boxcars 0.6.6 → 0.6.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 93deffd0bbce00b58479a93a7c20afdd59fa029bff798d3e90abafb37569ae83
-  data.tar.gz: 314f504e1b061856c90951730b9d3c8465174eaa007363a12c8805ee1e36fc32
+  metadata.gz: b025b4faf290736766df116e4146cdcea5285bfbd54b6b1c8eaae1e3fc220507
+  data.tar.gz: cdc51dd2f92609b3b2d52879de4716fe8a34e20a15c9bc5d9f27c005bd5f760f
 SHA512:
-  metadata.gz: 29fc4a409bd98e4528142e295a09727b3a5b0c156343d365601f07852f5f09333ebe6a4a1c798e2e0890420e7414c6be5fa17dd02aa40f528375a207fd74da04
-  data.tar.gz: a317992e392cda328242487910713cd82968ca371783072564379d3a806c1a8e831d00eace9ea6df80da4bc569526f97ecd1b1a6c133b4b5d7bb43ee8c1b305a
+  metadata.gz: 4267533089c1c3be564ac59d321a8adb670d4197d0548767be1291835049c1edf86f6994b93186d4b0c5ab49445f17d1bdba35bc3b983eafea150c0b05732775
+  data.tar.gz: 3d37938ea2c431fe3f1b839471421fc5af46e0c60ede688265874d8b1ac93a6e58e440509a66aea4c91375986ed0295a477cca3adca75f68ffc028f33976583a

data/CHANGELOG.md

@@ -1,9 +1,35 @@
 # Changelog
 
+## [v0.6.7](https://github.com/BoxcarsAI/boxcars/tree/v0.6.7) (2024-12-04)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.6.6...v0.6.7)
+
+## [v0.6.6](https://github.com/BoxcarsAI/boxcars/tree/v0.6.6) (2024-11-15)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.6.5...v0.6.6)
+
+**Merged pull requests:**
+
+- update anthropic prompt output [\#235](https://github.com/BoxcarsAI/boxcars/pull/235) ([francis](https://github.com/francis))
+- \[infra\] Bump rubocop-rspec from 3.1.0 to 3.2.0 [\#230](https://github.com/BoxcarsAI/boxcars/pull/230) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump rexml from 3.3.8 to 3.3.9 [\#229](https://github.com/BoxcarsAI/boxcars/pull/229) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump activerecord from 7.1.4 to 7.1.4.2 [\#228](https://github.com/BoxcarsAI/boxcars/pull/228) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump activesupport from 7.1.4 to 7.1.4.2 [\#227](https://github.com/BoxcarsAI/boxcars/pull/227) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump rubocop from 1.66.1 to 1.67.0 [\#224](https://github.com/BoxcarsAI/boxcars/pull/224) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump ruby-openai from 7.1.0 to 7.3.1 [\#223](https://github.com/BoxcarsAI/boxcars/pull/223) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump anthropic from 0.3.0 to 0.3.2 [\#220](https://github.com/BoxcarsAI/boxcars/pull/220) ([dependabot[bot]](https://github.com/apps/dependabot))
+- Dep auto [\#219](https://github.com/BoxcarsAI/boxcars/pull/219) ([francis](https://github.com/francis))
+- \[infra\] Bump webmock from 3.23.1 to 3.24.0 [\#218](https://github.com/BoxcarsAI/boxcars/pull/218) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump sqlite3 from 1.7.3 to 2.0.4 [\#206](https://github.com/BoxcarsAI/boxcars/pull/206) ([dependabot[bot]](https://github.com/apps/dependabot))
+
 ## [v0.6.5](https://github.com/BoxcarsAI/boxcars/tree/v0.6.5) (2024-10-04)
 
 [Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.6.4...v0.6.5)
 
+**Implemented enhancements:**
+
+- use sample sql data in the prompts [\#11](https://github.com/BoxcarsAI/boxcars/issues/11)
+
 **Closed issues:**
 
 - Consider methods to prevent hallucinations and incorrect answers [\#89](https://github.com/BoxcarsAI/boxcars/issues/89)

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    boxcars (0.6.6)
+    boxcars (0.6.8)
       anthropic (~> 0.1)
       google_search_results (~> 2.2)
       gpt4all (~> 0.0.4)
@@ -13,22 +13,24 @@ PATH
 GEM
   remote: https://rubygems.org/
   specs:
-    activemodel (7.1.4.2)
-      activesupport (= 7.1.4.2)
-    activerecord (7.1.4.2)
-      activemodel (= 7.1.4.2)
-      activesupport (= 7.1.4.2)
+    activemodel (7.2.2)
+      activesupport (= 7.2.2)
+    activerecord (7.2.2)
+      activemodel (= 7.2.2)
+      activesupport (= 7.2.2)
       timeout (>= 0.4.0)
-    activesupport (7.1.4.2)
+    activesupport (7.2.2)
       base64
+      benchmark (>= 0.3)
       bigdecimal
-      concurrent-ruby (~> 1.0, >= 1.0.2)
+      concurrent-ruby (~> 1.0, >= 1.3.1)
       connection_pool (>= 2.2.5)
       drb
       i18n (>= 1.6, < 2)
+      logger (>= 1.4.2)
       minitest (>= 5.1)
-      mutex_m
-      tzinfo (~> 2.0)
+      securerandom (>= 0.3)
+      tzinfo (~> 2.0, >= 2.0.5)
     addressable (2.8.7)
       public_suffix (>= 2.0.2, < 7.0)
     anthropic (0.3.2)
@@ -56,6 +58,7 @@ GEM
     async-pool (0.7.0)
       async (>= 1.25)
     base64 (0.2.0)
+    benchmark (0.4.0)
     bigdecimal (3.1.8)
     concurrent-ruby (1.3.4)
     connection_pool (2.4.1)
@@ -74,16 +77,16 @@ GEM
     dotenv (3.1.4)
     drb (2.2.1)
     event_stream_parser (1.0.0)
-    faraday (2.12.0)
-      faraday-net_http (>= 2.0, < 3.4)
+    faraday (2.12.1)
+      faraday-net_http (>= 2.0, < 3.5)
       json
       logger
     faraday-http-cache (2.5.1)
       faraday (>= 0.8)
     faraday-multipart (1.0.4)
       multipart-post (~> 2)
-    faraday-net_http (3.3.0)
-      net-http
+    faraday-net_http (3.4.0)
+      net-http (>= 0.5.0)
     faraday-retry (2.2.1)
       faraday (~> 2.0)
     fiber-annotation (0.2.0)
@@ -115,33 +118,32 @@ GEM
     irb (1.14.0)
       rdoc (>= 4.0.0)
       reline (>= 0.4.2)
-    json (2.7.5)
+    json (2.9.0)
     language_server-protocol (3.17.0.3)
-    logger (1.6.1)
+    logger (1.6.2)
     mime-types (3.5.2)
       mime-types-data (~> 3.2015)
     mime-types-data (3.2024.0702)
-    minitest (5.25.1)
+    minitest (5.25.4)
     multi_json (1.15.0)
     multipart-post (2.4.1)
-    mutex_m (0.2.0)
-    net-http (0.4.1)
+    net-http (0.6.0)
       uri
     netrc (0.11.0)
     nio4r (2.7.3)
-    nokogiri (1.16.7-arm64-darwin)
+    nokogiri (1.16.8-arm64-darwin)
       racc (~> 1.4)
-    nokogiri (1.16.7-x86_64-linux)
+    nokogiri (1.16.8-x86_64-linux)
       racc (~> 1.4)
     octokit (4.25.1)
       faraday (>= 1, < 3)
       sawyer (~> 0.9)
     os (1.1.4)
     parallel (1.26.3)
-    parser (3.3.5.0)
+    parser (3.3.6.0)
       ast (~> 2.4.1)
       racc
-    pg (1.5.8)
+    pg (1.5.9)
     pgvector (0.2.2)
     protocol-hpack (1.4.3)
     protocol-http (0.26.8)
@@ -158,7 +160,7 @@ GEM
     rake (13.2.1)
     rdoc (6.7.0)
       psych (>= 4.0.0)
-    regexp_parser (2.9.2)
+    regexp_parser (2.9.3)
     reline (0.5.9)
       io-console (~> 0.5)
     rest-client (2.1.0)
@@ -180,17 +182,17 @@ GEM
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.13.0)
     rspec-support (3.13.1)
-    rubocop (1.67.0)
+    rubocop (1.69.1)
       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)
-      rubocop-ast (>= 1.32.2, < 2.0)
+      regexp_parser (>= 2.9.3, < 3.0)
+      rubocop-ast (>= 1.36.2, < 2.0)
       ruby-progressbar (~> 1.7)
-      unicode-display_width (>= 2.4.0, < 3.0)
-    rubocop-ast (1.33.0)
+      unicode-display_width (>= 2.4.0, < 4.0)
+    rubocop-ast (1.36.2)
       parser (>= 3.3.1.0)
     rubocop-rake (0.6.0)
       rubocop (~> 1.0)
@@ -204,11 +206,12 @@ GEM
     sawyer (0.9.2)
       addressable (>= 2.3.5)
       faraday (>= 0.17.3, < 3)
+    securerandom (0.4.0)
     sqlite3 (2.0.4-arm64-darwin)
     sqlite3 (2.0.4-x86_64-linux-gnu)
     stringio (3.1.1)
     strings-ansi (0.2.0)
-    timeout (0.4.1)
+    timeout (0.4.2)
     timers (4.3.5)
     traces (0.11.1)
     tty-cursor (0.7.1)
@@ -221,7 +224,7 @@ GEM
     tzinfo (2.0.6)
       concurrent-ruby (~> 1.0)
     unicode-display_width (2.6.0)
-    uri (0.13.1)
+    uri (1.0.2)
     vcr (6.3.1)
       base64
     webmock (3.24.0)

data/lib/boxcars/engine/anthropic.rb

@@ -32,8 +32,8 @@ module Boxcars
       super(description: description, name: name)
     end
 
-    def conversation_model?(model)
-      @conversation_model ||= (extract_model_version(model) > 3.49)
+    def conversation_model?(_model)
+      true
     end
 
     def anthropic_client(anthropic_api_key: nil)
@@ -45,34 +45,27 @@ module Boxcars
     # @param anthropic_api_key [String] Optional api key to use when asking the engine.
     #   Defaults to Boxcars.configuration.anthropic_api_key.
     # @param kwargs [Hash] Additional parameters to pass to the engine if wanted.
-    # rubocop:disable Metrics/PerceivedComplexity
     def client(prompt:, inputs: {}, **kwargs)
       model_params = llm_params.merge(kwargs)
       api_key = Boxcars.configuration.anthropic_api_key(**kwargs)
       aclient = anthropic_client(anthropic_api_key: api_key)
       prompt = prompt.first if prompt.is_a?(Array)
-
-      if conversation_model?(model_params[:model])
-        params = convert_to_anthropic(prompt.as_messages(inputs).merge(model_params))
-        if Boxcars.configuration.log_prompts
-          if params[:messages].length < 2 && params[:system].present?
-            Boxcars.debug(">>>>>> Role: system <<<<<<\n#{params[:system]}")
-          end
-          Boxcars.debug(params[:messages].last(2).map { |p| ">>>>>> Role: #{p[:role]} <<<<<<\n#{p[:content]}" }.join("\n"), :cyan)
+      params = convert_to_anthropic(prompt.as_messages(inputs).merge(model_params))
+      if Boxcars.configuration.log_prompts
+        if params[:messages].length < 2 && params[:system].present?
+          Boxcars.debug(">>>>>> Role: system <<<<<<\n#{params[:system]}")
         end
-        response = aclient.messages(parameters: params)
-        response['completion'] = response.dig('content', 0, 'text')
-        response.delete('content')
-        response
-      else
-        params = prompt.as_prompt(inputs: inputs, prefixes: default_prefixes, show_roles: true).merge(model_params)
-        params[:prompt] = "\n\n#{params[:prompt]}" unless params[:prompt].start_with?("\n\n")
-        params[:stop_sequences] = params.delete(:stop) if params.key?(:stop)
-        Boxcars.debug("Prompt after formatting:#{params[:prompt]}", :cyan) if Boxcars.configuration.log_prompts
-        aclient.complete(parameters: params)
+        Boxcars.debug(params[:messages].last(2).map { |p| ">>>>>> Role: #{p[:role]} <<<<<<\n#{p[:content]}" }.join("\n"), :cyan)
       end
+      response = aclient.messages(parameters: params)
+      response['completion'] = response.dig('content', 0, 'text')
+      response.delete('content')
+      response
+    rescue StandardError => e
+      err = e.respond_to?(:response) ? e.response[:body] : e
+      Boxcars.warn("Anthropic Error #{e.class.name}: #{err}", :red)
+      raise
     end
-    # rubocop:enable Metrics/PerceivedComplexity
 
     # get an answer from the engine for a question.
     # @param question [String] The question to ask the engine.

data/lib/boxcars/engine/cohere.rb

@@ -50,9 +50,6 @@ module Boxcars
 
       # Make the API call
       response = connection.post { |req| req.body = params.to_json }
-
-      # response_data = JSON.parse(response.body, symbolize_names: true)
-      # response_data[:text]
       JSON.parse(response.body, symbolize_names: true)
     end
 

data/lib/boxcars/engine/gemini_ai.rb

@@ -0,0 +1,160 @@
+# frozen_string_literal: true
+
+# Boxcars - a framework for running a series of tools to get an answer to a question.
+module Boxcars
+  # A engine that uses Gemini's API.
+  class GeminiAi < Engine
+    attr_reader :prompts, :llm_params, :model_kwargs, :batch_size
+
+    # The default parameters to use when asking the engine.
+    DEFAULT_PARAMS = {
+      model: "gemini-1.5-flash-latest"
+    }.freeze
+
+    # the default name of the engine
+    DEFAULT_NAME = "Google Gemini AI engine"
+    # the default description of the engine
+    DEFAULT_DESCRIPTION = "useful for when you need to use Google Gemini AI to answer questions. " \
+                          "You should ask targeted questions"
+
+    # A engine is the driver for a single tool to run.
+    # @param name [String] The name of the engine. Defaults to "OpenAI engine".
+    # @param description [String] A description of the engine. Defaults to:
+    #        useful for when you need to use AI to answer questions. You should ask targeted questions".
+    # @param prompts [Array<String>] The prompts to use when asking the engine. Defaults to [].
+    def initialize(name: DEFAULT_NAME, description: DEFAULT_DESCRIPTION, prompts: [], **kwargs)
+      @llm_params = DEFAULT_PARAMS.merge(kwargs)
+      @prompts = prompts
+      @batch_size = 20
+      super(description: description, name: name)
+    end
+
+    def conversation_model?(_model)
+      true
+    end
+
+    def chat(params, gemini_api_key)
+      raise Boxcars::ConfigurationError('Google AI API key not set') if gemini_api_key.blank?
+
+      model_string = params.delete(:model_string)
+      raise Boxcars::ConfigurationError('Google AI API key not set') if model_string.blank?
+
+      # Define the API endpoint and parameters
+      api_endpoint = "https://generativelanguage.googleapis.com/v1beta/models/#{model_string}:generateContent?key=#{gemini_api_key}"
+
+      connection = Faraday.new(api_endpoint) do |faraday|
+        faraday.request :url_encoded
+        faraday.headers['Content-Type'] = 'application/json'
+      end
+
+      # Make the API call
+      response = connection.post { |req| req.body = params.to_json }
+
+      JSON.parse(response.body, symbolize_names: true)
+    end
+
+    # Get an answer from the engine.
+    # @param prompt [String] The prompt to use when asking the engine.
+    # @param gemini_api_key [String] Optional api key to use when asking the engine.
+    #   Defaults to Boxcars.configuration.gemini_api_key.
+    # @param kwargs [Hash] Additional parameters to pass to the engine if wanted.
+    def client(prompt:, inputs: {}, **kwargs)
+      api_key = Boxcars.configuration.gemini_api_key(**kwargs)
+      option_params = llm_params.merge(kwargs)
+      model_string = option_params.delete(:model) || DEFAULT_PARAMS[:model]
+      convo = prompt.as_messages(inputs: inputs)
+      # Convert conversation to Google Gemini format
+      params = to_google_gemini_format(convo[:messages], option_params)
+      params[:model_string] = model_string
+      Boxcars.debug("Prompt after formatting:#{params[:message]}", :cyan) if Boxcars.configuration.log_prompts
+      chat(params, api_key)
+    end
+
+    # get an answer from the engine for a question.
+    # @param question [String] The question to ask the engine.
+    # @param kwargs [Hash] Additional parameters to pass to the engine if wanted.
+    def run(question, **kwargs)
+      prompt = Prompt.new(template: question)
+      response = client(prompt: prompt, **kwargs)
+
+      raise Error, "GeminiAI: No response from API" unless response
+      raise Error, "GeminiAI: #{response[:error]}" if response[:error]
+
+      answer = response[:candidates].first[:content][:parts].first[:text]
+      Boxcars.debug(response, :yellow)
+      answer
+    end
+
+    # Get the default parameters for the engine.
+    def default_params
+      llm_params
+    end
+
+    # make sure we got a valid response
+    # @param response [Hash] The response to check.
+    # @param must_haves [Array<String>] The keys that must be in the response. Defaults to %w[choices].
+    # @raise [KeyError] if there is an issue with the access token.
+    # @raise [ValueError] if the response is not valid.
+    def check_response(response, must_haves: %w[completion])
+      if response['error']
+        code = response.dig('error', 'code')
+        msg = response.dig('error', 'message') || 'unknown error'
+        raise KeyError, "ANTHOPIC_API_KEY not valid" if code == 'invalid_api_key'
+
+        raise ValueError, "Gemini error: #{msg}"
+      end
+
+      must_haves.each do |key|
+        raise ValueError, "Expecting key #{key} in response" unless response.key?(key)
+      end
+    end
+
+    # the engine type
+    def engine_type
+      "claude"
+    end
+
+    # lookup the context size for a model by name
+    # @param modelname [String] The name of the model to lookup.
+    def modelname_to_contextsize(_modelname)
+      100000
+    end
+
+    # Calculate the maximum number of tokens possible to generate for a prompt.
+    # @param prompt_text [String] The prompt text to use.
+    # @return [Integer] the number of tokens possible to generate.
+    def max_tokens_for_prompt(prompt_text)
+      num_tokens = get_num_tokens(prompt_text)
+
+      # get max context size for model by name
+      max_size = modelname_to_contextsize(model_name)
+      max_size - num_tokens
+    end
+
+    def to_google_gemini_format(convo, option_params)
+      instructions = convo.shift.last if convo.first && convo.first[:role] == :system
+      system_instructions = instructions || "You are a helpful assistant."
+
+      # Convert conversation history to the format expected by Google
+      contents = convo.map { |message| { text: message[:content] } }
+
+      generation_config = {}
+      if option_params.length.positive?
+        generation_config.merge!(option_params)
+        generation_config[:stopSequences] = [generation_config.delete(:stop)] if generation_config[:stop].present?
+      end
+
+      rv = {
+        system_instruction: { parts: { text: system_instructions } }, # System instructions or context
+        contents: { parts: contents } # The chat messages
+      }
+
+      rv[:generationConfig] = generation_config if generation_config.length.positive?
+      rv
+    end
+
+    def default_prefixes
+      { system: "SYSTEM: ", user: "USER: ", assistant: "CHATBOT: ", history: :history }
+    end
+  end
+end

data/lib/boxcars/engine/openai.rb

@@ -45,7 +45,7 @@ module Boxcars
     def self.open_ai_client(openai_access_token: nil)
       access_token = Boxcars.configuration.openai_access_token(openai_access_token: openai_access_token)
       organization_id = Boxcars.configuration.organization_id
-      ::OpenAI::Client.new(access_token: access_token, organization_id: organization_id)
+      ::OpenAI::Client.new(access_token: access_token, organization_id: organization_id, log_errors: true)
     end
 
     def conversation_model?(model)
@@ -76,6 +76,10 @@ module Boxcars
         Boxcars.debug("Prompt after formatting:\n#{params[:prompt]}", :cyan) if Boxcars.configuration.log_prompts
         clnt.completions(parameters: params)
       end
+    rescue StandardError => e
+      err = e.respond_to?(:response) ? e.response[:body] : e
+      Boxcars.warn("OpenAI Error #{e.class.name}: #{err}", :red)
+      raise
     end
 
     # get an answer from the engine for a question.

data/lib/boxcars/engine.rb

@@ -79,3 +79,4 @@ require "boxcars/engine/ollama"
 require "boxcars/engine/openai"
 require "boxcars/engine/perplexityai"
 require "boxcars/engine/gpt4all_eng"
+require "boxcars/engine/gemini_ai"

data/lib/boxcars/version.rb

@@ -2,5 +2,5 @@
 
 module Boxcars
   # The current version of the gem.
-  VERSION = "0.6.6"
+  VERSION = "0.6.8"
 end

data/lib/boxcars/x_node.rb

@@ -54,7 +54,9 @@ module Boxcars
     end
 
     def xtext(path)
+      # rubocop:disable Style/SafeNavigationChainLength
       rv = xpath(path)&.text&.gsub(/[[:space:]]+/, " ")&.strip
+      # rubocop:enable Style/SafeNavigationChainLength
       return nil if rv.empty?
 
       rv

data/lib/boxcars.rb

@@ -62,6 +62,11 @@ module Boxcars
       key_lookup(:groq_api_key, kwargs)
     end
 
+    # @return [String] The Google AI API key either from arg or env.
+    def gemini_api_key(**kwargs)
+      key_lookup(:gemini_api_key, kwargs)
+    end
+
     private
 
     def check_key(key, val)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: boxcars
 version: !ruby/object:Gem::Version
-  version: 0.6.6
+  version: 0.6.8
 platform: ruby
 authors:
 - Francis Sullivan
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-11-15 00:00:00.000000000 Z
+date: 2024-12-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: anthropic
@@ -159,6 +159,7 @@ files:
 - lib/boxcars/engine/anthropic.rb
 - lib/boxcars/engine/cohere.rb
 - lib/boxcars/engine/engine_result.rb
+- lib/boxcars/engine/gemini_ai.rb
 - lib/boxcars/engine/gpt4all_eng.rb
 - lib/boxcars/engine/groq.rb
 - lib/boxcars/engine/ollama.rb