rasti-ai 3.2.0 → 3.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: fa07486ef7216a0a1408d1711ea21ca521283673c06ca2845ad63197dd1dbe8d
-  data.tar.gz: dd42552dfd7e3752dac0264b1c89159f04f230c6d067fa674dd4347aae01b0b1
+  metadata.gz: 58f15bbf9597b75e9dc32542801d07da3661a2e0477e13460ea9116e610501c5
+  data.tar.gz: '08f39b213898e52ee7fb378472e5e2ba0bc68a7823b000e086afeb9c95b14457'
 SHA512:
-  metadata.gz: 1cd0846a2f725823ba691f7d761c5abee299291d9307a7a870cc086f17a02ead88913acaf70cf257221b82a5b8684655595cdc4a8952d4825fa19f23ab7004e5
-  data.tar.gz: ea5880cc6e73795cac795189e6a6d192e27fdd11a252768f1b33933b13407d9e1f929e1aaf4020112164dfa880acbe4601708f202795737b434df0f5a63202d6
+  metadata.gz: 6237690a28cea19cfcd156f67dbf45674619dacfb0f8dd5a540da30aadf2020ec2a838a3a04439c959446faef65718ee8f83fff58524ed6674fbe4547cc55d81
+  data.tar.gz: e5e6f685ab358ddc17b006de0885c343ee370b8769b7cae8fb2a6031878643c19565e8b8994d7baaece9a050ada387ef59e4e3b1df8b52881d15d73fafb4e7cf

data/AGENTS.md

@@ -272,10 +272,11 @@ Thinking is inherited too: `reasoning_effort` is passed through as-is (same as O
 
 ## Thinking levels
 
-The base `Assistant` accepts `thinking: 'low' | 'medium' | 'high'` (validated on construction; nil = disabled). Each provider translates it in a private `thinking_config` method and passes the result to the client. The client includes it in the request body only if present.
+The base `Assistant` accepts `thinking: 'none' | 'low' | 'medium' | 'high'` (validated on construction). `nil` means the gem sends nothing and the model uses its own default — which for reasoning models (e.g. `gpt-6-luna`, `gemini-2.5-flash`, Huawei's DeepSeek/GLM) means reasoning *on*. `'none'` explicitly asks the provider to disable reasoning. Each provider translates it in a private `thinking_config` method and passes the result to the client. The client includes it in the request body only if present.
 
 | Level | OpenAI `reasoning_effort` | Anthropic `budget_tokens` | Gemini `thinking_budget` |
 |---|---|---|---|
+| `'none'` | `'none'` | _(field omitted)_ | `0` |
 | `'low'` | `'low'` | `1_024` | `1_024` |
 | `'medium'` | `'medium'` | `8_000` | `8_192` |
 | `'high'` | `'high'` | `16_000` | `24_576` |
@@ -284,7 +285,11 @@ The `THINKING_LEVELS` table lives in each **`Provider`** class (it's provider me
 
 For Gemini, `thinking_config` goes inside `generation_config` — the client doesn't need a new param. For OpenAI and Anthropic, it's a separate top-level param in the client method (`reasoning_effort:` and `thinking:` respectively).
 
-OpenRouter and Huawei MaaS inherit OpenAI's behavior: `thinking` is passed through as `reasoning_effort` with the same string values (`'low'`/`'medium'`/`'high'`). No `thinking_config` override is needed. Note that Huawei's models typically accept only `'high'` and `'max'` — `'high'` is the safe choice, and `'max'` is not expressible through the gem's universal levels.
+OpenRouter and Huawei MaaS inherit OpenAI's behavior: `thinking` is passed through as `reasoning_effort` with the same string values (`'low'`/`'medium'`/`'high'`). No `thinking_config` override is needed. Huawei MaaS rejects `reasoning_effort: 'none'` with a 400 (`deepseek-v4-flash` accepts `low|medium|high|xhigh|max`, `glm-5.3` only `low|high|max`), so `thinking: 'none'` fails there by design. Disabling reasoning on Huawei needs `thinking: {type: 'disabled'}` instead (works on DeepSeek and GLM ≤ 5.2, rejected by `glm-5.3`); the gem does not send it.
+
+`'none'` matters for `gpt-6-luna`: it reasons by default and rejects function tools on `/v1/chat/completions` unless `reasoning_effort: 'none'` is sent. Combining reasoning with tools there requires `/v1/responses`, which the gem does not support.
+
+Gemini's `thinking_budget: 0` is rejected by models that cannot turn reasoning off (e.g. `gemini-2.5-pro`).
 
 The loop does not change. Anthropic thinking blocks (`type: 'thinking'`) in responses are ignored by `parse_content` (looks for `type == 'text'`) and preserved automatically by `build_assistant_tool_calls_message` (passes full `response['content']` array).
 
@@ -315,10 +320,12 @@ Add an entry to the `PROVIDERS` table in `tasks/assistant.rake` so the interacti
 ```ruby
 PROVIDERS = {
   # existing providers ...
-  '<provider>' => {key: '<PROVIDER>_API_KEY', klass: -> { Rasti::AI::<Provider>::Assistant }}
+  '<provider>' => {key: '<PROVIDER>_API_KEY', klass: -> { Rasti::AI::<Provider>::Assistant }, thinking: 'none'}
 }.freeze
 ```
 
+`thinking:` is optional and passed as-is to the assistant. All providers use `'none'` except `huawei_maas`, which rejects it (see [Thinking levels](#thinking-levels)).
+
 The task name, description, env-key check, logger path and banner are all derived automatically from this entry.
 
 Add to `spec/minitest_helper.rb`:
@@ -392,7 +399,7 @@ All string constants use `.freeze`. Integer and array/hash literals that are alr
 USER = 'user'.freeze
 ASSISTANT = 'assistant'.freeze
 
-VALID_THINKING_LEVELS = %w[low medium high].freeze
+VALID_THINKING_LEVELS = %w[none low medium high].freeze
 
 THINKING_LEVELS = {
   'low' => {thinking_budget: 1_024}.freeze,

data/README.md

@@ -146,7 +146,7 @@ Both methods accept the same options:
 | `http_connect_timeout:`, `http_read_timeout:`, `http_max_retries:` | Per-call HTTP settings | ✓ | ✓ |
 | `client:` | A pre-built client (skips building one from the options above) | ✓ | ✓ |
 | `json_schema:` | Schema for structured responses | ✓ | ✓ |
-| `thinking:` | `'low'`, `'medium'` or `'high'` | ✓ | ✓ |
+| `thinking:` | `'none'`, `'low'`, `'medium'` or `'high'` | ✓ | ✓ |
 | `prompt:` | A single user message | ✓ | |
 | `messages:` | Conversation with generic roles — mutually exclusive with `prompt:` | ✓ | |
 | `system:` | System prompt. On `create_assistant` it's a shortcut for `state: AssistantState.new(context: ...)` | ✓ | ✓ |
@@ -330,6 +330,14 @@ assistant.call 'Solve this step by step: ...'
 
 The level controls how much computation the model can spend reasoning before responding. Higher levels may improve answer quality at the cost of more tokens and latency. Not all models support thinking — check your provider's documentation.
 
+Omitting `thinking:` leaves the decision to the model, and many current models reason by default. Pass `thinking: 'none'` to disable reasoning explicitly. Some models require it — `gpt-6-luna`, for example, rejects function tools unless reasoning is disabled:
+
+```ruby
+assistant = Rasti::AI.create_assistant provider: :open_ai, model: 'gpt-6-luna', tools: [GetCurrentWeather.new], thinking: 'none'
+```
+
+Huawei MaaS does not accept `'none'` and returns an error.
+
 ### Usage tracking
 
 Track token consumption across API calls (including tool calls):

data/lib/rasti/ai/anthropic/provider.rb

@@ -6,6 +6,7 @@ module Rasti
         STRUCTURED_OUTPUT_TOOL = 'structured_output'.freeze
 
         THINKING_LEVELS = {
+          'none' => nil,
           'low' => {type: 'enabled', budget_tokens: 1_024}.freeze,
           'medium' => {type: 'enabled', budget_tokens: 8_000}.freeze,
           'high' => {type: 'enabled', budget_tokens: 16_000}.freeze

data/lib/rasti/ai/assistant.rb

@@ -6,7 +6,7 @@ module Rasti
 
       attr_reader :state, :model, :thinking
 
-      VALID_THINKING_LEVELS = %w[low medium high].freeze
+      VALID_THINKING_LEVELS = %w[none low medium high].freeze
 
       def initialize(client:nil, provider:nil, json_schema:nil, state:nil, model:nil, thinking:nil, tools:[], mcp_servers:{}, logger:nil)
         raise ArgumentError, "Invalid thinking level '#{thinking}'. Valid: #{VALID_THINKING_LEVELS.join(', ')}" if thinking && !VALID_THINKING_LEVELS.include?(thinking)

data/lib/rasti/ai/gemini/provider.rb

@@ -4,6 +4,7 @@ module Rasti
       class Provider < Rasti::AI::Provider
 
         THINKING_LEVELS = {
+          'none' => {thinking_budget: 0}.freeze,
           'low' => {thinking_budget: 1_024}.freeze,
           'medium' => {thinking_budget: 8_192}.freeze,
           'high' => {thinking_budget: 24_576}.freeze

data/lib/rasti/ai/open_ai/assistant.rb

@@ -37,7 +37,7 @@ module Rasti
         end
 
         def thinking_config
-          thinking
+          Provider::THINKING_LEVELS[thinking]
         end
 
         def parse_tool_calls(response)

data/lib/rasti/ai/open_ai/provider.rb

@@ -4,6 +4,7 @@ module Rasti
       class Provider < Rasti::AI::Provider
 
         THINKING_LEVELS = {
+          'none' => 'none'.freeze,
           'low' => 'low'.freeze,
           'medium' => 'medium'.freeze,
           'high' => 'high'.freeze

data/lib/rasti/ai/version.rb

@@ -1,5 +1,5 @@
 module Rasti
   module AI
-    VERSION = '3.2.0'
+    VERSION = '3.2.1'
   end
 end

data/spec/anthropic/assistant_spec.rb

@@ -135,6 +135,20 @@ describe Rasti::AI::Anthropic::Assistant do
       assert_equal answer, response
     end
 
+    it 'Thinking none' do
+      body = read_json_resource('anthropic/basic_request.json', model: Rasti::AI.anthropic_default_model, prompt: question)
+
+      stub_request(:post, api_url)
+        .with(body: JSON.dump(body))
+        .to_return(body: read_resource('anthropic/basic_response.json', content: answer))
+
+      assistant = Rasti::AI::Anthropic::Assistant.new thinking: 'none'
+
+      response = assistant.call question
+
+      assert_equal answer, response
+    end
+
     it 'JSON Schema' do
       json_schema = {'answer' => 'Response answer'}
 

data/spec/anthropic/provider_spec.rb

@@ -58,6 +58,21 @@ describe Rasti::AI::Anthropic::Provider do
     assert_equal answer, result.content
   end
 
+  it 'Generate text with thinking none' do
+    stub_messages model: Rasti::AI.anthropic_default_model,
+                  max_tokens: 4096,
+                  messages: [
+                    {
+                      role: 'user',
+                      content: question
+                    }
+                  ]
+
+    result = provider.generate_text prompt: question, thinking: 'none'
+
+    assert_equal answer, result.content
+  end
+
   it 'Generate text with json schema' do
     json_schema = {player: 'string'}
 

data/spec/gemini/assistant_spec.rb

@@ -125,6 +125,21 @@ describe Rasti::AI::Gemini::Assistant do
       assert_equal answer, response
     end
 
+    it 'Thinking none' do
+      body = read_json_resource('gemini/basic_request.json', prompt: question)
+      body['generation_config'] = {'thinking_config' => {'thinking_budget' => 0}}
+
+      stub_request(:post, api_url)
+        .with(body: JSON.dump(body))
+        .to_return(body: read_resource('gemini/basic_response.json', content: answer))
+
+      assistant = Rasti::AI::Gemini::Assistant.new thinking: 'none'
+
+      response = assistant.call question
+
+      assert_equal answer, response
+    end
+
     it 'JSON Schema' do
       json_schema = {answer: 'Response answer'}
       json_answer = "{\\\"answer\\\": \\\"#{answer}\\\"}"

data/spec/gemini/provider_spec.rb

@@ -87,4 +87,18 @@ describe Rasti::AI::Gemini::Provider do
     assert_equal answer, provider.generate_text(prompt: question, json_schema: json_schema, thinking: 'medium').content
   end
 
+  it 'Generate text with thinking none' do
+    stub_generate_content contents: [
+                            {
+                              role: 'user',
+                              parts: [{text: question}]
+                            }
+                          ],
+                          generation_config: {
+                            thinking_config: {thinking_budget: 0}
+                          }
+
+    assert_equal answer, provider.generate_text(prompt: question, thinking: 'none').content
+  end
+
 end

data/spec/open_ai/assistant_spec.rb

@@ -127,6 +127,21 @@ describe Rasti::AI::OpenAI::Assistant do
       assert_equal answer, response
     end
 
+    it 'Thinking none' do
+      body = read_json_resource('open_ai/basic_request.json', model: Rasti::AI.openai_default_model, prompt: question)
+      body['reasoning_effort'] = 'none'
+
+      stub_request(:post, api_url)
+        .with(body: JSON.dump(body))
+        .to_return(body: read_resource('open_ai/basic_response.json', content: answer))
+
+      assistant = Rasti::AI::OpenAI::Assistant.new thinking: 'none'
+
+      response = assistant.call question
+
+      assert_equal answer, response
+    end
+
     it 'JSON Schema' do
       json_schema = {answer: 'Response answer'}
       json_answer = "{\\\"answer\\\": \\\"#{answer}\\\"}"

data/spec/open_ai/provider_spec.rb

@@ -62,6 +62,16 @@ describe Rasti::AI::OpenAI::Provider do
     assert_equal answer, result.content
   end
 
+  it 'Generate text with thinking none' do
+    body = basic_body(Rasti::AI.openai_default_model).merge(reasoning_effort: 'none')
+
+    stub_chat_completions 'https://api.openai.com/v1/chat/completions', body
+
+    result = Rasti::AI.provider(:open_ai).generate_text prompt: question, thinking: 'none'
+
+    assert_equal answer, result.content
+  end
+
   describe 'OpenAI compatible providers' do
 
     it 'OpenRouter' do

data/spec/rasti_ai_spec.rb

@@ -203,7 +203,7 @@ describe Rasti::AI do
         Rasti::AI.provider(:open_ai).generate_text prompt: question, thinking: 'invalid'
       end
 
-      assert_equal "Invalid thinking level 'invalid'. Valid: low, medium, high", error.message
+      assert_equal "Invalid thinking level 'invalid'. Valid: none, low, medium, high", error.message
     end
 
   end

data/tasks/assistant.rake

@@ -19,11 +19,11 @@ require 'logger'
 WEATHER_MCP_URL = 'https://gateway.pipeworx.io/weather/mcp'.freeze
 
 PROVIDERS = {
-  'openai'     => {key: 'OPENAI_API_KEY',       klass: -> { Rasti::AI::OpenAI::Assistant }},
-  'gemini'     => {key: 'GEMINI_API_KEY',       klass: -> { Rasti::AI::Gemini::Assistant }},
-  'anthropic'  => {key: 'ANTHROPIC_API_KEY',    klass: -> { Rasti::AI::Anthropic::Assistant }},
-  'openrouter' => {key: 'OPENROUTER_API_KEY',   klass: -> { Rasti::AI::OpenRouter::Assistant }},
-  'huawei_maas'=> {key: 'HUAWEI_MAAS_API_KEY',  klass: -> { Rasti::AI::HuaweiMaaS::Assistant }}
+  'openai' => {key: 'OPENAI_API_KEY', klass: -> { Rasti::AI::OpenAI::Assistant }, thinking: 'none'},
+  'gemini' => {key: 'GEMINI_API_KEY', klass: -> { Rasti::AI::Gemini::Assistant }, thinking: 'none'},
+  'anthropic' => {key: 'ANTHROPIC_API_KEY', klass: -> { Rasti::AI::Anthropic::Assistant }, thinking: 'none'},
+  'openrouter' => {key: 'OPENROUTER_API_KEY', klass: -> { Rasti::AI::OpenRouter::Assistant }, thinking: 'none'},
+  'huawei_maas' => {key: 'HUAWEI_MAAS_API_KEY', klass: -> { Rasti::AI::HuaweiMaaS::Assistant }}
 }.freeze
 
 def build_weather_mcp
@@ -70,7 +70,7 @@ def chat_loop(assistant)
   puts "\nGoodbye!"
 end
 
-def start_assistant(provider, env_key, assistant_klass)
+def start_assistant(provider, env_key, assistant_klass, thinking)
   require 'rasti-ai'
 
   abort "[Error] #{env_key} is not set." unless ENV[env_key]
@@ -81,7 +81,7 @@ def start_assistant(provider, env_key, assistant_klass)
   mcp = build_weather_mcp
   print_banner provider.capitalize, mcp.list_tools.map { |t| t['name'] }
 
-  chat_loop assistant_klass.call.new(mcp_servers: {weather: mcp})
+  chat_loop assistant_klass.call.new(mcp_servers: {weather: mcp}, thinking: thinking)
 end
 
 namespace :assistant do
@@ -89,7 +89,7 @@ namespace :assistant do
   PROVIDERS.each do |provider, config|
     desc "Chat with #{provider.capitalize} assistant using Pipeworx weather MCP (requires #{config[:key]})"
     task provider.to_sym do
-      start_assistant(provider, config[:key], config[:klass])
+      start_assistant(provider, config[:key], config[:klass], config[:thinking])
     end
   end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rasti-ai
 version: !ruby/object:Gem::Version
-  version: 3.2.0
+  version: 3.2.1
 platform: ruby
 authors:
 - Gabriel Naiman
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-07-28 00:00:00.000000000 Z
+date: 2026-10-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: multi_require