patient_llm 0.2.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7ba366158096a79e3180cb63c1a03741d88536fb787b42852f2bcb5dc61b0181
-  data.tar.gz: e507e79956d2e3534c5de85646bbfb0973c7253cb4db266ac4c721cea9201cf0
+  metadata.gz: 87bcd18afe8c5a4c46e191946f51201e077b355d223ac9d1817bd4041041143f
+  data.tar.gz: 51a67045f05fc5f55ec1aee9391006eecf60170c3d3e47f3248d2169046ffa0e
 SHA512:
-  metadata.gz: edfb6fc58609f0b7b7733b9e5729877318f2bbe7bdec15a7aeddff05f0a1d75f1c8c47683a3431154a5267e8f5099c1a0271b4520cc3b56fec4df9e1b3228048
-  data.tar.gz: 5075e4c627595ee0d99795726a0071378e6529ca2db9ac11228740781e8e0764d44e44fa624156c30ce4675aa85978f6bdaf93d4693e8068e5305791413cadbf
+  metadata.gz: 594c2deb985655a5396c587e6bce19f603f2f2ae9af3c6df1d8f268f5ac81d9a2542bedeed966f84700ed6f971cc1e1916bf0be3f217397fd07840a794a1f005
+  data.tar.gz: 2e4d20a58e2ce23bb882c1035a98aca401644b85171661b039aa1bc0b7ac5a55a207095ca5facf54723e6ff6baa06e39e752948411dbf3100183bcb041c082f5

data/CHANGELOG.md

@@ -4,6 +4,23 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## 0.4.0
+
+### Added
+
+- Added `preprocessors:` option to provider configuration and `PatientLLM.ask` for applying named PatientHttp request preprocessors (e.g., AWS SigV4 request signing) at dispatch time. A per-request value replaces the provider default; pass an empty array to clear it for a single request.
+
+### Changed
+
+- Requires patient_http version 1.2 or higher.
+
+## 0.3.0
+
+### Changed
+
+- Renamed the `completion_path` argument to `path` on both `PatientLLM.ask` and provider configuration. The `completion_path` name is still accepted as a deprecated alias and now emits a deprecation warning.
+- The path is now joined with the base URL using `URI.join` to ensure proper handling of relative paths in the `path` argument. If the path starts with a slash, it will be treated as an absolute path and will replace the base URL's path. If it does not start with a slash, it will be treated as a relative path and will be appended to the base URL's path.
+
 ## 0.2.0
 
 ### Added

data/README.md

@@ -52,6 +52,38 @@ end
 > [!NOTE]
 > You can also set up encryption for your job payloads to ensure the entire serialized payload is always encrypted in the job queue. See the documentation for [patient_http-sidekiq](https://github.com/bdurand/patient_http-sidekiq#sensitive-data-handling) or [patient_http-solid_queue](https://github.com/bdurand/patient_http-solid_queue#sensitive-data-handling) for details.
 
+### Request signing (preprocessors)
+
+Some providers require request signing rather than a static authentication header — for example, AWS Bedrock with SigV4, where a signature is computed over the final outgoing request. For these, register a [request preprocessor](https://github.com/bdurand/patient_http#request-preprocessors) on the PatientHttp configuration and reference it by name from the provider. Like secrets, only the preprocessor name is serialized into the job queue; the signing logic and credentials stay on the processor side.
+
+```ruby
+PatientLLM.configure do |config|
+  config.provider :bedrock,
+    url: "https://bedrock-runtime.us-east-1.amazonaws.com",
+    serializer: :converse,
+    preprocessors: :aws_sigv4
+end
+
+PatientHttp::Sidekiq.configure do |config|
+  config.register_preprocessor(:aws_sigv4) do |request|
+    signer = Aws::Sigv4::Signer.new(
+      service: "bedrock",
+      region: "us-east-1",
+      credentials_provider: Aws::CredentialProviderChain.new.resolve
+    )
+    signature = signer.sign_request(
+      http_method: request.http_method.to_s.upcase,
+      url: request.url,
+      headers: request.headers.to_h,
+      body: request.body.to_s
+    )
+    signature.headers.each { |name, value| request.headers[name] = value }
+  end
+end
+```
+
+Multiple preprocessors can be given as an array; they run in order at dispatch time.
+
 ### Creating a Callback Class
 
 Create a callback class with `on_complete` and `on_error` methods. Callbacks receive
@@ -185,34 +217,37 @@ session.max_output_tokens = 1000
 PatientLLM.ask(session,
   provider: :openai,
   callback: LLMCallback,
-  url: "http://localhost:1234",           # Override the provider's base URL
+  url: "http://localhost:1234",            # Override the provider's base URL
   serializer: :messages,                   # Override the API format
-  completion_path: "/chat/completions",    # Override the endpoint path
+  path: "/chat/completions",               # Override the endpoint path
   headers: {"X-Custom" => "value"},        # Additional HTTP headers
-  params: {max_completion_tokens: 1000}    # Additional request parameters
+  params: {max_completion_tokens: 1000},   # Additional request parameters
+  preprocessors: :aws_sigv4                # Replace the provider's request preprocessors
 )
 ```
 
+Note that `headers` and `params` are merged on top of the provider's configured values, while `preprocessors` (like `url`, `serializer`, and `path`) replaces the provider default. Pass `preprocessors: []` to clear a provider-level preprocessor for a single request.
+
 ### URL composition
 
-The full request URL is built by concatenating the base URL (from the provider registry or the `url:` option) with the `completion_path`. When you don't set `completion_path`, it defaults to the path for the active serializer (`/v1/chat/completions` for `:chat_completion`, `/v1/responses` for `:open_responses`, `/v1/messages` for `:messages`, `/converse` for `:converse`, `/v1beta/models/{model}:generateContent` for `:gemini`). A `{model}` placeholder in the path is replaced with the session's model at dispatch time, which is how the Gemini default targets Google's `/v1beta/models/{model}:generateContent` endpoint. Trailing slashes on the base and leading slashes on the path are normalized, so:
+The full request URL is built by concatenating the base URL (from the provider registry or the `url:` option) with the `path`. When you don't set `path`, it defaults to the path for the active serializer (`/v1/chat/completions` for `:chat_completion`, `/v1/responses` for `:open_responses`, `/v1/messages` for `:messages`, `/converse` for `:converse`, `/v1beta/models/{model}:generateContent` for `:gemini`). A `{model}` placeholder in the path is replaced with the session's model at dispatch time, which is how the Gemini default targets Google's `/v1beta/models/{model}:generateContent` endpoint. Trailing slashes on the base and leading slashes on the path are normalized, so:
 
 ```
-url = "https://api.openai.com"            completion_path = "/v1/chat/completions"
+url = "https://api.openai.com"            path = "/v1/chat/completions"
 -> https://api.openai.com/v1/chat/completions
 
-url = "http://localhost:1234"             completion_path = "/v1/chat/completions"
+url = "http://localhost:1234"             path = "/v1/chat/completions"
 -> http://localhost:1234/v1/chat/completions
 ```
 
-If your base URL already includes a `/v1` prefix, override the completion path to avoid duplication:
+If your base URL already includes a `/v1` prefix, override the path to avoid duplication:
 
 ```ruby
 PatientLLM.ask(session,
   provider: :openai,
   callback: LLMCallback,
   url: "https://my-gateway.internal/openai/v1",
-  completion_path: "/chat/completions"
+  path: "chat/completions"
 )
 ```
 

data/VERSION

@@ -1 +1 @@
-0.2.0
+0.4.0

data/lib/patient_llm/callback.rb

@@ -213,9 +213,10 @@ module PatientLLM
       # Restore per-request overrides
       ask_kwargs[:url] = request_options["url"] if request_options["url"]
       ask_kwargs[:serializer] = request_options["serializer"].to_sym if request_options["serializer"]
-      ask_kwargs[:completion_path] = request_options["completion_path"] if request_options["completion_path"]
+      ask_kwargs[:path] = request_options["path"] if request_options["path"]
       ask_kwargs[:headers] = request_options["headers"] if request_options["headers"]
       ask_kwargs[:params] = request_options["params"] if request_options["params"]
+      ask_kwargs[:preprocessors] = request_options["preprocessors"] if request_options["preprocessors"]
 
       PatientLLM.ask(session, **ask_kwargs)
     end

data/lib/patient_llm/configuration.rb

@@ -26,10 +26,19 @@ module PatientLLM
     # @param url [String] Base URL for the provider API
     # @param headers [Hash] Default headers for requests
     # @param serializer [Symbol] API format (:chat_completion, :open_responses, :messages, :converse, :gemini)
-    # @param completion_path [String, nil] Override the default endpoint path
+    # @param path [String, nil] Override the default endpoint path
+    # @param completion_path [String, nil] Deprecated alias for +path+
     # @param params [Hash] Additional parameters to merge into every request payload
+    # @param preprocessors [String, Symbol, Array<String, Symbol>, nil] Names of request
+    #   preprocessors to apply to every request (e.g. for request signing). Names must be
+    #   registered on the PatientHttp configuration with `register_preprocessor`.
     # @return [void]
-    def provider(name, url:, headers: {}, serializer: :chat_completion, completion_path: nil, params: {})
+    def provider(name, url:, headers: {}, serializer: :chat_completion, path: nil, completion_path: nil, params: {}, preprocessors: nil)
+      if completion_path
+        warn "PatientLLM::Configuration#provider: the `completion_path:` argument is deprecated; use `path:` instead", uplevel: 1
+        path ||= completion_path
+      end
+
       sym = serializer.to_sym
       unless PatientLLM::VALID_SERIALIZERS.include?(sym)
         raise ArgumentError, "Unknown serializer: #{sym.inspect}. Valid options: #{PatientLLM::VALID_SERIALIZERS.map(&:inspect).join(", ")}"
@@ -41,8 +50,9 @@ module PatientLLM
         url: url,
         headers: headers,
         serializer: sym,
-        completion_path: completion_path,
-        params: params
+        path: path,
+        params: params,
+        preprocessors: preprocessors
       }
     end
 

data/lib/patient_llm.rb

@@ -2,6 +2,7 @@
 
 require "patient_http"
 require "prompt_builder"
+require "uri"
 
 module PatientLLM
   VERSION = File.read(File.join(__dir__, "../VERSION")).strip
@@ -16,11 +17,11 @@ module PatientLLM
   # dispatch time, matching Google's `/v1beta/models/{model}:generateContent`
   # endpoint.
   SERIALIZER_PATHS = {
-    chat_completion: "/v1/chat/completions",
-    open_responses: "/v1/responses",
-    messages: "/v1/messages",
-    converse: "/converse",
-    gemini: "/v1beta/models/{model}:generateContent"
+    chat_completion: "v1/chat/completions",
+    open_responses: "v1/responses",
+    messages: "v1/messages",
+    converse: "converse",
+    gemini: "v1beta/models/{model}:generateContent"
   }.freeze
 
   # Required version header for the Anthropic Messages API.
@@ -62,11 +63,20 @@ module PatientLLM
     # @param callback_args [Hash] Custom arguments passed through to the callback
     # @param url [String, nil] Override the provider's base URL for this request
     # @param serializer [Symbol, nil] Override the provider's serializer for this request
-    # @param completion_path [String, nil] Override the endpoint path for this request
+    # @param path [String, nil] Override the endpoint path for this request
+    # @param completion_path [String, nil] Deprecated alias for +path+
     # @param headers [Hash, nil] Additional headers merged on top of provider headers
     # @param params [Hash, nil] Additional params merged into the request payload
+    # @param preprocessors [String, Symbol, Array<String, Symbol>, nil] Names of request
+    #   preprocessors to apply to this request. Replaces the provider's configured
+    #   preprocessors; pass an empty array to clear them for this request.
     # @return [Object] Handler-specific identifier for the enqueued request
-    def ask(session, provider:, callback:, callback_args: {}, url: nil, serializer: nil, completion_path: nil, headers: nil, params: nil, tool_iteration: 0, original_request_id: nil) # :nodoc: tool_iteration and original_request_id are internal
+    def ask(session, provider:, callback:, callback_args: {}, url: nil, serializer: nil, path: nil, completion_path: nil, headers: nil, params: nil, preprocessors: nil, tool_iteration: 0, original_request_id: nil) # :nodoc: tool_iteration and original_request_id are internal
+      if completion_path
+        warn "PatientLLM.ask: the `completion_path:` argument is deprecated; use `path:` instead", uplevel: 1
+        path ||= completion_path
+      end
+
       provider_config = self.provider(provider) || {}
       provider_name = provider.to_s
 
@@ -79,32 +89,35 @@ module PatientLLM
 
       resolved_serializer = (serializer || provider_config[:serializer] || :chat_completion).to_sym
       validate_serializer!(resolved_serializer)
-      resolved_completion_path = completion_path || provider_config[:completion_path] || SERIALIZER_PATHS[resolved_serializer] || "/v1/chat/completions"
-      if resolved_completion_path.include?("{model}")
-        resolved_completion_path = resolved_completion_path.gsub("{model}", session.model.to_s)
+      resolved_path = path || provider_config[:path] || SERIALIZER_PATHS[resolved_serializer] || "/v1/chat/completions"
+      if resolved_path.include?("{model}")
+        resolved_path = resolved_path.gsub("{model}", session.model.to_s)
       end
       resolved_headers = (provider_config[:headers] || {}).merge(headers || {})
       if resolved_serializer == :messages && !resolved_headers.key?("anthropic-version")
         resolved_headers = {"anthropic-version" => ANTHROPIC_VERSION}.merge(resolved_headers)
       end
       resolved_params = (provider_config[:params] || {}).merge(params || {})
+      resolved_preprocessors = preprocessors || provider_config[:preprocessors]
 
       payload = session.request_payload(resolved_serializer)
       payload = deep_merge(payload, deep_stringify_keys(resolved_params)) unless resolved_params.empty?
 
-      request_url = join_url(resolved_url, resolved_completion_path)
+      request_url = join_url(resolved_url, resolved_path)
 
       request_options = {}
       request_options["url"] = url if url
       request_options["serializer"] = serializer.to_s if serializer
-      request_options["completion_path"] = completion_path if completion_path
+      request_options["path"] = path if path
       request_options["headers"] = headers if headers && !headers.empty?
       request_options["params"] = params if params && !params.empty?
+      request_options["preprocessors"] = preprocessors if preprocessors
 
       PatientHttp.post(
         request_url,
         json: payload,
         headers: resolved_headers,
+        preprocessors: resolved_preprocessors,
         raise_error_responses: true,
         callback: PatientLLM::Callback,
         callback_args: {
@@ -128,7 +141,9 @@ module PatientLLM
     end
 
     def join_url(base, path)
-      "#{base.sub(%r{/\z}, "")}/#{path.to_s.sub(%r{\A/}, "")}"
+      base_uri = URI.parse(base)
+      base_uri.path = "#{base_uri.path}/" unless base_uri.path&.end_with?("/")
+      URI.join(base_uri, path).to_s
     end
 
     def deep_merge(hash1, hash2)

data/patient_llm.gemspec

@@ -37,7 +37,7 @@ Gem::Specification.new do |spec|
 
   spec.required_ruby_version = ">= 3.0"
 
-  spec.add_dependency "patient_http", "~> 1.1"
+  spec.add_dependency "patient_http", "~> 1.2"
   spec.add_dependency "prompt_builder"
 
   spec.add_development_dependency "bundler"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: patient_llm
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Brian Durand
@@ -15,14 +15,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
+        version: '1.2'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
+        version: '1.2'
 - !ruby/object:Gem::Dependency
   name: prompt_builder
   requirement: !ruby/object:Gem::Requirement