ai-lite 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2fbb65bf2c01cdc7950a76cc5a273c0b495174bb9d671f7b5d32d49a79274b34
-  data.tar.gz: e8e909ff9c107f552e1f6a81772b3800bb2b1904bf9a5da50cfe6a2ca2380bd4
+  metadata.gz: cd1b2ead0a8ae48d9617b9e73460649f817fc158430e11cc7b9ac798a86846d2
+  data.tar.gz: f3dad03f56f3907e537dc8c72d63bf7e57e069d4b28ef666151d348b98f6b3e6
 SHA512:
-  metadata.gz: ef2a0f01ea606b121b583134f6026c7d23388b3a58732a5a1f197275cb7178d2547ffea5040956955fd39452f944f1703833a4ba4925d0d2314ef2693372b904
-  data.tar.gz: 272108490b47700c06f49fb817d684fa6a2dbd003e02eff35fa59f19db8fe120e41983d941c32a50c640ef6776aee00038bc5ed841bef60cdde04fee854ea0c4
+  metadata.gz: 6e70409efd1d697b2578ace5b878e74ef4e6214efbcbae2c295aae066e0e9ce4887cd39c95fcdde8bc73e0f6b0cfe60434dc7f1ca2611c77e649cc645b94351a
+  data.tar.gz: 1faba64c32c14038f18c1d50a2f73bd92a0c940d8b303b3eecb89c2193d944d3b189334c5ecde81f55e8268db01e01e562686b74f20111105f0c0fbbeea387c6

data/README.md

@@ -18,6 +18,7 @@ It is not meant to replace the official OpenAI SDK. It is a small wrapper for pr
 ```ruby
 ai.chat("Say hello")
 ai.moderate("User submitted text")
+ai.embed("Text to vectorize")
 ```
 
 ## Usage
@@ -49,6 +50,7 @@ AiLite.configure do |config|
   config.api_key = ENV["OPENAI_API_KEY"]
   config.model = "gpt-5.5"
   config.moderation_model = "omni-moderation-latest"
+  config.embedding_model = "text-embedding-3-small"
   config.timeout = 120
   config.max_output_tokens = 2000
 end
@@ -95,16 +97,43 @@ The OpenAI API URL is fixed to `https://api.openai.com/v1/responses`.
 
 ## Moderation
 
-Use `moderate` to classify text or images for potentially harmful content:
+Use `moderate` to classify user-submitted text or images for potentially harmful content before saving, publishing, or sending it into another AI call.
+
+A common use case is checking a comment before it is shown publicly:
 
 ```ruby
-result = ai.moderate("User submitted text")
+comment = "User submitted comment"
+result = ai.moderate(comment)
 
 if result["content"]["flagged"]
-  # hide, block, or route to review
+  # hide, block, or route the comment to review
+else
+  # publish the comment
 end
 ```
 
+Moderation responses include an overall `flagged` value plus category booleans and scores:
+
+```ruby
+{
+  "content" => {
+    "flagged" => false,
+    "categories" => {
+      "hate" => false,
+      "violence" => false
+    },
+    "category_scores" => {
+      "hate" => 0.01,
+      "violence" => 0.02
+    }
+  },
+  "response_id" => "modr_...",
+  "status" => 200,
+  "error" => nil,
+  "raw" => nil
+}
+```
+
 `moderate` sends a `POST` request to `/v1/moderations` with:
 
 - `model`
@@ -144,6 +173,72 @@ result = ai.moderate([
 ])
 ```
 
+## Embeddings
+
+Use `embed` to turn text into vectors that can be stored and compared for semantic search, recommendations, duplicate detection, or retrieval-augmented generation.
+
+```ruby
+result = ai.embed("How do I reset my password?")
+vector = result["content"]
+```
+
+For one string input, `content` is the embedding vector:
+
+```ruby
+{
+  "content" => [0.012, -0.44, 0.203],
+  "response_id" => nil,
+  "status" => 200,
+  "error" => nil,
+  "raw" => nil
+}
+```
+
+For multiple inputs, `content` is an array of vectors in the same order:
+
+```ruby
+result = ai.embed([
+  "How do I reset my password?",
+  "How do I update my billing card?"
+])
+
+vectors = result["content"]
+first_vector = vectors[0]
+```
+
+```ruby
+{
+  "content" => [
+    [0.012, -0.44, 0.203],
+    [0.332, 0.021, -0.118]
+  ],
+  "response_id" => nil,
+  "status" => 200,
+  "error" => nil,
+  "raw" => nil
+}
+```
+
+`embed` sends a `POST` request to `/v1/embeddings` with:
+
+- `model`
+- `input`
+- optional `dimensions`
+- optional `encoding_format`
+- optional `debug`
+- optional extra `options`
+
+The default embedding model is `text-embedding-3-small`.
+
+Pass `debug: true` to include the raw OpenAI response, including token usage:
+
+```ruby
+result = ai.embed("Hello", debug: true)
+
+result["content"]       # embedding vector
+result["raw"]["usage"]  # token usage
+```
+
 ## Multi-Turn Chat
 
 Responses include a `response_id` that can be passed back through `options` as `previous_response_id`:
@@ -189,22 +284,6 @@ JSON-looking model output:
 }
 ```
 
-Moderation output:
-
-```ruby
-{
-  "content" => {
-    "flagged" => false,
-    "categories" => { ... },
-    "category_scores" => { ... }
-  },
-  "response_id" => "modr_...",
-  "status" => 200,
-  "error" => nil,
-  "raw" => nil
-}
-```
-
 Failure:
 
 ```ruby

data/lib/ai_lite/version.rb

@@ -1,3 +1,3 @@
 class AiLite
-  VERSION = "0.2.0".freeze
+  VERSION = "0.3.0".freeze
 end

data/lib/ai_lite.rb

@@ -8,6 +8,7 @@ class AiLite
   API_BASE_URL = "https://api.openai.com/v1".freeze
   DEFAULT_MODEL = "gpt-5.5".freeze
   DEFAULT_MODERATION_MODEL = "omni-moderation-latest".freeze
+  DEFAULT_EMBEDDING_MODEL = "text-embedding-3-small".freeze
   DEFAULT_TIMEOUT = 120
   DEFAULT_MAX_OUTPUT_TOKENS = 2000
   IMAGE_MIME_TYPES = {
@@ -19,12 +20,13 @@ class AiLite
   }.freeze
 
   class Configuration
-    attr_accessor :api_key, :model, :moderation_model, :timeout, :max_output_tokens
+    attr_accessor :api_key, :model, :moderation_model, :embedding_model, :timeout, :max_output_tokens
 
     def initialize
       @api_key = nil
       @model = DEFAULT_MODEL
       @moderation_model = DEFAULT_MODERATION_MODEL
+      @embedding_model = DEFAULT_EMBEDDING_MODEL
       @timeout = DEFAULT_TIMEOUT
       @max_output_tokens = DEFAULT_MAX_OUTPUT_TOKENS
     end
@@ -59,19 +61,24 @@ class AiLite
       client.moderate(input, **kwargs)
     end
 
+    def embed(input, **kwargs)
+      client.embed(input, **kwargs)
+    end
+
     def reset_client!
       @client = nil
     end
   end
 
-  attr_reader :api_key, :model, :moderation_model, :timeout, :max_output_tokens, :headers
+  attr_reader :api_key, :model, :moderation_model, :embedding_model, :timeout, :max_output_tokens, :headers
 
-  def initialize(api_key: nil, model: nil, moderation_model: nil, timeout: nil, max_output_tokens: nil)
+  def initialize(api_key: nil, model: nil, moderation_model: nil, embedding_model: nil, timeout: nil, max_output_tokens: nil)
     @api_key = api_key || self.class.configuration.api_key || ENV["OPENAI_API_KEY"] || ENV["OPEN_AI_TOKEN"]
     raise ArgumentError, "Missing OpenAI API key" if @api_key.to_s.strip.empty?
 
     @model = model || self.class.configuration.model
     @moderation_model = moderation_model || self.class.configuration.moderation_model
+    @embedding_model = embedding_model || self.class.configuration.embedding_model
     @timeout = timeout || self.class.configuration.timeout
     @max_output_tokens = max_output_tokens || self.class.configuration.max_output_tokens
     @headers = {
@@ -104,6 +111,19 @@ class AiLite
     prettify_data(status: "unknown", error: e.message, raw: nil, debug: debug)
   end
 
+  def embed(input, model: nil, dimensions: nil, encoding_format: nil, debug: false, options: {})
+    payload = options.merge(
+      model: model || embedding_model,
+      input: input
+    )
+    payload[:dimensions] = dimensions if dimensions
+    payload[:encoding_format] = encoding_format if encoding_format
+
+    extract_embedding(post(payload, endpoint: embedding_endpoint), multiple: input.is_a?(Array), debug: debug)
+  rescue => e
+    prettify_data(status: "unknown", error: e.message, raw: nil, debug: debug)
+  end
+
   private
 
   def post(payload, endpoint: response_endpoint)
@@ -131,6 +151,10 @@ class AiLite
     "#{API_BASE_URL}/moderations"
   end
 
+  def embedding_endpoint
+    "#{API_BASE_URL}/embeddings"
+  end
+
   def extract_content(response, debug: false)
     status = response.code.to_i
     parsed_response = JSON.parse(response.body)
@@ -187,6 +211,33 @@ class AiLite
     prettify_data(status: response_status(response), error: e.message, raw: nil, debug: debug)
   end
 
+  def extract_embedding(response, multiple:, debug: false)
+    status = response.code.to_i
+    parsed_response = JSON.parse(response.body)
+
+    unless success_status?(status)
+      return prettify_data(
+        status: status,
+        error: error_message(parsed_response),
+        response_id: parsed_response["id"],
+        raw: parsed_response,
+        debug: debug
+      )
+    end
+
+    prettify_data(
+      status: status,
+      content: embedding_content(parsed_response, multiple: multiple),
+      response_id: parsed_response["id"],
+      raw: parsed_response,
+      debug: debug
+    )
+  rescue JSON::ParserError => e
+    prettify_data(status: response_status(response), error: e.message, raw: response&.body, debug: debug)
+  rescue => e
+    prettify_data(status: response_status(response), error: e.message, raw: nil, debug: debug)
+  end
+
   def extract_output_text(raw)
     Array(raw["output"]).flat_map do |item|
       next [] unless item.is_a?(Hash) && item["type"] == "message"
@@ -253,6 +304,14 @@ class AiLite
     results.length == 1 ? results.first : results
   end
 
+  def embedding_content(raw, multiple:)
+    embeddings = Array(raw["data"]).map do |item|
+      item["embedding"] if item.is_a?(Hash)
+    end.compact
+
+    multiple ? embeddings : embeddings.first
+  end
+
   def success_status?(status)
     status >= 200 && status < 300
   end

data/test/ai_lite_test.rb

@@ -39,6 +39,7 @@ class AiLiteTest < Minitest::Test
       assert_equal "explicit-key", client.api_key
       assert_equal "gpt-test", client.model
       assert_equal "omni-moderation-latest", client.moderation_model
+      assert_equal "text-embedding-3-small", client.embedding_model
       assert_equal 10, client.timeout
       assert_equal 2000, client.max_output_tokens
       assert_equal "Bearer explicit-key", client.headers["Authorization"]
@@ -52,6 +53,7 @@ class AiLiteTest < Minitest::Test
         config.api_key = "configured-key"
         config.model = "gpt-config"
         config.moderation_model = "omni-moderation-test"
+        config.embedding_model = "text-embedding-test"
         config.timeout = 15
         config.max_output_tokens = 750
       end
@@ -61,6 +63,7 @@ class AiLiteTest < Minitest::Test
       assert_equal "configured-key", client.api_key
       assert_equal "gpt-config", client.model
       assert_equal "omni-moderation-test", client.moderation_model
+      assert_equal "text-embedding-test", client.embedding_model
       assert_equal 15, client.timeout
       assert_equal 750, client.max_output_tokens
       assert_same client, AiLite.client
@@ -83,6 +86,7 @@ class AiLiteTest < Minitest::Test
       config.api_key = "configured-key"
       config.model = "gpt-config"
       config.moderation_model = "omni-moderation-config"
+      config.embedding_model = "text-embedding-config"
       config.timeout = 15
       config.max_output_tokens = 750
     end
@@ -91,6 +95,7 @@ class AiLiteTest < Minitest::Test
       api_key: "explicit-key",
       model: "gpt-explicit",
       moderation_model: "omni-moderation-explicit",
+      embedding_model: "text-embedding-explicit",
       timeout: 5,
       max_output_tokens: 300
     )
@@ -98,6 +103,7 @@ class AiLiteTest < Minitest::Test
     assert_equal "explicit-key", client.api_key
     assert_equal "gpt-explicit", client.model
     assert_equal "omni-moderation-explicit", client.moderation_model
+    assert_equal "text-embedding-explicit", client.embedding_model
     assert_equal 5, client.timeout
     assert_equal 300, client.max_output_tokens
   end
@@ -427,6 +433,98 @@ class AiLiteTest < Minitest::Test
     assert_nil result["raw"]
   end
 
+  def test_embed_sends_post_to_embeddings_with_default_payload
+    client = AiLite.new(api_key: "token-abc")
+
+    with_stubbed_http(embedding_response([[0.12, -0.34, 0.56]])) do |captured, _response|
+      result = client.embed("Text to vectorize")
+      request = captured[:http].last_request
+      payload = JSON.parse(request.body)
+
+      assert_equal [0.12, -0.34, 0.56], result["content"]
+      assert_nil result["response_id"]
+      assert_equal 200, result["status"]
+      assert_nil result["error"]
+      assert_nil result["raw"]
+      assert_equal "api.openai.com", captured[:host]
+      assert_equal 443, captured[:port]
+      assert_equal true, captured[:use_ssl]
+      assert_instance_of Net::HTTP::Post, request
+      assert_equal "/v1/embeddings", request.path
+      assert_equal "Bearer token-abc", request["Authorization"]
+      assert_equal "application/json", request["Content-Type"]
+      assert_equal "text-embedding-3-small", payload["model"]
+      assert_equal "Text to vectorize", payload["input"]
+    end
+  end
+
+  def test_embed_returns_vector_list_for_multiple_inputs
+    client = AiLite.new(api_key: "token-abc")
+    embeddings = [
+      [0.11, 0.22, 0.33],
+      [0.44, 0.55, 0.66]
+    ]
+
+    with_stubbed_http(embedding_response(embeddings)) do |captured, _response|
+      result = client.embed(["First text", "Second text"])
+      payload = JSON.parse(captured[:http].last_request.body)
+
+      assert_equal ["First text", "Second text"], payload["input"]
+      assert_equal embeddings, result["content"]
+      assert_equal 200, result["status"]
+      assert_nil result["error"]
+    end
+  end
+
+  def test_embed_includes_options_dimensions_encoding_format_and_model
+    client = AiLite.new(api_key: "token-abc")
+
+    with_stubbed_http(embedding_response([[0.12, -0.34]])) do |captured, _response|
+      client.embed(
+        "Short text",
+        model: "text-embedding-3-large",
+        dimensions: 2,
+        encoding_format: "float",
+        options: {
+          user: "user-123"
+        }
+      )
+      payload = JSON.parse(captured[:http].last_request.body)
+
+      assert_equal "text-embedding-3-large", payload["model"]
+      assert_equal "Short text", payload["input"]
+      assert_equal 2, payload["dimensions"]
+      assert_equal "float", payload["encoding_format"]
+      assert_equal "user-123", payload["user"]
+    end
+  end
+
+  def test_embed_uses_class_level_configured_client
+    AiLite.configure do |config|
+      config.api_key = "configured-key"
+      config.embedding_model = "text-embedding-config"
+    end
+
+    with_stubbed_http(embedding_response([[0.12, -0.34, 0.56]])) do |captured, _response|
+      AiLite.embed("Use configured defaults")
+      payload = JSON.parse(captured[:http].last_request.body)
+
+      assert_equal "text-embedding-config", payload["model"]
+      assert_equal "Bearer configured-key", captured[:http].last_request["Authorization"]
+    end
+  end
+
+  def test_embed_debug_true_returns_raw_usage
+    client = AiLite.new(api_key: "token-abc")
+
+    with_stubbed_http(embedding_response([[0.12, -0.34, 0.56]])) do |_captured, _response|
+      result = client.embed("Text to vectorize", debug: true)
+
+      assert_equal [0.12, -0.34, 0.56], result["content"]
+      assert_equal({ "prompt_tokens" => 4, "total_tokens" => 4 }, result["raw"]["usage"])
+    end
+  end
+
   def test_http_errors_return_standard_envelope
     client = AiLite.new(api_key: "token-abc")
     body = JSON.generate("error" => { "message" => "Invalid API key" })
@@ -542,6 +640,25 @@ class AiLiteTest < Minitest::Test
     }
   end
 
+  def embedding_response(embeddings)
+    body = JSON.generate(
+      "object" => "list",
+      "data" => embeddings.each_with_index.map do |embedding, index|
+        {
+          "object" => "embedding",
+          "index" => index,
+          "embedding" => embedding
+        }
+      end,
+      "model" => "text-embedding-3-small",
+      "usage" => {
+        "prompt_tokens" => 4,
+        "total_tokens" => 4
+      }
+    )
+    FakeResponse.new("200", body)
+  end
+
   def with_env(values)
     originals = {}
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ai-lite
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - William Basmayor
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2026-06-19 00:00:00.000000000 Z
+date: 2026-07-07 00:00:00.000000000 Z
 dependencies: []
 description: AI Lite is a dependency-light Ruby client for Rails apps and plain Ruby
   projects that need simple OpenAI Responses API calls.