boxcars 0.6.4 → 0.6.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c7288e4aaea4065096105c84b78e9425330847eba37b55b6479466c55ae840a8
-  data.tar.gz: 6ff9ec0d0c69113dc70360750fa8fe5460236401c75a3b4e681009efad733301
+  metadata.gz: 164d55b95e134a691b4525f778a83f6d06cdaa1948028b87371cb6cb8a16fd43
+  data.tar.gz: 2707ac063bbc1127d75398be8770812ae258447ba021c0b9527cd3fbf113ff02
 SHA512:
-  metadata.gz: e48e4debf178dca7647a08357eaabbdd8ceef6edc3ae40d49a25a31b8ee21fbc41ded409fd55c2b840762868546f1edb0ada72729178c15953e6f48cc0d832fe
-  data.tar.gz: 3717955095a68207dd7c0ee914132573d84126c01ea4e53cec0d5a46985f512a59da3fb1b78ededac29c0d72d16e5f70301629ab326aa20ec2feb4b6664ae343
+  metadata.gz: 8b0d7909ef7c049d7c393b646abd072c01dd4b295a7af909bbe2f85fe48b8425c946f956440548fa3902182c636d3e57e81279332d331a9d7c4df4181cd4e4d5
+  data.tar.gz: ed1c903fb5575ae563ccab8adb0e80cfed9c00e529a5850fc7b89acaf165eecaf42e69cdd548cdec3622855645b5bf88591b3d6a6cb1af47bfbe97d5d6d16ad9

data/.ruby-version

@@ -1 +1 @@
-3.3.3
+3.3.5

data/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## [v0.6.4](https://github.com/BoxcarsAI/boxcars/tree/v0.6.4) (2024-07-27)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.6.3...v0.6.4)
+
+**Merged pull requests:**
+
+- Add Ollama Engine [\#200](https://github.com/BoxcarsAI/boxcars/pull/200) ([francis](https://github.com/francis))
+
 ## [v0.6.3](https://github.com/BoxcarsAI/boxcars/tree/v0.6.3) (2024-07-26)
 
 [Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.6.2...v0.6.3)

data/Gemfile

@@ -15,13 +15,13 @@ gem "sqlite3", "~> 1.7"
 
 gem "async", "~>1.32.1"
 
-gem "activerecord", "~> 7.0"
+gem "activerecord", "~> 7.1"
 
 gem "github_changelog_generator", "~> 1.16"
 
 gem "faraday-retry", "~> 2.0"
 
-gem "activesupport", "~> 7.0"
+gem "activesupport", "~> 7.1"
 
 gem "rest-client", "~> 2.1"
 
@@ -32,9 +32,9 @@ gem "pgvector", "~> 0.2.2"
 
 group :development, :test do
   gem "rspec", "~> 3.13"
-  gem "rubocop", "~> 1.64"
-  gem "vcr", "~> 6.2.0"
+  gem "rubocop", "~> 1.66"
+  gem "vcr", "~> 6.3.1"
   gem "webmock", "~> 3.23.1"
   gem "rubocop-rake", "~> 0.6.0"
-  gem "rubocop-rspec", "~> 3.0"
+  gem "rubocop-rspec", "~> 3.1"
 end

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    boxcars (0.6.4)
+    boxcars (0.6.5)
       anthropic (~> 0.1)
       google_search_results (~> 2.2)
       gpt4all (~> 0.0.4)
@@ -13,13 +13,13 @@ PATH
 GEM
   remote: https://rubygems.org/
   specs:
-    activemodel (7.1.3.4)
-      activesupport (= 7.1.3.4)
-    activerecord (7.1.3.4)
-      activemodel (= 7.1.3.4)
-      activesupport (= 7.1.3.4)
+    activemodel (7.1.4)
+      activesupport (= 7.1.4)
+    activerecord (7.1.4)
+      activemodel (= 7.1.4)
+      activesupport (= 7.1.4)
       timeout (>= 0.4.0)
-    activesupport (7.1.3.4)
+    activesupport (7.1.4)
       base64
       bigdecimal
       concurrent-ruby (~> 1.0, >= 1.0.2)
@@ -57,7 +57,7 @@ GEM
       async (>= 1.25)
     base64 (0.2.0)
     bigdecimal (3.1.8)
-    concurrent-ruby (1.3.3)
+    concurrent-ruby (1.3.4)
     connection_pool (2.4.1)
     console (1.27.0)
       fiber-annotation
@@ -71,7 +71,7 @@ GEM
       reline (>= 0.3.8)
     diff-lcs (1.5.1)
     domain_name (0.6.20240107)
-    dotenv (3.1.2)
+    dotenv (3.1.4)
     drb (2.2.1)
     event_stream_parser (1.0.0)
     faraday (2.10.0)
@@ -120,7 +120,7 @@ GEM
     mime-types (3.5.2)
       mime-types-data (~> 3.2015)
     mime-types-data (3.2024.0702)
-    minitest (5.24.1)
+    minitest (5.25.1)
     multi_json (1.15.0)
     multipart-post (2.4.1)
     mutex_m (0.2.0)
@@ -128,19 +128,19 @@ GEM
       uri
     netrc (0.11.0)
     nio4r (2.7.3)
-    nokogiri (1.16.6-arm64-darwin)
+    nokogiri (1.16.7-arm64-darwin)
       racc (~> 1.4)
-    nokogiri (1.16.6-x86_64-linux)
+    nokogiri (1.16.7-x86_64-linux)
       racc (~> 1.4)
     octokit (4.25.1)
       faraday (>= 1, < 3)
       sawyer (~> 0.9)
     os (1.1.4)
-    parallel (1.25.1)
-    parser (3.3.4.0)
+    parallel (1.26.3)
+    parser (3.3.5.0)
       ast (~> 2.4.1)
       racc
-    pg (1.5.6)
+    pg (1.5.8)
     pgvector (0.2.2)
     protocol-hpack (1.4.3)
     protocol-http (0.26.8)
@@ -152,7 +152,7 @@ GEM
     psych (5.1.2)
       stringio
     public_suffix (6.0.0)
-    racc (1.8.0)
+    racc (1.8.1)
     rainbow (3.1.1)
     rake (13.2.1)
     rdoc (6.7.0)
@@ -165,8 +165,7 @@ GEM
       http-cookie (>= 1.0.2, < 2.0)
       mime-types (>= 1.16, < 4.0)
       netrc (~> 0.8)
-    rexml (3.3.2)
-      strscan
+    rexml (3.3.8)
     rspec (3.13.0)
       rspec-core (~> 3.13.0)
       rspec-expectations (~> 3.13.0)
@@ -180,22 +179,21 @@ GEM
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
     rspec-support (3.13.1)
-    rubocop (1.65.0)
+    rubocop (1.66.1)
       json (~> 2.3)
       language_server-protocol (>= 3.17.0)
       parallel (~> 1.10)
       parser (>= 3.3.0.2)
       rainbow (>= 2.2.2, < 4.0)
       regexp_parser (>= 2.4, < 3.0)
-      rexml (>= 3.2.5, < 4.0)
-      rubocop-ast (>= 1.31.1, < 2.0)
+      rubocop-ast (>= 1.32.2, < 2.0)
       ruby-progressbar (~> 1.7)
       unicode-display_width (>= 2.4.0, < 3.0)
-    rubocop-ast (1.31.3)
+    rubocop-ast (1.32.3)
       parser (>= 3.3.1.0)
     rubocop-rake (0.6.0)
       rubocop (~> 1.0)
-    rubocop-rspec (3.0.3)
+    rubocop-rspec (3.1.0)
       rubocop (~> 1.61)
     ruby-openai (7.1.0)
       event_stream_parser (>= 0.3.0, < 2.0.0)
@@ -209,7 +207,6 @@ GEM
     sqlite3 (1.7.3-x86_64-linux)
     stringio (3.1.1)
     strings-ansi (0.2.0)
-    strscan (3.1.0)
     timeout (0.4.1)
     timers (4.3.5)
     traces (0.11.1)
@@ -222,9 +219,10 @@ GEM
     tty-screen (0.8.2)
     tzinfo (2.0.6)
       concurrent-ruby (~> 1.0)
-    unicode-display_width (2.5.0)
+    unicode-display_width (2.6.0)
     uri (0.13.0)
-    vcr (6.2.0)
+    vcr (6.3.1)
+      base64
     webmock (3.23.1)
       addressable (>= 2.8.0)
       crack (>= 0.3.2)
@@ -233,11 +231,12 @@ GEM
 PLATFORMS
   arm64-darwin-22
   arm64-darwin-23
+  arm64-darwin-24
   x86_64-linux
 
 DEPENDENCIES
-  activerecord (~> 7.0)
-  activesupport (~> 7.0)
+  activerecord (~> 7.1)
+  activesupport (~> 7.1)
   async (~> 1.32.1)
   boxcars!
   debug (~> 1.9)
@@ -250,11 +249,11 @@ DEPENDENCIES
   rake (~> 13.2)
   rest-client (~> 2.1)
   rspec (~> 3.13)
-  rubocop (~> 1.64)
+  rubocop (~> 1.66)
   rubocop-rake (~> 0.6.0)
-  rubocop-rspec (~> 3.0)
+  rubocop-rspec (~> 3.1)
   sqlite3 (~> 1.7)
-  vcr (~> 6.2.0)
+  vcr (~> 6.3.1)
   webmock (~> 3.23.1)
 
 BUNDLED WITH

data/lib/boxcars/boxcar/json_engine_boxcar.rb

@@ -50,12 +50,14 @@ module Boxcars
     # @param engine_output [String] The output from the engine.
     # @return [Result] The result.
     def get_answer(engine_output)
-      # sometimes the LLM adds text in front of the JSON output, so let's strip it here
-      json_start = engine_output.index("{")
-      json_end = engine_output.rindex("}")
-      extract_answer(JSON.parse(engine_output[json_start..json_end], symbolize_names: symbolize))
+      json_string = extract_json(engine_output)
+      reply = JSON.parse(json_string, symbolize_names: symbolize)
+      Result.new(status: :ok, answer: reply, explanation: reply)
+    rescue JSON::ParserError => e
+      Boxcars.debug "JSON: #{engine_output}", :red
+      Result.from_error("JSON parsing error: #{e.message}")
     rescue StandardError => e
-      Result.from_error("Error: #{e.message}:\n#{engine_output}")
+      Result.from_error("Unexpected error: #{e.message}")
     end
 
     # get answer from parsed JSON
@@ -63,14 +65,29 @@ module Boxcars
     # @return [Result] The result.
     def extract_answer(data)
       reply = data
+      Result.new(status: :ok, answer: reply, explanation: reply)
+    end
+
+    private
+
+    def extract_json(text)
+      # Escape control characters (U+0000 to U+001F)
+      text = text.gsub(/[\u0000-\u001F]/, '')
+      # first strip hidden characters
+      # text = text.encode('UTF-8', invalid: :replace, undef: :replace, replace: '')
+
+      # sometimes the LLM adds text in front of the JSON output, so let's strip it here
+      json_start = text.index("{")
+      json_end = text.rindex("}")
+      text[json_start..json_end]
+    end
+
+    def extract_json2(text)
+      # Match the outermost JSON object
+      match = text.match(/\{(?:[^{}]|\{(?:[^{}]|\{[^{}]*\})*\})*\}/)
+      raise StandardError, "No valid JSON object found in the output" unless match
 
-      if reply.present?
-        Result.new(status: :ok, answer: reply, explanation: reply)
-      else
-        # we have an unexpected output from the engine
-        Result.new(status: :error, answer: nil,
-                   explanation: "You gave me an improperly formatted answer. I was expecting a valid reply.")
-      end
+      match[0]
     end
   end
 end

data/lib/boxcars/boxcar.rb

@@ -166,7 +166,7 @@ module Boxcars
         output = call(inputs: inputs)
       rescue StandardError => e
         Boxcars.error "Error in #{name} boxcar#call: #{e}\nbt:#{e.backtrace[0..5].join("\n   ")}", :red
-        Boxcars.error("Response Body: #{e.response[:body]}", :red) if e.respond_to?(:response)
+        Boxcars.error("Response Body: #{e.response[:body]}", :red) if e.respond_to?(:response) && e.response.present?
         raise e
       end
       validate_outputs(outputs: output.keys)

data/lib/boxcars/engine/openai.rb

@@ -28,6 +28,11 @@ module Boxcars
     # @param batch_size [Integer] The number of prompts to send to the engine at once. Defaults to 20.
     def initialize(name: DEFAULT_NAME, description: DEFAULT_DESCRIPTION, prompts: [], batch_size: 20, **kwargs)
       @open_ai_params = DEFAULT_PARAMS.merge(kwargs)
+      if @open_ai_params[:model] =~ /^o/ && @open_ai_params[:max_tokens].present?
+        @open_ai_params[:max_completion_tokens] = @open_ai_params.delete(:max_tokens)
+        @open_ai_params.delete(:temperature)
+      end
+
       @prompts = prompts
       @batch_size = batch_size
       super(description: description, name: name)
@@ -44,7 +49,7 @@ module Boxcars
     end
 
     def conversation_model?(model)
-      !!(model =~ /(^gpt-4)|(-turbo\b)/)
+      !!(model =~ /(^gpt-4)|(-turbo\b)|(^o\d)/)
     end
 
     # Get an answer from the engine.
@@ -57,6 +62,10 @@ module Boxcars
       params = open_ai_params.merge(kwargs)
       if conversation_model?(params[:model])
         prompt = prompt.first if prompt.is_a?(Array)
+        if params[:model] =~ /^o/
+          params.delete(:response_format)
+          params.delete(:stop)
+        end
         params = prompt.as_messages(inputs).merge(params)
         if Boxcars.configuration.log_prompts
           Boxcars.debug(params[:messages].last(2).map { |p| ">>>>>> Role: #{p[:role]} <<<<<<\n#{p[:content]}" }.join("\n"), :cyan)

data/lib/boxcars/engine/perplexityai.rb

@@ -40,6 +40,7 @@ module Boxcars
         faraday.request :json
         faraday.response :json
         faraday.response :raise_error
+        # faraday.options.timeout = 180 # 3 minutes
       end
 
       response = conn.post do |req|
@@ -122,4 +123,51 @@ module Boxcars
   def max_tokens_for_prompt(_prompt_text)
     8096
   end
+
+  # Get generation informaton
+  # @param sub_choices [Array<Hash>] The choices to get generation info for.
+  # @return [Array<Generation>] The generation information.
+  def generation_info(sub_choices)
+    sub_choices.map do |choice|
+      Generation.new(
+        text: choice.dig("message", "content") || choice["text"],
+        generation_info: {
+          finish_reason: choice.fetch("finish_reason", nil),
+          logprobs: choice.fetch("logprobs", nil)
+        }
+      )
+    end
+  end
+
+  # Call out to endpoint with k unique prompts.
+  # @param prompts [Array<String>] The prompts to pass into the model.
+  # @param inputs [Array<String>] The inputs to subsitite into the prompt.
+  # @param stop [Array<String>] Optional list of stop words to use when generating.
+  # @return [EngineResult] The full engine output.
+  def generate(prompts:, stop: nil)
+    params = {}
+    params[:stop] = stop if stop
+    choices = []
+    token_usage = {}
+    # Get the token usage from the response.
+    # Includes prompt, completion, and total tokens used.
+    inkeys = %w[completion_tokens prompt_tokens total_tokens].freeze
+    prompts.each_slice(batch_size) do |sub_prompts|
+      sub_prompts.each do |sprompts, inputs|
+        response = client(prompt: sprompts, inputs: inputs, **params)
+        check_response(response)
+        choices.concat(response["choices"])
+        usage_keys = inkeys & response["usage"].keys
+        usage_keys.each { |key| token_usage[key] = token_usage[key].to_i + response["usage"][key] }
+      end
+    end
+
+    n = params.fetch(:n, 1)
+    generations = []
+    prompts.each_with_index do |_prompt, i|
+      sub_choices = choices[i * n, (i + 1) * n]
+      generations.push(generation_info(sub_choices))
+    end
+    EngineResult.new(generations: generations, engine_output: { token_usage: token_usage })
+  end
 end

data/lib/boxcars/engine.rb

@@ -68,7 +68,6 @@ module Boxcars
       end
       EngineResult.new(generations: generations, engine_output: { token_usage: token_usage })
     end
-    # rubocop:enable Metrics/AbcSize
   end
 end
 

data/lib/boxcars/version.rb

@@ -2,5 +2,5 @@
 
 module Boxcars
   # The current version of the gem.
-  VERSION = "0.6.4"
+  VERSION = "0.6.5"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: boxcars
 version: !ruby/object:Gem::Version
-  version: 0.6.4
+  version: 0.6.5
 platform: ruby
 authors:
 - Francis Sullivan
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-07-27 00:00:00.000000000 Z
+date: 2024-10-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: anthropic
@@ -219,7 +219,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.11
+rubygems_version: 3.5.20
 signing_key:
 specification_version: 4
 summary: Boxcars is a gem that enables you to create new systems with AI composability.