecap="round" stroke-linejoin="round" d="M19.5 13.5 12 21m0 0-7.5-7.5M12 21V3" />
+        </svg>
+
+    </button>
+</template>
+
+<script setup>
+import { ref } from 'vue'
+
+const bottomTarget = ref(null)
+
+const scrollToBottom = () => {
+    bottomTarget.value?.scrollIntoView({ behavior: 'smooth' })
+}
+
+defineExpose({ bottomTarget })
+</script>