ready to merge

LlmLaraHub · Mar 29, 2024 · 69fdbc3 · 69fdbc3
1 parent a8ff8c2
commit 69fdbc3
Show file tree

Hide file tree

Showing 12 changed files with 148 additions and 54 deletions.
diff --git a/README.md b/README.md
@@ -27,3 +27,12 @@ And just thinking how to make some of those flows work in a Laravel environment
 
 Per the Laravel docs https://laravel.com/docs/11.x/reverb
 
+
+
+## Local Dev
+
+```bash
+php artisan horizon:watch
+php artisan reverb:start --debug
+npm run dev
+```
diff --git a/app/Domains/Messages/SearchOrSummarizeChatRepo.php b/app/Domains/Messages/SearchOrSummarizeChatRepo.php
@@ -23,7 +23,7 @@ public function search(Chat $chat, string $input): string
 
         /** @var EmbeddingsResponseDto $embedding */
         $embedding = LlmDriverFacade::driver(
-            $chat->chatable->getEmddingDriver()
+            $chat->chatable->getEmbeddingDriver()
         )->embedData($input);
 
         $results = DocumentChunk::query()
@@ -40,12 +40,12 @@ public function search(Chat $chat, string $input): string
         $content = [];
 
         foreach ($results as $result) {
-            $content[] = reduce_text_size($result->content);
+            $content[] = remove_ascii(reduce_text_size($result->content)); //reduce_text_size seem to mess up Claude?
         }
 
         $content = implode(' ', $content);
 
-        $content = 'This is data from the search results when entering the users prompt please use this for context and only this: '.$content;
+        $content = 'This is data from the search results when entering the users prompt please use this for context and only this and return as markdown so I can render it: '.$content;
 
         $chat->addInput(
             message: $content,

diff --git a/app/LlmDriver/ClaudeClient.php b/app/LlmDriver/ClaudeClient.php
@@ -11,7 +11,6 @@
 
 class ClaudeClient
 {
-
     protected string $baseUrl = 'https://api.anthropic.com/v1';
 
     protected string $version = '2023-06-01';
@@ -20,48 +19,66 @@ class ClaudeClient
 
     public function embedData(string $data): EmbeddingsResponseDto
     {
-
-        Log::info('LlmDriver::ClaudeClient::embedData');
-
-        return EmbeddingsResponseDto::from([
-            'embedding' => data_get($data, 'data.0.embedding'),
-            'token_count' => 1000,
-        ]);
+        throw new \Exception('Not implemented');
     }
 
     /**
      * @param  MessageInDto[]  $messages
      */
     public function chat(array $messages): CompletionResponse
     {
-        $model = $this->getConfig('claude')['models']['completion_model']; 
-        $maxTokens = $this->getConfig('claude')['max_tokens']; 
+        $model = $this->getConfig('claude')['models']['completion_model'];
+        $maxTokens = $this->getConfig('claude')['max_tokens'];
 
         Log::info('LlmDriver::Claude::completion');
 
-        $messages = collect($messages)->map(function($item) {
-            if($item->role === 'system') {
+        /**
+         * I need to iterate over each item
+         * then if there are two rows with role assistant I need to insert
+         * in betwee a user row with some copy to make it work like "And the user search results had"
+         * using the Laravel Collection library
+         */
+        $messages = collect($messages)->map(function ($item) {
+            if ($item->role === 'system') {
                 $item->role = 'assistant';
             }
 
             return $item->toArray();
-        })->reverse()->values()->all();
-
+        })->reverse()->values();
+
+        $messages = $messages->flatMap(function ($item, $index) use ($messages) {
+            if ($index > 0 && $item['role'] === 'assistant' && optional($messages->get($index + 1))['role'] === 'assistant') {
+                return [
+                    $item,
+                    ['role' => 'user', 'content' => 'Continuation of search results'],
+                ];
+            }
+
+            return [$item];
+        })->toArray();
+
+        put_fixture('claude_messages_debug.json', $messages);
+
         $results = $this->getClient()->post('/messages', [
             'model' => $model,
-            "max_tokens" => $maxTokens,
+            'system' => 'Return a markdown response.',
+            'max_tokens' => $maxTokens,
             'messages' => $messages,
         ]);
 
-        if(!$results->ok()) {
+        if (! $results->ok()) {
             $error = $results->json()['error']['type'];
-            Log::error('Claude API Error ' . $error);
-            throw new \Exception('Claude API Error ' . $error);
+            $message = $results->json()['error']['message'];
+            Log::error('Claude API Error ', [
+                'type' => $error,
+                'message' => $message,
+            ]);
+            throw new \Exception('Claude API Error '.$message);
         }
 
         $data = null;
 
-        foreach($results->json()['content'] as $content) {
+        foreach ($results->json()['content'] as $content) {
             $data = $content['text'];
         }
 
@@ -72,32 +89,31 @@ public function chat(array $messages): CompletionResponse
 
     public function completion(string $prompt): CompletionResponse
     {
-        $model = $this->getConfig('claude')['models']['completion_model']; 
-        $maxTokens = $this->getConfig('claude')['max_tokens']; 
+        $model = $this->getConfig('claude')['models']['completion_model'];
+        $maxTokens = $this->getConfig('claude')['max_tokens'];
 
         Log::info('LlmDriver::Claude::completion');
 
         $results = $this->getClient()->post('/messages', [
             'model' => $model,
-            "max_tokens" => $maxTokens,
+            'max_tokens' => $maxTokens,
             'messages' => [
                 [
-                    'role' => "user",
-                    'content' => $prompt
-                ]
+                    'role' => 'user',
+                    'content' => $prompt,
+                ],
             ],
         ]);
 
-
-        if(!$results->ok()) {
+        if (! $results->ok()) {
             $error = $results->json()['error']['type'];
-            Log::error('Claude API Error ' . $error);
-            throw new \Exception('Claude API Error ' . $error);
+            Log::error('Claude API Error '.$error);
+            throw new \Exception('Claude API Error '.$error);
         }
 
         $data = null;
 
-        foreach($results->json()['content'] as $content) {
+        foreach ($results->json()['content'] as $content) {
             $data = $content['text'];
         }
 
@@ -106,13 +122,15 @@ public function completion(string $prompt): CompletionResponse
         ]);
     }
 
-    protected function getError(Response $response) {
+    protected function getError(Response $response)
+    {
         return $response->json()['error']['type'];
     }
 
-    protected function getClient() {
+    protected function getClient()
+    {
         $api_token = $this->getConfig('claude')['api_key'];
-        if(!$api_token) {
+        if (! $api_token) {
             throw new \Exception('Claude API Token not found');
         }
 

diff --git a/app/LlmDriver/DriversEnum.php b/app/LlmDriver/DriversEnum.php
@@ -1,13 +1,13 @@
-<?php 
+<?php
 
 namespace App\LlmDriver;
 
-enum DriversEnum : string {
-
+enum DriversEnum: string
+{
     case Mock = 'mock';
     case OpenAi = 'openai';
     case OpenAiAzure = 'openai_azure';
     case Ollama = 'ollama';
     case Gemini = 'gemini';
     case Claude = 'claude';
-}
+}
diff --git a/app/LlmDriver/HasDrivers.php b/app/LlmDriver/HasDrivers.php
@@ -1,10 +1,10 @@
-<?php 
+<?php
 
 namespace App\LlmDriver;
 
-
 interface HasDrivers
 {
     public function getDriver(): string;
+
     public function getEmbeddingDriver(): string;
-}
+}
diff --git a/app/Models/Collection.php b/app/Models/Collection.php
@@ -2,13 +2,13 @@
 
 namespace App\Models;
 
+use App\LlmDriver\DriversEnum;
+use App\LlmDriver\HasDrivers;
 use Illuminate\Database\Eloquent\Factories\HasFactory;
 use Illuminate\Database\Eloquent\Model;
 use Illuminate\Database\Eloquent\Relations\BelongsTo;
 use Illuminate\Database\Eloquent\Relations\HasMany;
 use Illuminate\Database\Eloquent\Relations\MorphMany;
-use App\LlmDriver\DriversEnum;
-use App\LlmDriver\HasDrivers;
 
 /**
  * Class Project
@@ -32,10 +32,9 @@ class Collection extends Model implements HasDrivers
     protected $casts = [
         'active' => 'boolean',
         'driver' => DriversEnum::class,
-        'embedding_driver' => DriversEnum::class
+        'embedding_driver' => DriversEnum::class,
     ];
 
-
     public function team(): BelongsTo
     {
         return $this->belongsTo(Team::class);

diff --git a/app/Models/Document.php b/app/Models/Document.php
@@ -30,7 +30,6 @@ class Document extends Model implements HasDrivers
         'summary_status' => StatusEnum::class,
     ];
 
-
     public function collection(): BelongsTo
     {
         return $this->belongsTo(Collection::class);
@@ -63,8 +62,8 @@ public function getDriver(): string
         return $this->collection->driver->value;
     }
 
-
-    public function getEmbeddingDriver(): string { 
+    public function getEmbeddingDriver(): string
+    {
         return $this->collection->embedding_driver->value;
     }
 }
diff --git a/app/helpers.php b/app/helpers.php
@@ -25,7 +25,7 @@ function put_fixture($file_name, $content = [], $json = true)
 if (! function_exists('remove_ascii')) {
     function remove_ascii($string): string
     {
-        return preg_replace('/[^\x00-\x7F]+/', '', $string);
+        return str_replace("\u2019", ' ', preg_replace('/[^\x00-\x7F]+/', '', $string));
     }
 }
 

diff --git a/config/llmdriver.php b/config/llmdriver.php
@@ -23,7 +23,7 @@
             'models' => [
                 //@see https://www.anthropic.com/news/claude-3-family
                 'completion_model' => env('CLAUDE_COMPLETION_MODEL', 'claude-3-opus-20240229'),
-            ]
+            ],
         ],
         'ollama' => [
 

diff --git a/tests/Feature/ClaudeClientTest.php b/tests/Feature/ClaudeClientTest.php
@@ -3,7 +3,6 @@
 namespace Tests\Feature;
 
 use App\LlmDriver\ClaudeClient;
-use App\LlmDriver\MockClient;
 use App\LlmDriver\Requests\MessageInDto;
 use App\LlmDriver\Responses\CompletionResponse;
 use App\LlmDriver\Responses\EmbeddingsResponseDto;
@@ -18,6 +17,8 @@ class ClaudeClientTest extends TestCase
     public function test_embeddings(): void
     {
 
+        $this->markTestSkipped('@TODO: Requires another server');
+
         $client = new ClaudeClient();
 
         $results = $client->embedData('test');
@@ -68,9 +69,52 @@ public function test_chat(): void
         Http::assertSent(function ($request) {
             $message1 = $request->data()['messages'][0]['role'];
             $message2 = $request->data()['messages'][1]['role'];
+
             return $message2 === 'assistant' &&
                 $message1 === 'user';
         });
 
     }
+
+    public function test_chat_with_multiple_assistant_messages(): void
+    {
+        $client = new ClaudeClient();
+
+        $data = get_fixture('claude_completion.json');
+
+        Http::fake([
+            'api.anthropic.com/*' => Http::response($data, 200),
+        ]);
+
+        $results = $client->chat([
+            MessageInDto::from([
+                'content' => 'test 3',
+                'role' => 'assistant',
+            ]),
+            MessageInDto::from([
+                'content' => 'test 2',
+                'role' => 'assistant',
+            ]),
+            MessageInDto::from([
+                'content' => 'test 1',
+                'role' => 'assistant',
+            ]),
+            MessageInDto::from([
+                'content' => 'test',
+                'role' => 'user',
+            ]),
+        ]);
+
+        $this->assertInstanceOf(CompletionResponse::class, $results);
+
+        Http::assertSent(function ($request) {
+            $message1 = $request->data()['messages'][0]['role'];
+            $message2 = $request->data()['messages'][1]['role'];
+            $message3 = $request->data()['messages'][1]['role'];
+
+            return $message2 === 'assistant' &&
+                $message1 === 'user' && $message3 === 'assistant';
+        });
+
+    }
 }
diff --git a/tests/Feature/Http/Controllers/CollectionControllerTest.php b/tests/Feature/Http/Controllers/CollectionControllerTest.php
@@ -10,7 +10,6 @@
 use Illuminate\Support\Facades\Queue;
 use Illuminate\Support\Facades\Storage;
 use Inertia\Testing\AssertableInertia as Assert;
-use Laravel\Pennant\Contracts\Driver;
 use Tests\TestCase;
 
 class CollectionControllerTest extends TestCase

diff --git a/tests/fixtures/claude_messages_debug.json b/tests/fixtures/claude_messages_debug.json
@@ -0,0 +1,26 @@
+[
+    {
+        "content": "test",
+        "role": "user"
+    },
+    {
+        "content": "test 1",
+        "role": "assistant"
+    },
+    {
+        "role": "user",
+        "content": "Continuation of search results"
+    },
+    {
+        "content": "test 2",
+        "role": "assistant"
+    },
+    {
+        "role": "user",
+        "content": "Continuation of search results"
+    },
+    {
+        "content": "test 3",
+        "role": "assistant"
+    }
+]