boxcars 0.6.1 → 0.6.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 71b72b7d30bda72b5dd0b3150773a233eac72e15ab3efa66d7643d5278770793
-  data.tar.gz: '080addca8e8c89f982db5ae91d9f12e67db15fb37f58d4787857b209848c1006'
+  metadata.gz: 8c7137b49fcea80018efc8f6dc19361fa02b74a575f7e71218784986b366c350
+  data.tar.gz: 9b2bba4216ab4f50024d158cc7b18977569fa7b9bada1a4734fb24f7b9e4a3be
 SHA512:
-  metadata.gz: 330aa8094b7069084a7e06c7d656b30bd1935f1e12cfefded4788f23d92c62dbbbb92458328e37b95cbf5e446afd73b9170916329f409490e532ae329ac20a79
-  data.tar.gz: 1730d2356b2ddae5e19ee6d0ec0b4b6150eb9a9efb2305863b8e946ccd0184124a3ab22ef6fa37c4cfdacebed7a8fbf520f2d18243f9d05de7f547764d3625ef
+  metadata.gz: 18b58fc68b3837e2f8c7ca1e77e1daf49655843903d39a355233ff1d8c7a68caa885530ba08bcfc3842189b070c547ce152e13ee1a1b5cdb4cdc0424fe8e8ddc
+  data.tar.gz: e6240effad988b8cb819070ea37f54eab03dd99b32eff0caa07aa167e0f854492c12843d781e9614f429ffd3a311f426bc58c9569af22f7424bd9b65f60244fa

data/.ruby-version

@@ -1 +1 @@
-3.2.2
+3.3.3

data/CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## [v0.6.1](https://github.com/BoxcarsAI/boxcars/tree/v0.6.1) (2024-07-19)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.5.1...v0.6.1)
+
+**Merged pull requests:**
+
+- various updates with Claude 3.5 support [\#197](https://github.com/BoxcarsAI/boxcars/pull/197) ([francis](https://github.com/francis))
+- \[infra\] Bump rubocop-rspec from 2.30.0 to 3.0.2 [\#195](https://github.com/BoxcarsAI/boxcars/pull/195) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump nokogiri from 1.16.5 to 1.16.6 [\#194](https://github.com/BoxcarsAI/boxcars/pull/194) ([dependabot[bot]](https://github.com/apps/dependabot))
+- \[infra\] Bump ruby-openai from 7.0.1 to 7.1.0 [\#193](https://github.com/BoxcarsAI/boxcars/pull/193) ([dependabot[bot]](https://github.com/apps/dependabot))
+
 ## [v0.5.1](https://github.com/BoxcarsAI/boxcars/tree/v0.5.1) (2024-06-14)
 
 [Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.4.10...v0.5.1)

data/Gemfile.lock

@@ -1,14 +1,14 @@
 PATH
   remote: .
   specs:
-    boxcars (0.6.1)
+    boxcars (0.6.3)
       anthropic (~> 0.1)
       google_search_results (~> 2.2)
       gpt4all (~> 0.0.4)
       hnswlib (~> 0.8)
       nokogiri (~> 1.16)
       pgvector (~> 0.2)
-      ruby-openai (>= 4.2, < 8.0)
+      ruby-openai (>= 7.1, < 8.0)
 
 GEM
   remote: https://rubygems.org/
@@ -81,7 +81,7 @@ GEM
       faraday (>= 0.8)
     faraday-multipart (1.0.4)
       multipart-post (~> 2)
-    faraday-net_http (3.1.0)
+    faraday-net_http (3.1.1)
       net-http
     faraday-retry (2.2.1)
       faraday (~> 2.0)

data/boxcars.gemspec

@@ -37,7 +37,7 @@ Gem::Specification.new do |spec|
   spec.add_dependency "hnswlib", "~> 0.8"
   spec.add_dependency "nokogiri", "~> 1.16"
   spec.add_dependency "pgvector", "~> 0.2"
-  spec.add_dependency "ruby-openai", ">= 4.2", "< 8.0"
+  spec.add_dependency "ruby-openai", ">= 7.1", "< 8.0"
 
   # For more information and examples about making a new gem, checkout our
   # guide at: https://bundler.io/guides/creating_gem.html

data/lib/boxcars/boxcar/json_engine_boxcar.rb

@@ -5,20 +5,21 @@ module Boxcars
   # For Boxcars that use an engine to do their work.
   class JSONEngineBoxcar < EngineBoxcar
     # A JSON Engine Boxcar is a container for a single tool to run.
-    attr_accessor :wanted_data, :data_description, :important
+    attr_accessor :wanted_data, :data_description, :important, :symbolize
 
     # @param prompt [Boxcars::Prompt] The prompt to use for this boxcar with sane defaults.
     # @param wanted_data [String] The data to extract from.
     # @param data_description [String] The description of the data.
     # @param important [String] Any important instructions you want to give the LLM.
+    # @param symbolize [Boolean] Symbolize the JSON results if true
     # @param kwargs [Hash] Additional arguments
-    def initialize(prompt: nil, wanted_data: nil, data_description: nil, important: nil, **kwargs)
+    def initialize(prompt: nil, wanted_data: nil, data_description: nil, important: nil, symbolize: false, **kwargs)
       @wanted_data = wanted_data || "summarize the pertinent facts from the input data"
       @data_description = data_description || "the input data"
       @important = important
       the_prompt = prompt || default_prompt
-      the_prompt.append("\n\nImportant: #{important}\n") if important.present?
       kwargs[:description] ||= "JSON Engine Boxcar"
+      @symbolize = symbolize
       super(prompt: the_prompt, **kwargs)
     end
 
@@ -34,6 +35,8 @@ module Boxcars
             %<wanted_data>s
           }
       SYSPR
+      stock_prompt += "\n\nImportant:\n#{important}\n" if important.present?
+
       sprompt = format(stock_prompt, wanted_data: wanted_data, data_description: data_description)
       ctemplate = [
         Boxcar.syst(sprompt),
@@ -50,7 +53,7 @@ module Boxcars
       # 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]))
+      extract_answer(JSON.parse(engine_output[json_start..json_end], symbolize_names: symbolize))
     rescue StandardError => e
       Result.from_error("Error: #{e.message}:\n#{engine_output}")
     end

data/lib/boxcars/engine/groq.rb

@@ -0,0 +1,174 @@
+# frozen_string_literal: true
+
+# Boxcars is a framework for running a series of tools to get an answer to a question.
+module Boxcars
+  # A engine that uses Groq's API.
+  class Groq < Engine
+    attr_reader :prompts, :groq_parmas, :model_kwargs, :batch_size
+
+    # The default parameters to use when asking the engine.
+    DEFAULT_PARAMS = {
+      model: "llama3-70b-8192",
+      temperature: 0.1,
+      max_tokens: 4096
+    }.freeze
+
+    # the default name of the engine
+    DEFAULT_NAME = "Groq engine"
+    # the default description of the engine
+    DEFAULT_DESCRIPTION = "useful for when you need to use AI to answer questions. " \
+                          "You should ask targeted questions"
+
+    # A engine is a container for a single tool to run.
+    # @param name [String] The name of the engine. Defaults to "Groq 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 [].
+    # @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)
+      @groq_parmas = DEFAULT_PARAMS.merge(kwargs)
+      @prompts = prompts
+      @batch_size = batch_size
+      super(description: description, name: name)
+    end
+
+    # Get the OpenAI API client
+    # @param groq_api_key [String] The access token to use when asking the engine.
+    #   Defaults to Boxcars.configuration.groq_api_key
+    # @return [OpenAI::Client] The OpenAI API gem client.
+    def self.open_ai_client(groq_api_key: nil)
+      access_token = Boxcars.configuration.groq_api_key(groq_api_key: groq_api_key)
+      ::OpenAI::Client.new(access_token: access_token, uri_base: "https://api.groq.com/openai")
+    end
+
+    def conversation_model?(_model)
+      true
+    end
+
+    # Get an answer from the engine.
+    # @param prompt [String] The prompt to use when asking the engine.
+    # @param groq_api_key [String] The access token to use when asking the engine.
+    #   Defaults to Boxcars.configuration.groq_api_key.
+    # @param kwargs [Hash] Additional parameters to pass to the engine if wanted.
+    def client(prompt:, inputs: {}, groq_api_key: nil, **kwargs)
+      clnt = Groq.open_ai_client(groq_api_key: groq_api_key)
+      params = groq_parmas.merge(kwargs)
+      prompt = prompt.first if prompt.is_a?(Array)
+      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)
+      end
+      clnt.chat(parameters: params)
+    rescue => e
+      Boxcars.error(e, :red)
+      raise
+    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, "Groq: No response from API" unless response
+      raise Error, "Groq: #{response['error']}" if response["error"]
+
+      answer = response["choices"].map { |c| c.dig("message", "content") || c["text"] }.join("\n").strip
+      puts answer
+      answer
+    end
+
+    # Get the default parameters for the engine.
+    def default_params
+      groq_parmas
+    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
+
+    # 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[choices])
+      if response['error']
+        code = response.dig('error', 'code')
+        msg = response.dig('error', 'message') || 'unknown error'
+        raise KeyError, "OPENAI_ACCESS_TOKEN not valid" if code == 'invalid_api_key'
+
+        raise ValueError, "Groq error: #{msg}"
+      end
+
+      must_haves.each do |key|
+        raise ValueError, "Expecting key #{key} in response" unless response.key?(key)
+      end
+    end
+
+    # Call out to Groq's 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
+    # rubocop:enable Metrics/AbcSize
+  end
+
+  # the engine type
+  def engine_type
+    "groq"
+  end
+
+  # calculate the number of tokens used
+  def get_num_tokens(text:)
+    text.split.length # TODO: hook up to token counting gem
+  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 = 8096
+    max_size - num_tokens
+  end
+end

data/lib/boxcars/engine.rb

@@ -22,6 +22,7 @@ end
 require "boxcars/engine/engine_result"
 require "boxcars/engine/anthropic"
 require "boxcars/engine/cohere"
+require "boxcars/engine/groq"
 require "boxcars/engine/openai"
 require "boxcars/engine/perplexityai"
 require "boxcars/engine/gpt4all_eng"

data/lib/boxcars/version.rb

@@ -2,5 +2,5 @@
 
 module Boxcars
   # The current version of the gem.
-  VERSION = "0.6.1"
+  VERSION = "0.6.3"
 end

data/lib/boxcars.rb

@@ -27,7 +27,7 @@ module Boxcars
 
   # Configuration contains gem settings
   class Configuration
-    attr_writer :openai_access_token, :serpapi_api_key
+    attr_writer :openai_access_token, :serpapi_api_key, :groq_api_key
     attr_accessor :organization_id, :logger, :log_prompts, :log_generated, :default_train, :default_engine
 
     def initialize
@@ -52,11 +52,16 @@ module Boxcars
       key_lookup(:anthropic_api_key, kwargs)
     end
 
-    # @return [String] The Anthropic API key either from arg or env.
+    # @return [String] The Cohere API key either from arg or env.
     def cohere_api_key(**kwargs)
       key_lookup(:cohere_api_key, kwargs)
     end
 
+    # @return [String] The Groq API key either from arg or env.
+    def groq_api_key(**kwargs)
+      key_lookup(:groq_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: 0.6.1
+  version: 0.6.3
 platform: ruby
 authors:
 - Francis Sullivan
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-07-19 00:00:00.000000000 Z
+date: 2024-07-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: anthropic
@@ -101,7 +101,7 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '4.2'
+        version: '7.1'
     - - "<"
       - !ruby/object:Gem::Version
         version: '8.0'
@@ -111,7 +111,7 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '4.2'
+        version: '7.1'
     - - "<"
       - !ruby/object:Gem::Version
         version: '8.0'
@@ -160,6 +160,7 @@ files:
 - lib/boxcars/engine/cohere.rb
 - lib/boxcars/engine/engine_result.rb
 - lib/boxcars/engine/gpt4all_eng.rb
+- lib/boxcars/engine/groq.rb
 - lib/boxcars/engine/openai.rb
 - lib/boxcars/engine/perplexityai.rb
 - lib/boxcars/generation.rb