From 66f90e9aca486e0c709aeb44efc3f36b2b8a0016 Mon Sep 17 00:00:00 2001
From: Christopher Hertel <mail@christopher-hertel.de>
Date: Sat, 21 Sep 2024 23:26:40 +0200
Subject: [PATCH] feat: adding youtube transcription tool

---
 README.md                               |  13 ++
 composer.json                           |   6 +-
 composer.lock                           | 201 +++++++++++++++++++++++-
 examples/toolchain-youtube.php          |  28 ++++
 src/ToolBox/Tool/YouTubeTranscriber.php |  67 ++++++++
 5 files changed, 313 insertions(+), 2 deletions(-)
 create mode 100755 examples/toolchain-youtube.php
 create mode 100644 src/ToolBox/Tool/YouTubeTranscriber.php

diff --git a/README.md b/README.md
index 0a353e7..c9b101a 100644
--- a/README.md
+++ b/README.md
@@ -54,6 +54,7 @@ Provided Tools
 * [x] Clock
 * [x] Wikipedia
 * [x] Weather
+* [x] YouTube Transcriber
 
 Usage Examples
 --------------
@@ -105,6 +106,18 @@ Depending on the example you need to export needed environment variables for API
    php examples/toolchain-serpapi.php
    ```
 
+1. Weather Tool
+   ```bash
+   export OPENAI_API_KEY=sk-...
+   php examples/toolchain-weather.php
+   ```
+
+1. YouTube Transcriber Tool
+   ```bash
+   export OPENAI_API_KEY=sk-...
+   php examples/toolchain-youtube.php
+   ```
+
 ### Structured Output
 
 1. Structured Output Example: OpenAI's GPT
diff --git a/composer.json b/composer.json
index aa4eb17..0b18e6b 100644
--- a/composer.json
+++ b/composer.json
@@ -29,12 +29,16 @@
         "probots-io/pinecone-php": "^1.0",
         "symfony/clock": "^6.4 || ^7.1",
         "symfony/console": "^6.4 || ^7.1",
+        "symfony/css-selector": "^6.4 || ^7.1",
+        "symfony/dom-crawler": "^6.4 || ^7.1",
         "symfony/var-dumper": "^6.4 || ^7.1"
     },
     "suggest": {
         "codewithkyrian/chromadb-php": "For using the ChromaDB as retrieval vector store.",
         "probots-io/pinecone-php": "For using the Pinecone as retrieval vector store.",
-        "symfony/clock": "For using the clock tool."
+        "symfony/clock": "For using the clock tool.",
+        "symfony/css-selector": "For using the YouTube transcription tool.",
+        "symfony/dom-crawler": "For using the YouTube transcription tool."
     },
     "config": {
         "sort-packages": true
diff --git a/composer.lock b/composer.lock
index c6e68e2..b08c2fd 100644
--- a/composer.lock
+++ b/composer.lock
@@ -4,7 +4,7 @@
         "Read more about it at https://getcomposer.org/doc/01-basic-usage.md#installing-dependencies",
         "This file is @generated automatically"
     ],
-    "content-hash": "71fd4420733a7abfe8b647efc5bac701",
+    "content-hash": "317e154b61bb699eae1c61f5bac6038d",
     "packages": [
         {
             "name": "doctrine/deprecations",
@@ -2090,6 +2090,73 @@
             ],
             "time": "2024-07-18T11:15:46+00:00"
         },
+        {
+            "name": "masterminds/html5",
+            "version": "2.9.0",
+            "source": {
+                "type": "git",
+                "url": "https://github.com/Masterminds/html5-php.git",
+                "reference": "f5ac2c0b0a2eefca70b2ce32a5809992227e75a6"
+            },
+            "dist": {
+                "type": "zip",
+                "url": "https://api.github.com/repos/Masterminds/html5-php/zipball/f5ac2c0b0a2eefca70b2ce32a5809992227e75a6",
+                "reference": "f5ac2c0b0a2eefca70b2ce32a5809992227e75a6",
+                "shasum": ""
+            },
+            "require": {
+                "ext-dom": "*",
+                "php": ">=5.3.0"
+            },
+            "require-dev": {
+                "phpunit/phpunit": "^4.8.35 || ^5.7.21 || ^6 || ^7 || ^8 || ^9"
+            },
+            "type": "library",
+            "extra": {
+                "branch-alias": {
+                    "dev-master": "2.7-dev"
+                }
+            },
+            "autoload": {
+                "psr-4": {
+                    "Masterminds\\": "src"
+                }
+            },
+            "notification-url": "https://packagist.org/downloads/",
+            "license": [
+                "MIT"
+            ],
+            "authors": [
+                {
+                    "name": "Matt Butcher",
+                    "email": "technosophos@gmail.com"
+                },
+                {
+                    "name": "Matt Farina",
+                    "email": "matt@mattfarina.com"
+                },
+                {
+                    "name": "Asmir Mustafic",
+                    "email": "goetas@gmail.com"
+                }
+            ],
+            "description": "An HTML5 parser and serializer.",
+            "homepage": "http://masterminds.github.io/html5-php",
+            "keywords": [
+                "HTML5",
+                "dom",
+                "html",
+                "parser",
+                "querypath",
+                "serializer",
+                "xml"
+            ],
+            "support": {
+                "issues": "https://github.com/Masterminds/html5-php/issues",
+                "source": "https://github.com/Masterminds/html5-php/tree/2.9.0"
+            },
+            "time": "2024-03-31T07:05:07+00:00"
+        },
         {
             "name": "myclabs/deep-copy",
             "version": "1.12.0",
@@ -4340,6 +4407,138 @@
             ],
             "time": "2024-09-20T08:28:38+00:00"
         },
+        {
+            "name": "symfony/css-selector",
+            "version": "v7.1.1",
+            "source": {
+                "type": "git",
+                "url": "https://github.com/symfony/css-selector.git",
+                "reference": "1c7cee86c6f812896af54434f8ce29c8d94f9ff4"
+            },
+            "dist": {
+                "type": "zip",
+                "url": "https://api.github.com/repos/symfony/css-selector/zipball/1c7cee86c6f812896af54434f8ce29c8d94f9ff4",
+                "reference": "1c7cee86c6f812896af54434f8ce29c8d94f9ff4",
+                "shasum": ""
+            },
+            "require": {
+                "php": ">=8.2"
+            },
+            "type": "library",
+            "autoload": {
+                "psr-4": {
+                    "Symfony\\Component\\CssSelector\\": ""
+                },
+                "exclude-from-classmap": [
+                    "/Tests/"
+                ]
+            },
+            "notification-url": "https://packagist.org/downloads/",
+            "license": [
+                "MIT"
+            ],
+            "authors": [
+                {
+                    "name": "Fabien Potencier",
+                    "email": "fabien@symfony.com"
+                },
+                {
+                    "name": "Jean-François Simon",
+                    "email": "jeanfrancois.simon@sensiolabs.com"
+                },
+                {
+                    "name": "Symfony Community",
+                    "homepage": "https://symfony.com/contributors"
+                }
+            ],
+            "description": "Converts CSS selectors to XPath expressions",
+            "homepage": "https://symfony.com",
+            "support": {
+                "source": "https://github.com/symfony/css-selector/tree/v7.1.1"
+            },
+            "funding": [
+                {
+                    "url": "https://symfony.com/sponsor",
+                    "type": "custom"
+                },
+                {
+                    "url": "https://github.com/fabpot",
+                    "type": "github"
+                },
+                {
+                    "url": "https://tidelift.com/funding/github/packagist/symfony/symfony",
+                    "type": "tidelift"
+                }
+            ],
+            "time": "2024-05-31T14:57:53+00:00"
+        },
+        {
+            "name": "symfony/dom-crawler",
+            "version": "v7.1.5",
+            "source": {
+                "type": "git",
+                "url": "https://github.com/symfony/dom-crawler.git",
+                "reference": "b92af238457a7cdd2738f941cd525d76313e8283"
+            },
+            "dist": {
+                "type": "zip",
+                "url": "https://api.github.com/repos/symfony/dom-crawler/zipball/b92af238457a7cdd2738f941cd525d76313e8283",
+                "reference": "b92af238457a7cdd2738f941cd525d76313e8283",
+                "shasum": ""
+            },
+            "require": {
+                "masterminds/html5": "^2.6",
+                "php": ">=8.2",
+                "symfony/polyfill-ctype": "~1.8",
+                "symfony/polyfill-mbstring": "~1.0"
+            },
+            "require-dev": {
+                "symfony/css-selector": "^6.4|^7.0"
+            },
+            "type": "library",
+            "autoload": {
+                "psr-4": {
+                    "Symfony\\Component\\DomCrawler\\": ""
+                },
+                "exclude-from-classmap": [
+                    "/Tests/"
+                ]
+            },
+            "notification-url": "https://packagist.org/downloads/",
+            "license": [
+                "MIT"
+            ],
+            "authors": [
+                {
+                    "name": "Fabien Potencier",
+                    "email": "fabien@symfony.com"
+                },
+                {
+                    "name": "Symfony Community",
+                    "homepage": "https://symfony.com/contributors"
+                }
+            ],
+            "description": "Eases DOM navigation for HTML and XML documents",
+            "homepage": "https://symfony.com",
+            "support": {
+                "source": "https://github.com/symfony/dom-crawler/tree/v7.1.5"
+            },
+            "funding": [
+                {
+                    "url": "https://symfony.com/sponsor",
+                    "type": "custom"
+                },
+                {
+                    "url": "https://github.com/fabpot",
+                    "type": "github"
+                },
+                {
+                    "url": "https://tidelift.com/funding/github/packagist/symfony/symfony",
+                    "type": "tidelift"
+                }
+            ],
+            "time": "2024-09-15T06:48:17+00:00"
+        },
         {
             "name": "symfony/polyfill-php83",
             "version": "v1.31.0",
diff --git a/examples/toolchain-youtube.php b/examples/toolchain-youtube.php
new file mode 100755
index 0000000..28c5ab9
--- /dev/null
+++ b/examples/toolchain-youtube.php
@@ -0,0 +1,28 @@
+<?php
+
+use PhpLlm\LlmChain\Chain;
+use PhpLlm\LlmChain\Message\Message;
+use PhpLlm\LlmChain\Message\MessageBag;
+use PhpLlm\LlmChain\OpenAI\Model\Gpt;
+use PhpLlm\LlmChain\OpenAI\Model\Gpt\Version;
+use PhpLlm\LlmChain\OpenAI\Runtime\OpenAI;
+use PhpLlm\LlmChain\ToolBox\ParameterAnalyzer;
+use PhpLlm\LlmChain\ToolBox\Registry;
+use PhpLlm\LlmChain\ToolBox\Tool\YouTubeTranscriber;
+use PhpLlm\LlmChain\ToolBox\ToolAnalyzer;
+use Symfony\Component\HttpClient\HttpClient;
+
+require_once dirname(__DIR__).'/vendor/autoload.php';
+
+$httpClient = HttpClient::create();
+$runtime = new OpenAI($httpClient, getenv('OPENAI_API_KEY'));
+$llm = new Gpt($runtime, Version::GPT_4o_MINI);
+
+$transcriber = new YouTubeTranscriber($httpClient);
+$registry = new Registry(new ToolAnalyzer(new ParameterAnalyzer()), [$transcriber]);
+$chain = new Chain($llm, $registry);
+
+$messages = new MessageBag(Message::ofUser('Please summarize this video for me: https://www.youtube.com/watch?v=6uXW-ulpj0s'));
+$response = $chain->call($messages);
+
+echo $response.PHP_EOL;
diff --git a/src/ToolBox/Tool/YouTubeTranscriber.php b/src/ToolBox/Tool/YouTubeTranscriber.php
new file mode 100644
index 0000000..60abce6
--- /dev/null
+++ b/src/ToolBox/Tool/YouTubeTranscriber.php
@@ -0,0 +1,67 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpLlm\LlmChain\ToolBox\Tool;
+
+use PhpLlm\LlmChain\ToolBox\AsTool;
+use Symfony\Component\CssSelector\CssSelectorConverter;
+use Symfony\Component\DomCrawler\Crawler;
+use Symfony\Contracts\HttpClient\HttpClientInterface;
+
+#[AsTool('youtube_transcript', 'Fetches the transcript of a YouTube video')]
+final readonly class YouTubeTranscriber
+{
+    public function __construct(
+        private HttpClientInterface $client,
+    ) {
+        if (!class_exists(Crawler::class)) {
+            throw new \LogicException('The Symfony DomCrawler component is required to use this tool.');
+        }
+        if (!class_exists(CssSelectorConverter::class)) {
+            throw new \LogicException('The Symfony CSS Selector component is required to use this tool.');
+        }
+    }
+
+    /**
+     * @param string $videoId The ID of the YouTube video
+     */
+    public function __invoke(string $videoId): string
+    {
+        // Fetch the HTML content of the YouTube video page
+        $htmlResponse = $this->client->request('GET', 'https://youtube.com/watch?v='.$videoId);
+        $html = $htmlResponse->getContent();
+
+        // Use DomCrawler to parse the HTML
+        $crawler = new Crawler($html);
+
+        // Extract the script containing the ytInitialPlayerResponse
+        $scriptContent = $crawler->filter('script')->reduce(function (Crawler $node) {
+            return str_contains($node->text(), 'var ytInitialPlayerResponse = {');
+        })->text();
+
+        // Extract and parse the JSON data from the script
+        $start = strpos($scriptContent, 'var ytInitialPlayerResponse = ') + strlen('var ytInitialPlayerResponse = ');
+        $dataString = substr($scriptContent, $start);
+        $dataString = substr($dataString, 0, strrpos($dataString, ';') ?: null);
+        $data = json_decode(trim($dataString), true);
+
+        // Extract the URL for the captions
+        if (!isset($data['captions']['playerCaptionsTracklistRenderer']['captionTracks'][0]['baseUrl'])) {
+            throw new \Exception('Captions are not available for this video.');
+        }
+        $captionsUrl = $data['captions']['playerCaptionsTracklistRenderer']['captionTracks'][0]['baseUrl'];
+
+        // Fetch and parse the captions XML
+        $xmlResponse = $this->client->request('GET', $captionsUrl);
+        $xmlContent = $xmlResponse->getContent();
+        $xmlCrawler = new Crawler($xmlContent);
+
+        // Collect all text elements from the captions
+        $transcript = $xmlCrawler->filter('text')->each(function (Crawler $node) {
+            return $node->text().' ';
+        });
+
+        return implode(PHP_EOL, $transcript);
+    }
+}