lex-llm-vertex 0.2.0 → 0.2.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 35a2ffd46ca20c21c1c0d280688794b6702285b996d0f1386c37de565a409b57
-  data.tar.gz: 5ae360df40fe27bd2c43d37cd15ee36449303422df3a65550e596fc51f22ae3e
+  metadata.gz: 278b657e5c2050cb2208929e9ae1e872666cbccc5ccfbf6cda2253550d9aca89
+  data.tar.gz: 25ce6f8fdf008778892ff3ec8ca4940b07e3f53e7c5c847baa2ed622c04e8055
 SHA512:
-  metadata.gz: ac9e04f6a5ffdf57bd83392cbb039220871eac5d006256bf90c0a095bae241722a2aa8a05216ab663751bc0d5f144eec95ffcf6f6b5921f2135309938bd3de77
-  data.tar.gz: 928812748673dd57855e0ad4e718d2de1fa0f331d2d60df190d96b072de49cd4b317f4eb905c59a9eccaaf38a85c5ea2b87879bcaab00d5a7c4446db6295258d
+  metadata.gz: ca1e2604260d332f28565e7d492a9e38ee1b983f8306d78bdfc36c071f7bd00665465c7462d216c44f319164771b3272e5e93a1a4737ccddd58d5feec39df659
+  data.tar.gz: bf2fb3a0389015cda88dd1a934c178e1431fddff9bdc482314737e9823e2fa16d378c103571c2e9fd67214947fc898b93e2a4bd567c55eac370ce6f90eef01f1

data/.github/workflows/ci.yml

@@ -8,8 +8,20 @@ jobs:
   ci:
     uses: LegionIO/.github/.github/workflows/ci.yml@main
 
+  excluded-files:
+    uses: LegionIO/.github/.github/workflows/excluded-files.yml@main
+
+  security:
+    uses: LegionIO/.github/.github/workflows/security-scan.yml@main
+
+  version-changelog:
+    uses: LegionIO/.github/.github/workflows/version-changelog.yml@main
+
+  dependency-review:
+    uses: LegionIO/.github/.github/workflows/dependency-review.yml@main
+
   release:
-    needs: ci
+    needs: [ci, excluded-files, security]
     if: github.event_name == 'push' && github.ref == 'refs/heads/main'
     uses: LegionIO/.github/.github/workflows/release.yml@main
     secrets:

data/CHANGELOG.md

@@ -1,5 +1,42 @@
 # Changelog
 
+## 0.2.6 - 2026-05-06
+
+- Load provider-owned fleet actors through the LegionIO subscription base and the canonical Vertex provider root.
+- Keep fleet runners anchored on the provider root namespace so provider constants and instance discovery are always loaded.
+- Preserve configured transport and tier metadata when Vertex builds routing offerings.
+- Remove throwaway unused-argument allocation in provider request methods.
+- Gate release publishing on the shared security workflow.
+
+## 0.2.5 - 2026-05-06
+
+- Use the shared `lex-llm` fleet provider responder helper for provider-owned fleet workers.
+- Remove the runtime `legion-llm` dependency and require `lex-llm >= 0.4.3` for responder-side fleet execution.
+
+## 0.2.4 - 2026-05-06
+
+- Keep clean CI installs on published RubyGems dependency floors while preserving local path overrides for unreleased sibling integration testing.
+- Add a `stream_chat` compatibility alias so Vertex exposes the shared provider streaming surface even when running against older published `lex-llm` versions.
+- Register Vertex configuration options directly when the installed `lex-llm` does not expose `Configuration.register_provider_options`.
+- Make the provider-owned fleet responder bridge load only when the installed `legion-llm` exposes `Legion::LLM::Fleet::ProviderResponder`; fleet actors stay disabled instead of breaking gem load when that helper is unavailable.
+- Refresh README dependency, fleet responder, file-map, license, and development-command guidance.
+
+## 0.2.3 - 2026-05-06
+
+- Remove require-time provider self-registration; `legion-llm` now owns adapter creation and registry writes from loaded provider discovery metadata.
+- Bump dependency floors to `lex-llm >= 0.4.1` and `legion-llm >= 0.9.1`.
+
+## 0.2.2 - 2026-05-06
+
+- Enforce the shared keyword-only `lex-llm` provider contract for chat, embeddings, and token counting.
+- Move Vertex defaults back to `Legion::Extensions::Llm.provider_settings` with credentials/provider metadata under the default instance and instance-level fleet responder settings.
+- Add provider-owned fleet responder actor and runner backed by `legion-llm` fleet policy execution.
+- Bump the transport dependency floor to `legion-transport >= 1.4.14`.
+
+## 0.2.1 - 2026-05-03
+
+- Normalize generic settings keys to Vertex provider config keys during instance discovery.
+
 ## 0.2.0 - 2026-05-01
 
 - Add auto-discovery via CredentialSources and AutoRegistration from lex-llm 0.3.0

data/Gemfile

@@ -4,6 +4,8 @@ source 'https://rubygems.org'
 
 group :test do
   llm_base_path = ENV.fetch('LEX_LLM_PATH', File.expand_path('../lex-llm', __dir__))
+  transport_path = ENV.fetch('LEGION_TRANSPORT_PATH', File.expand_path('../../legion-transport', __dir__))
+  gem 'legion-transport', path: transport_path if File.directory?(transport_path)
   gem 'lex-llm', path: llm_base_path if File.directory?(llm_base_path)
 end
 

data/README.md

@@ -2,7 +2,7 @@
 
 Google Cloud Vertex AI provider extension for `Legion::Extensions::Llm`.
 
-This gem adds a hosted Vertex AI provider surface for Legion LLM routing without depending on the old `legion-llm` gem. It keeps discovery offline by default, preserves full Vertex publisher model resource names for routing, and exposes project/location instance metadata for multi-region provider fleets. It requires `lex-llm >= 0.1.5` for the shared model offering, alias, readiness, and fleet lane contract.
+This gem adds a hosted Vertex AI provider surface for Legion LLM routing. It keeps discovery offline by default, preserves full Vertex publisher model resource names for routing, and exposes project/location instance metadata for multi-region provider fleets. It installs against the current published `lex-llm` gem, while the `Gemfile` can use local sibling checkouts for unreleased provider-contract testing.
 
 ## Install
 
@@ -32,6 +32,27 @@ Default settings expose `env://` references and keep live discovery disabled:
 Legion::Extensions::Llm::Vertex.default_settings
 ```
 
+## Fleet Responder
+
+Provider instances can opt in to consuming Legion LLM fleet requests. The provider-owned fleet actor only starts when at least one configured instance enables `respond_to_requests`.
+
+Fleet request execution is delegated to `Legion::Extensions::Llm::Fleet::ProviderResponder` from `lex-llm`. Request-side routing and reply orchestration remain owned by `legion-llm`; this provider only needs `lex-llm` and `legion-transport` to consume fleet jobs on a responder node.
+
+```yaml
+extensions:
+  llm:
+    vertex:
+      instances:
+        local:
+          fleet:
+            enabled: true
+            respond_to_requests: true
+            capabilities:
+              - chat
+              - stream_chat
+              - embed
+```
+
 ## Provider Surface
 
 ```ruby
@@ -40,10 +61,10 @@ provider = Legion::Extensions::Llm::Vertex::Provider.new(Legion::Extensions::Llm
 provider.discover_offerings(live: false)
 provider.offering_for(model: 'gemini-2.5-flash')
 provider.health(live: false)
-provider.chat(messages, model: model)
-provider.stream(messages, model: model) { |chunk| chunk.content }
-provider.embed('hello', model: 'gemini-embedding-001')
-provider.count_tokens(messages, model: model)
+provider.chat(messages:, model:)
+provider.stream_chat(messages:, model:) { |chunk| chunk.content }
+provider.embed(text: 'hello', model: 'gemini-embedding-001')
+provider.count_tokens(messages:, model:)
 ```
 
 `discover_offerings(live: false)` returns a conservative static catalog for routing defaults and unit tests. `discover_offerings(live: true)` calls the Vertex publisher models listing endpoint and maps returned model data into `Legion::Extensions::Llm::Routing::ModelOffering` records.
@@ -82,11 +103,9 @@ When transport is available, the `RegistryPublisher` publishes best-effort readi
 |------|---------|
 | `lib/legion/extensions/llm/vertex.rb` | Namespace module, default settings, provider registration |
 | `lib/legion/extensions/llm/vertex/provider.rb` | Vertex AI provider: chat, stream, embed, count_tokens, health, discovery |
-| `lib/legion/extensions/llm/vertex/registry_publisher.rb` | Async best-effort llm.registry event publisher |
-| `lib/legion/extensions/llm/vertex/registry_event_builder.rb` | Builds sanitized registry event envelopes |
+| `lib/legion/extensions/llm/vertex/actors/fleet_worker.rb` | Legion subscription actor for provider-owned fleet request consumption |
+| `lib/legion/extensions/llm/vertex/runners/fleet_worker.rb` | Runner entrypoint that delegates fleet request execution to `lex-llm` |
 | `lib/legion/extensions/llm/vertex/version.rb` | `VERSION` constant |
-| `lib/legion/extensions/llm/vertex/transport/exchanges/llm_registry.rb` | `llm.registry` topic exchange definition |
-| `lib/legion/extensions/llm/vertex/transport/messages/registry_event.rb` | Transport message for registry events |
 
 ## Observability
 
@@ -111,14 +130,13 @@ Provider-specific request bodies are not guessed. Partner raw-predict chat reque
 
 ```bash
 bundle install
-bundle exec rspec       # 0 failures
-bundle exec rubocop -A  # auto-fix
-bundle exec rubocop     # lint check
+bundle exec rspec --format json --out tmp/rspec_results.json --format progress --out tmp/rspec_progress.txt
+bundle exec rubocop -A
 ```
 
 ## License
 
-Apache-2.0
+MIT
 
 ## References
 

data/lex-llm-vertex.gemspec

@@ -26,5 +26,6 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'legion-json', '>= 1.2.1'
   spec.add_dependency 'legion-logging', '>= 1.3.2'
   spec.add_dependency 'legion-settings', '>= 1.3.14'
-  spec.add_dependency 'lex-llm', '>= 0.3.0'
+  spec.add_dependency 'legion-transport', '>= 1.4.14'
+  spec.add_dependency 'lex-llm', '>= 0.4.3'
 end

data/lib/legion/extensions/llm/vertex/actors/fleet_worker.rb

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+
+begin
+  require 'legion/extensions/actors/subscription'
+rescue LoadError => e
+  warn(e.message) if $VERBOSE
+end
+
+unless defined?(Legion::Extensions::Actors::Subscription)
+  raise LoadError, 'LegionIO actor runtime is required for Vertex fleet worker'
+end
+
+require 'legion/extensions/llm/vertex'
+require 'legion/extensions/llm/fleet/provider_responder'
+
+module Legion
+  module Extensions
+    module Llm
+      module Vertex
+        module Actor
+          # Subscription actor for Vertex fleet request consumption.
+          class FleetWorker < Legion::Extensions::Actors::Subscription
+            def runner_class
+              'Legion::Extensions::Llm::Vertex::Runners::FleetWorker'
+            end
+
+            def runner_function
+              'handle_fleet_request'
+            end
+
+            def use_runner?
+              false
+            end
+
+            def enabled?
+              Legion::Extensions::Llm::Fleet::ProviderResponder.enabled_for?(Vertex.discover_instances)
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/legion/extensions/llm/vertex/provider.rb

@@ -177,7 +177,16 @@ module Legion
             end
           end
 
-          def chat(messages, model:, temperature: nil, max_tokens: nil, tools: {}, tool_prefs: nil, params: {})
+          def chat(
+            messages:,
+            model:,
+            temperature: nil,
+            max_tokens: nil,
+            tools: {},
+            tool_prefs: nil,
+            params: {},
+            **_provider_options
+          )
             model_id = model_id(model)
             log.info { "chat model=#{model_id} messages=#{messages.size}" }
             @model = model_id
@@ -187,7 +196,8 @@ module Legion
             parse_chat_response(response, model: model_id)
           end
 
-          def stream(messages, model:, temperature: nil, max_tokens: nil, tools: {}, tool_prefs: nil, params: {})
+          def stream(messages:, model:, temperature: nil, max_tokens: nil, tools: {}, tool_prefs: nil, params: {},
+                     **_provider_options)
             model_id = model_id(model)
             log.info { "stream model=#{model_id} messages=#{messages.size}" }
             @model = model_id
@@ -199,7 +209,16 @@ module Legion
             parse_chat_response(response, model: model_id)
           end
 
-          def count_tokens(messages, model:, params: {})
+          def stream_chat(messages:, model:, tools: {}, temperature: nil, max_tokens: nil, params: {}, tool_prefs: nil,
+                          **provider_options, &)
+            stream(messages:, model:, temperature:, max_tokens:, tools:, tool_prefs:, params:, **provider_options, &)
+          end
+
+          def count_tokens(
+            messages:,
+            model:,
+            params: {}
+          )
             model_id = model_id(model)
             log.info { "count_tokens model=#{model_id}" }
             unless generate_content_model?(model_id)
@@ -216,7 +235,15 @@ module Legion
             { input_tokens: response.body['totalTokens'], raw: response.body }
           end
 
-          def embed(text, model:, dimensions: nil, task_type: nil, title: nil, params: {})
+          def embed(
+            text:,
+            model:,
+            dimensions: nil,
+            task_type: nil,
+            title: nil,
+            params: {},
+            **_provider_options
+          )
             model_id = model_id(model)
             log.info { "embed model=#{model_id} inputs=#{Array(text).size}" }
             unless Capabilities.embeddings?(model_id)
@@ -236,9 +263,9 @@ module Legion
             payload[:generationConfig] = Utils.deep_merge(payload[:generationConfig] || {},
                                                           generation_config(temperature, schema, thinking))
             if block_given?
-              stream(messages, model:, temperature:, tools:, tool_prefs:, params: payload, &)
+              stream(messages:, model:, temperature:, tools:, tool_prefs:, params: payload, &)
             else
-              chat(messages, model:, temperature:, tools:, tool_prefs:, params: payload)
+              chat(messages:, model:, temperature:, tools:, tool_prefs:, params: payload)
             end
           end
 
@@ -293,8 +320,8 @@ module Legion
             Legion::Extensions::Llm::Routing::ModelOffering.new(
               provider_family: :vertex,
               instance_id: instance_id,
-              transport: :http,
-              tier: :frontier,
+              transport: configured_transport(:http),
+              tier: configured_tier(:frontier),
               model: model,
               usage_type: usage_type,
               capabilities: default_capabilities(model, api:),
@@ -310,6 +337,14 @@ module Legion
             )
           end
 
+          def configured_transport(default)
+            config.respond_to?(:transport) ? config.transport : default
+          end
+
+          def configured_tier(default)
+            config.respond_to?(:tier) ? config.tier : default
+          end
+
           def publisher_parent
             "projects/#{project}/locations/#{location}/publishers/#{DEFAULT_PUBLISHER}/models"
           end

data/lib/legion/extensions/llm/vertex/runners/fleet_worker.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+require 'legion/extensions/llm/fleet/provider_responder'
+require 'legion/extensions/llm/vertex'
+
+module Legion
+  module Extensions
+    module Llm
+      module Vertex
+        module Runners
+          # Runner entrypoint for Vertex fleet request execution.
+          module FleetWorker
+            module_function
+
+            def handle_fleet_request(payload, delivery: nil, properties: nil)
+              Legion::Extensions::Llm::Fleet::ProviderResponder.call(
+                payload: payload,
+                provider_family: Vertex::PROVIDER_FAMILY,
+                provider_class: Vertex::Provider,
+                provider_instances: -> { Vertex.discover_instances },
+                delivery: delivery,
+                properties: properties
+              )
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/legion/extensions/llm/vertex/version.rb

@@ -4,7 +4,7 @@ module Legion
   module Extensions
     module Llm
       module Vertex
-        VERSION = '0.2.0'
+        VERSION = '0.2.6'
       end
     end
   end

data/lib/legion/extensions/llm/vertex.rb

@@ -16,17 +16,33 @@ module Legion
         PROVIDER_FAMILY = :vertex
 
         def self.default_settings
-          {
-            enabled: false,
-            default_model: nil,
-            project: nil,
-            location: 'us-central1',
-            model_whitelist: [],
-            model_blacklist: [],
-            model_cache_ttl: 3600,
-            tls: { enabled: false, verify: :peer },
-            instances: {}
-          }
+          ::Legion::Extensions::Llm.provider_settings(
+            family: PROVIDER_FAMILY,
+            instance: {
+              endpoint: nil,
+              tier: :frontier,
+              transport: :http,
+              credentials: {
+                access_token: nil,
+                credentials: nil
+              },
+              provider: {
+                project: nil,
+                location: Provider::DEFAULT_LOCATION,
+                model_aliases: {}
+              },
+              usage: { inference: true, embedding: true, image: false },
+              limits: { concurrency: 4 },
+              fleet: {
+                enabled: false,
+                respond_to_requests: false,
+                capabilities: %i[chat stream_chat embed],
+                lanes: [],
+                concurrency: 4,
+                queue_suffix: nil
+              }
+            }
+          )
         end
 
         def self.provider_class
@@ -44,7 +60,7 @@ module Legion
           cfg = CredentialSources.setting(:extensions, :llm, :vertex)
           return unless cfg.is_a?(Hash) && vertex_credentials_present?(cfg)
 
-          instances[:settings] = cfg.except(:instances, 'instances').merge(tier: :cloud)
+          instances[:settings] = normalize_instance_config(cfg).merge(tier: :cloud)
         end
 
         def self.discover_named_instances(instances)
@@ -57,7 +73,7 @@ module Legion
           named.each do |name, config|
             next unless config.is_a?(Hash) && vertex_credentials_present?(config)
 
-            instances[name.to_sym] = config.merge(tier: :cloud)
+            instances[name.to_sym] = normalize_instance_config(config).merge(tier: :cloud)
           end
         end
 
@@ -70,14 +86,33 @@ module Legion
           !(token.nil? && creds.nil?)
         end
 
-        private_class_method :discover_default_instance, :discover_named_instances, :vertex_credentials_present?
+        def self.normalize_instance_config(config)
+          normalized = config.to_h.transform_keys { |key| key.respond_to?(:to_sym) ? key.to_sym : key }
+          normalized[:vertex_project] ||= normalized.delete(:project)
+          normalized[:vertex_location] ||= normalized.delete(:location)
+          normalized[:vertex_api_base] ||= normalized.delete(:base_url)
+          normalized[:vertex_api_base] ||= normalized.delete(:api_base)
+          normalized[:vertex_api_base] ||= normalized.delete(:endpoint)
+          normalized[:vertex_access_token] ||= normalized.delete(:access_token)
+          normalized[:vertex_credentials] ||= normalized.delete(:credentials)
+          normalized[:vertex_model_aliases] ||= normalized.delete(:model_aliases)
+          normalized.compact.except(:instances)
+        end
+
+        def self.register_provider_options
+          configuration = Legion::Extensions::Llm::Configuration
+          if configuration.respond_to?(:register_provider_options)
+            configuration.register_provider_options(Provider.configuration_options)
+          elsif configuration.respond_to?(:option, true)
+            Provider.configuration_options.each { |key| configuration.send(:option, key) }
+          end
+        end
+
+        private_class_method :discover_default_instance, :discover_named_instances, :vertex_credentials_present?,
+                             :normalize_instance_config, :register_provider_options
+
+        register_provider_options
       end
     end
   end
 end
-
-Legion::Extensions::Llm::Configuration.register_provider_options(
-  Legion::Extensions::Llm::Vertex::Provider.configuration_options
-)
-
-Legion::Extensions::Llm::Vertex.register_discovered_instances

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: lex-llm-vertex
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.6
 platform: ruby
 authors:
 - LegionIO
@@ -51,20 +51,34 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 1.3.14
+- !ruby/object:Gem::Dependency
+  name: legion-transport
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.4.14
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.4.14
 - !ruby/object:Gem::Dependency
   name: lex-llm
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.3.0
+        version: 0.4.3
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.3.0
+        version: 0.4.3
 description: Google Cloud Vertex AI provider integration for the LegionIO LLM routing
   framework.
 email:
@@ -84,7 +98,9 @@ files:
 - README.md
 - lex-llm-vertex.gemspec
 - lib/legion/extensions/llm/vertex.rb
+- lib/legion/extensions/llm/vertex/actors/fleet_worker.rb
 - lib/legion/extensions/llm/vertex/provider.rb
+- lib/legion/extensions/llm/vertex/runners/fleet_worker.rb
 - lib/legion/extensions/llm/vertex/version.rb
 homepage: https://github.com/LegionIO/lex-llm-vertex
 licenses: