Embedding Foundations & OpenAI Embeddings (#70)

Author: sixlive

Diff for: docs/.vitepress/config.mts

@@ -102,6 +102,10 @@ export default defineConfig({
                 text: "Tool & Function Calling",
                 link: "/core-concepts/tools-function-calling",
               },
+              {
+                text: "Embeddings",
+                link: "/core-concepts/embeddings",
+              },
               {
                 text: "Prism Server",
                 link: "/core-concepts/prism-server",

Diff for: docs/core-concepts/embeddings.md

@@ -0,0 +1,107 @@
+# Embeddings
+
+Transform your text into powerful vector representations! Embeddings let you add semantic search, recommendation systems, and other advanced natural language features to your applications.
+
+## Quick Start
+
+Here's how to generate embeddings with just a few lines of code:
+
+```php
+use EchoLabs\Prism\Facades\Prism;
+use EchoLabs\Prism\Enums\Provider;
+
+$response = Prism::embeddings()
+    ->using(Provider::OpenAI, 'text-embedding-3-large')
+    ->fromInput('Your text goes here')
+    ->generate();
+
+// Get your embeddings vector
+$embeddings = $response->embeddings;
+
+// Check token usage
+echo $response->usage->tokens;
+```
+
+## Input Methods
+
+You've got two convenient ways to feed text into the embeddings generator:
+
+### Direct Text Input
+
+```php
+$response = Prism::embeddings()
+    ->using(Provider::OpenAI, 'text-embedding-3-large')
+    ->fromInput('Analyze this text')
+    ->generate();
+```
+
+### From File
+
+Need to analyze a larger document? No problem:
+
+```php
+$response = Prism::embeddings()
+    ->using(Provider::OpenAI, 'text-embedding-3-large')
+    ->fromFile('/path/to/your/document.txt')
+    ->generate();
+```
+
+> [!NOTE]
+> Make sure your file exists and is readable. The generator will throw a helpful `PrismException` if there's any issue accessing the file.
+
+## Common Settings
+
+Just like with text generation, you can fine-tune your embeddings requests:
+
+```php
+$response = Prism::embeddings()
+    ->using(Provider::OpenAI, 'text-embedding-3-large')
+    ->fromInput('Your text here')
+    ->withClientOptions(['timeout' => 30]) // Adjust request timeout
+    ->withClientRetry(3, 100) // Add automatic retries
+    ->generate();
+```
+
+## Response Handling
+
+The embeddings response gives you everything you need:
+
+```php
+// Get the embeddings vector
+$vector = $response->embeddings;
+
+// Check token usage
+$tokenCount = $response->usage->tokens;
+```
+
+## Error Handling
+
+Always handle potential errors gracefully:
+
+```php
+use EchoLabs\Prism\Exceptions\PrismException;
+
+try {
+    $response = Prism::embeddings()
+        ->using(Provider::OpenAI, 'text-embedding-3-large')
+        ->fromInput('Your text here')
+        ->generate();
+} catch (PrismException $e) {
+    Log::error('Embeddings generation failed:', [
+        'error' => $e->getMessage()
+    ]);
+}
+```
+
+## Pro Tips 🌟
+
+**Vector Storage**: Consider using a vector database like Milvus, Qdrant, or pgvector to store and query your embeddings efficiently.
+
+**Text Preprocessing**: For best results, clean and normalize your text before generating embeddings. This might include:
+   - Removing unnecessary whitespace
+   - Converting to lowercase
+   - Removing special characters
+   - Handling Unicode normalization
+
+> [!IMPORTANT]
+> Different providers and models produce vectors of different dimensions. Always check your provider's documentation for specific details about the embedding model you're using.

Diff for: src/Contracts/Provider.php

@@ -4,6 +4,8 @@
 
 namespace EchoLabs\Prism\Contracts;
 
+use EchoLabs\Prism\Embeddings\Request as EmbeddingsRequest;
+use EchoLabs\Prism\Embeddings\Response as EmbeddingsResponse;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Request as StructuredRequest;
 use EchoLabs\Prism\Text\Request as TextRequest;
@@ -13,4 +15,6 @@ interface Provider
     public function text(TextRequest $request): ProviderResponse;
 
     public function structured(StructuredRequest $request): ProviderResponse;
+
+    public function embeddings(EmbeddingsRequest $request): EmbeddingsResponse;
 }

Diff for: src/Embeddings/Generator.php

@@ -0,0 +1,74 @@
+<?php
+
+declare(strict_types=1);
+
+namespace EchoLabs\Prism\Embeddings;
+
+use Closure;
+use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Enums\Provider as ProviderEnum;
+use EchoLabs\Prism\Exceptions\PrismException;
+use EchoLabs\Prism\PrismManager;
+
+class Generator
+{
+    protected string $input = '';
+
+    /** @var array<string, mixed> */
+    protected array $clientOptions = [];
+
+    /** @var array{0: array<int, int>|int, 1?: Closure|int, 2?: ?callable, 3?: bool} */
+    protected array $clientRetry = [0];
+
+    protected Provider $provider;
+
+    protected string $model;
+
+    public function using(string|ProviderEnum $provider, string $model): self
+    {
+        $this->provider = resolve(PrismManager::class)
+            ->resolve($provider);
+
+        $this->model = $model;
+
+        return $this;
+    }
+
+    public function fromInput(string $input): self
+    {
+        $this->input = $input;
+
+        return $this;
+    }
+
+    public function fromFile(string $path): self
+    {
+        if (! is_file($path)) {
+            throw new PrismException(sprintf('%s is not a valid file', $path));
+        }
+
+        $contents = file_get_contents($path);
+
+        if ($contents === false) {
+            throw new PrismException(sprintf('%s contents could not be read', $path));
+        }
+
+        $this->input = $contents;
+
+        return $this;
+    }
+
+    public function generate(): Response
+    {
+        if ($this->input === '' || $this->input === '0') {
+            throw new PrismException('Embeddings input is required');
+        }
+
+        return $this->provider->embeddings(new Request(
+            model: $this->model,
+            input: $this->input,
+            clientOptions: $this->clientOptions,
+            clientRetry: $this->clientRetry,
+        ));
+    }
+}

Diff for: src/Embeddings/Request.php

@@ -0,0 +1,21 @@
+<?php
+
+declare(strict_types=1);
+
+namespace EchoLabs\Prism\Embeddings;
+
+use Closure;
+
+class Request
+{
+    /**
+     * @param  array<string, mixed>  $clientOptions
+     * @param  array{0: array<int, int>|int, 1?: Closure|int, 2?: ?callable, 3?: bool}  $clientRetry
+     */
+    public function __construct(
+        public readonly string $model,
+        public readonly string $input,
+        public readonly array $clientOptions,
+        public readonly array $clientRetry,
+    ) {}
+}

Diff for: src/Embeddings/Response.php

@@ -0,0 +1,18 @@
+<?php
+
+declare(strict_types=1);
+
+namespace EchoLabs\Prism\Embeddings;
+
+use EchoLabs\Prism\ValueObjects\EmbeddingsUsage;
+
+class Response
+{
+    /**
+     * @param  array<int, int|string>  $embeddings
+     */
+    public function __construct(
+        public readonly array $embeddings,
+        public readonly EmbeddingsUsage $usage,
+    ) {}
+}

Diff for: src/Prism.php

@@ -5,6 +5,7 @@
 namespace EchoLabs\Prism;
 
 use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Embeddings\Generator as EmbeddingsGenerator;
 use EchoLabs\Prism\Enums\Provider as ProviderEnum;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Generator as StructuredGenerator;
@@ -44,4 +45,9 @@ public static function structured(): StructuredGenerator
     {
         return new StructuredGenerator;
     }
+
+    public static function embeddings(): \EchoLabs\Prism\Embeddings\Generator
+    {
+        return new EmbeddingsGenerator;
+    }
 }

Diff for: src/Providers/Anthropic/Anthropic.php

@@ -5,6 +5,8 @@
 namespace EchoLabs\Prism\Providers\Anthropic;
 
 use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Embeddings\Request as EmbeddingRequest;
+use EchoLabs\Prism\Embeddings\Response as EmbeddingResponse;
 use EchoLabs\Prism\Providers\Anthropic\Handlers\Text;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Request as StructuredRequest;
@@ -33,6 +35,12 @@ public function structured(StructuredRequest $request): ProviderResponse
         throw new \Exception(sprintf('%s does not support structured mode', class_basename($this)));
     }
 
+    #[\Override]
+    public function embeddings(EmbeddingRequest $request): EmbeddingResponse
+    {
+        throw new \Exception(sprintf('%s does not support embeddings', class_basename($this)));
+    }
+
     /**
      * @param  array<string, mixed>  $options
      * @param  array<mixed>  $retry

Diff for: src/Providers/Groq/Groq.php

@@ -5,6 +5,8 @@
 namespace EchoLabs\Prism\Providers\Groq;
 
 use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Embeddings\Request as EmbeddingRequest;
+use EchoLabs\Prism\Embeddings\Response as EmbeddingResponse;
 use EchoLabs\Prism\Providers\Groq\Handlers\Text;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Request as StructuredRequest;
@@ -33,6 +35,12 @@ public function structured(StructuredRequest $request): ProviderResponse
         throw new \Exception(sprintf('%s does not support structured mode', class_basename($this)));
     }
 
+    #[\Override]
+    public function embeddings(EmbeddingRequest $request): EmbeddingResponse
+    {
+        throw new \Exception(sprintf('%s does not support embeddings', class_basename($this)));
+    }
+
     /**
      * @param  array<string, mixed>  $options
      * @param  array<mixed>  $retry

Diff for: src/Providers/Mistral/Mistral.php

@@ -5,6 +5,8 @@
 namespace EchoLabs\Prism\Providers\Mistral;
 
 use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Embeddings\Request as EmbeddingRequest;
+use EchoLabs\Prism\Embeddings\Response as EmbeddingResponse;
 use EchoLabs\Prism\Providers\Mistral\Handlers\Text;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Request as StructuredRequest;
@@ -33,6 +35,12 @@ public function structured(StructuredRequest $request): ProviderResponse
         throw new \Exception(sprintf('%s does not support structured mode', class_basename($this)));
     }
 
+    #[\Override]
+    public function embeddings(EmbeddingRequest $request): EmbeddingResponse
+    {
+        throw new \Exception(sprintf('%s does not support embeddings', class_basename($this)));
+    }
+
     /**
      * @param  array<string, mixed>  $options
      * @param  array<mixed>  $retry

Diff for: src/Providers/Ollama/Ollama.php

@@ -5,6 +5,8 @@
 namespace EchoLabs\Prism\Providers\Ollama;
 
 use EchoLabs\Prism\Contracts\Provider;
+use EchoLabs\Prism\Embeddings\Request as EmbeddingRequest;
+use EchoLabs\Prism\Embeddings\Response as EmbeddingResponse;
 use EchoLabs\Prism\Providers\Ollama\Handlers\Text;
 use EchoLabs\Prism\Providers\ProviderResponse;
 use EchoLabs\Prism\Structured\Request as StructuredRequest;
@@ -33,6 +35,12 @@ public function structured(StructuredRequest $request): ProviderResponse
         throw new \Exception(sprintf('%s does not support structured mode', class_basename($this)));
     }
 
+    #[\Override]
+    public function embeddings(EmbeddingRequest $request): EmbeddingResponse
+    {
+        throw new \Exception(sprintf('%s does not support embeddings', class_basename($this)));
+    }
+
     /**
      * @param  array<string, mixed>  $options
      * @param  array<mixed>  $retry