boxcars 0.1.4 → 0.1.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 0fc42578d07a962cd91fbab547ba657842c131729070e57bcf7789aa1648aaee
-  data.tar.gz: 3e3009663c45d58ee616d2d705c74f939b12922b24bceff74d231f5aeaff573d
+  metadata.gz: 666f9b91451e8b6db61eb1162cb6508a566a46b676d646f1fb0000175714cede
+  data.tar.gz: bee7b55be2288bf9361012bd562f88d91f64ce7972b0cf5c495a5fce87eea77c
 SHA512:
-  metadata.gz: d5a2664fb2b4973b57e41f682a9b194394bc196ada00b9c8cf9106a7876b92065cd9f40ff270650219c282143f812241e16fc450a24f61567acaaf413f44150f
-  data.tar.gz: 1d73a0c1519a3834852ffe7108e43e03460cc6e1b99dbcc9c16bf73c7824016d71628878b32394a33b548a0c84de7f9a41feefa1dc922628957eb0251269cd7a
+  metadata.gz: '0681f5bd00bbfe2afd67011d76c561640a0575a9656b20913d3c60193208ea33361003ee652d01802b49702b106bd4f3d5bc91c1c7f2023bc2e8224edca14a8d'
+  data.tar.gz: 794fe20c2f9b8f14402db177bdee3d3e5318ab652e69c02929f9ae79d6daca096ae386a4e85730da4004927ab31aff53cc3a8d960ca661d6a5dd192432ffeabb

data/CHANGELOG.md

@@ -1,5 +1,45 @@
 # Changelog
 
+## [Unreleased](https://github.com/BoxcarsAI/boxcars/tree/HEAD)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.1.5...HEAD)
+
+**Implemented enhancements:**
+
+- Add a callback function for Boxcars::ActiveRecord to approve changes  [\#24](https://github.com/BoxcarsAI/boxcars/issues/24)
+
+**Merged pull requests:**
+
+- Add approval callback function for Boxcars::ActiveRecord for changes to the data [\#25](https://github.com/BoxcarsAI/boxcars/pull/25) ([francis](https://github.com/francis))
+- \[fix\] Fixed specs which required a key [\#23](https://github.com/BoxcarsAI/boxcars/pull/23) ([AKovtunov](https://github.com/AKovtunov))
+
+## [v0.1.5](https://github.com/BoxcarsAI/boxcars/tree/v0.1.5) (2023-02-22)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.1.4...v0.1.5)
+
+**Implemented enhancements:**
+
+- Make Boxcars::ActiveRecord read\_only by default [\#20](https://github.com/BoxcarsAI/boxcars/issues/20)
+
+**Merged pull requests:**
+
+- Active Record readonly [\#21](https://github.com/BoxcarsAI/boxcars/pull/21) ([francis](https://github.com/francis))
+
+## [v0.1.4](https://github.com/BoxcarsAI/boxcars/tree/v0.1.4) (2023-02-22)
+
+[Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.1.3...v0.1.4)
+
+**Implemented enhancements:**
+
+- Extend Sql concept to produce and run ActiveRecord code instead of SQL [\#9](https://github.com/BoxcarsAI/boxcars/issues/9)
+
+**Merged pull requests:**
+
+- first pass at an ActiveRecord boxcar [\#18](https://github.com/BoxcarsAI/boxcars/pull/18) ([francis](https://github.com/francis))
+- change Boxcars::default\_train to Boxcars::train to improve code reada… [\#17](https://github.com/BoxcarsAI/boxcars/pull/17) ([francis](https://github.com/francis))
+- rename class Serp to GoogleSearch [\#16](https://github.com/BoxcarsAI/boxcars/pull/16) ([francis](https://github.com/francis))
+- Update README.md [\#15](https://github.com/BoxcarsAI/boxcars/pull/15) ([tabrez-syed](https://github.com/tabrez-syed))
+
 ## [v0.1.3](https://github.com/BoxcarsAI/boxcars/tree/v0.1.3) (2023-02-17)
 
 [Full Changelog](https://github.com/BoxcarsAI/boxcars/compare/v0.1.2...v0.1.3)

data/Gemfile

@@ -24,3 +24,5 @@ gem "sqlite3", "~> 1.6"
 gem "activerecord", "~> 7.0"
 
 gem "github_changelog_generator", "~> 1.16"
+
+gem "faraday-retry", "~> 2.0"

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    boxcars (0.1.4)
+    boxcars (0.1.6)
       google_search_results (~> 2.2)
       ruby-openai (~> 3.0)
 
@@ -56,6 +56,8 @@ GEM
     faraday-http-cache (2.4.1)
       faraday (>= 0.8)
     faraday-net_http (3.0.2)
+    faraday-retry (2.0.0)
+      faraday (~> 2.0)
     fiber-local (1.0.0)
     github_changelog_generator (1.16.4)
       activesupport
@@ -171,6 +173,7 @@ DEPENDENCIES
   boxcars!
   debug (~> 1.1)
   dotenv (~> 2.8)
+  faraday-retry (~> 2.0)
   github_changelog_generator (~> 1.16)
   rake (~> 13.0)
   rspec (~> 3.2)

data/README.md

@@ -4,16 +4,16 @@
   <a href="https://www.boxcars.ai">Website</a> |
   <a href="https://www.boxcars.ai/roadmap">Roadmap</a> |
   <a href="https://www.boxcars.ai/blog">Blog</a> |
-  <a href="https://www.boxcars.ai/en/introduction/">Documentation</a> 
+  <a href="https://github.com/BoxcarsAI/boxcars/wiki">Documentation</a> 
 </h4>
 
 <p align="center">
   <a href="https://github.com/BoxcarsAI/boxcars/blob/main/LICENSE.txt"><img src="https://img.shields.io/badge/license-MIT-informational" alt="License"></a>
 </p>
 
-Boxcars is a gem that enables you to create new systems with AI composability, using various concepts such as OpenAI, Search, SQL, Rails Active Record and more (including your concepts).
+Boxcars is a gem that enables you to create new systems with AI composability, using various concepts such as OpenAI, Search, SQL, Rails Active Record and more. This can even be extended with your concepts as well.(including your concepts).
 
-This work was inspired by the popular Python library Langchain. However, we wanted to give it a Ruby spin and make it more user-friendly for beginners to get started.
+This gem was inspired by the popular Python library Langchain. However, we wanted to give it a Ruby spin and make it more user-friendly for beginners to get started.
 
 ## Concepts
 All of these concepts are in a module named Boxcars:
@@ -79,10 +79,11 @@ Note that since Openai is currently the most used Engine, if you do not pass in
 calc = Boxcars::Calculator.new # just use the default Engine
 puts calc.run "what is pi to the forth power divided by 22.1?"
 ```
+You can change the default_engine with `Boxcars::configuration.default_engine = NewDefaultEngine`
 ### Boxcars currently implemmented
 
 Here is what we have so far, but please put up a PR with your new ideas.
-- Search: uses the SERP API to do seaches
+- GoogleSearch: uses the SERP API to do seaches
 - Calculator: uses an Engine to generate ruby code to do math
 - SQL: given an ActiveRecord connection, it will generate and run sql statments from a prompt.
 

data/boxcars.gemspec

@@ -8,8 +8,8 @@ Gem::Specification.new do |spec|
   spec.authors = ["Francis Sullivan", "Tabrez Syed"]
   spec.email = ["hi@boxcars.ai"]
 
-  spec.summary = "Boxcars provide an API to connect together Boxcars and then conduct them. Inspired by python langchain."
-  spec.description = "You simply give a number of boxcars to a train, and it does the magic."
+  spec.summary = "Boxcars is a gem that enables you to create new systems with AI composability. Inspired by python langchain."
+  spec.description = "You simply set an OpenAI key, give a number of Boxcars to a Train, and magic ensues when you run it."
   spec.homepage = "https://github.com/BoxcarsAI/boxcars"
   spec.license = "MIT"
   spec.required_ruby_version = ">= 2.6.0"

data/lib/boxcars/boxcar/active_record.rb

@@ -7,26 +7,22 @@ module Boxcars
     # the description of this engine boxcar
     ARDESC = "useful for when you need to query a database for an application named %<name>s."
     LOCKED_OUT_MODELS = %w[ActiveRecord::SchemaMigration ActiveRecord::InternalMetadata ApplicationRecord].freeze
-    attr_accessor :connection, :input_key, :requested_models
+    attr_accessor :connection, :input_key, :requested_models, :read_only, :approval_callback
     attr_reader :except_models
 
     # @param engine [Boxcars::Engine] The engine to user for this boxcar. Can be inherited from a train if nil.
     # @param models [Array<ActiveRecord::Model>] The models to use for this boxcar. Will use all if nil.
-    # @param input_key [Symbol] The key to use for the input. Defaults to :question.
-    # @param output_key [Symbol] The key to use for the output. Defaults to :answer.
-    # @param kwargs [Hash] Any other keyword arguments to pass to the parent class. This can include
-    #   :name, :description and :prompt
-    def initialize(engine: nil, models: nil, input_key: :question, output_key: :answer, **kwargs)
-      if models.is_a?(Array) && models.length.positive?
-        @requested_models = models
-        models.each do |m|
-          raise ArgumentError, "model #{m} needs to be an Active Record model" unless m.ancestors.include?(::ActiveRecord::Base)
-        end
-      elsif models
-        raise ArgumentError, "models needs to be an array of Active Record models"
-      end
+    # @param read_only [Boolean] Whether to use read only models. Defaults to true unless you pass an approval function.
+    # @param approval_callback [Proc] A function to call to approve changes. Defaults to nil.
+    # @param kwargs [Hash] Any other keyword arguments. These can include:
+    #   :name, :description, :prompt, :input_key, :output_key and :except_models
+    def initialize(engine: nil, models: nil, read_only: nil, approval_callback: nil, **kwargs)
+      check_models(models)
       @except_models = LOCKED_OUT_MODELS + kwargs[:except_models].to_a
-      @input_key = input_key
+      @approval_callback = approval_callback
+      @read_only = read_only.nil? ? !approval_callback : read_only
+      @input_key = kwargs[:input_key] || :question
+      @output_key = kwargs[:output_key] || :answer
       the_prompt = kwargs[prompt] || my_prompt
       name = kwargs[:name] || "Data"
       super(name: name,
@@ -60,6 +56,21 @@ module Boxcars
 
     private
 
+    def read_only?
+      read_only
+    end
+
+    def check_models(models)
+      if models.is_a?(Array) && models.length.positive?
+        @requested_models = models
+        models.each do |m|
+          raise ArgumentError, "model #{m} needs to be an Active Record model" unless m.ancestors.include?(::ActiveRecord::Base)
+        end
+      elsif models
+        raise ArgumentError, "models needs to be an array of Active Record models"
+      end
+    end
+
     def wanted_models
       the_models = requested_models || ::ActiveRecord::Base.descendants
       the_models.reject { |m| except_models.include?(m.name) }
@@ -75,21 +86,94 @@ module Boxcars
       models.pretty_inspect
     end
 
-    def get_active_record_answer(text)
-      code = text[/^ARCode: (.*)/, 1]
+    # to be safe, we wrap the code in a transaction and rollback
+    def rollback_after_running
+      rv = nil
+      ::ActiveRecord::Base.transaction do
+        rv = yield
+      ensure
+        raise ::ActiveRecord::Rollback
+      end
+      rv
+    end
+
+    # check for dangerous code that is outside of ActiveRecord
+    def safe_to_run?(code)
+      bad_words = %w[commit drop_constraint drop_constraint! drop_extension drop_extension! drop_foreign_key drop_foreign_key! \
+                     drop_index drop_index! drop_join_table drop_join_table! drop_materialized_view drop_materialized_view! \
+                     drop_partition drop_partition! drop_schema drop_schema! drop_table drop_table! drop_trigger drop_trigger! \
+                     drop_view drop_view! eval execute reset revoke rollback truncate].freeze
+      without_strings = code.gsub(/('([^'\\]*(\\.[^'\\]*)*)'|"([^"\\]*(\\.[^"\\]*)*"))/, 'XX')
+      word_list = without_strings.split(/[.,()]/)
+
+      bad_words.each do |w|
+        if word_list.include?(w)
+          puts "code included destructive instruction: #{w} #{code}".colorize(:red)
+          return false
+        end
+      end
+
+      true
+    end
+
+    def evaluate_input(code)
+      raise SecurityError, "Found unsafe code while evaluating: #{code}" unless safe_to_run?(code)
+
+      # rubocop:disable Security/Eval
+      eval code
+      # rubocop:enable Security/Eval
+    end
+
+    def change_count(changes_code)
+      return 0 unless changes_code
+
+      rollback_after_running do
+        puts "computing change count with: #{changes_code}".colorize(:yellow)
+        evaluate_input changes_code
+      end
+    end
+
+    def approved?(changes_code, code)
+      # find out how many changes there are
+      changes = change_count(changes_code)
+      return true unless changes&.positive?
+
+      puts "Pending Changes: #{changes}".colorize(:yellow, style: :bold)
+      change_str = "#{changes} change#{'s' if changes.to_i > 1}"
+      raise SecurityError, "Can not run code that makes #{change_str} in read-only mode" if read_only?
+
+      return approval_callback.call(changes, code) if approval_callback.is_a?(Proc)
+
+      true
+    end
+
+    def run_active_record_code(code)
       puts code.colorize(:yellow)
-      begin
-        # rubocop:disable Security/Eval
-        output = eval code
-        # rubocop:enable Security/Eval
-        output = output.first if output.is_a?(Array) && output.length == 1
-        "Answer: #{output.inspect}"
-      rescue StandardError => e
-        "Error: #{e.message}"
+      if read_only?
+        rollback_after_running do
+          evaluate_input code
+        end
+      else
+        evaluate_input code
       end
     end
 
+    def get_active_record_answer(text)
+      code = text[/^ARCode: (.*)/, 1]
+      changes_code = text[/^ARChanges: (.*)/, 1]
+      raise SecurityError, "Permission to run code that makes changes denied" unless approved?(changes_code, code)
+
+      output = run_active_record_code(code)
+      output = 0 if output.is_a?(Array) && output.empty?
+      output = output.first if output.is_a?(Array) && output.length == 1
+      output = output[output.keys.first] if output.is_a?(Hash) && output.length == 1
+      "Answer: #{output.inspect}"
+    rescue StandardError => e
+      "Error: #{e.message}"
+    end
+
     def get_answer(text)
+      # debugger
       case text
       when /^ARCode:/
         get_active_record_answer(text)
@@ -114,6 +198,7 @@ module Boxcars
       Use the following format:
       Question: "Question here"
       ARCode: "Active Record code to run"
+      ARChanges: "Active Record code to compute the number of records going to change" - Only add this line if the ARCode on the line before will make data changes
       Answer: "Final answer here"
 
       Only use the following Active Record models:

data/lib/boxcars/boxcar/google_search.rb

@@ -13,7 +13,7 @@ module Boxcars
     # @param name [String] The name of the boxcar. Defaults to classname.
     # @param description [String] A description of the boxcar. Defaults to SERPDESC.
     # @param serpapi_api_key [String] The API key to use for the SerpAPI. Defaults to Boxcars.configuration.serpapi_api_key.
-    def initialize(name: "Search", description: SERPDESC, serpapi_api_key: "not set")
+    def initialize(name: "Search", description: SERPDESC, serpapi_api_key: nil)
       super(name: name, description: description)
       api_key = Boxcars.configuration.serpapi_api_key(serpapi_api_key: serpapi_api_key)
       ::GoogleSearch.api_key = api_key

data/lib/boxcars/engine/openai.rb

@@ -38,7 +38,7 @@ module Boxcars
     # @param openai_access_token [String] The access token to use when asking the engine.
     #   Defaults to Boxcars.configuration.openai_access_token.
     # @param kwargs [Hash] Additional parameters to pass to the engine if wanted.
-    def client(prompt:, openai_access_token: 'not set', **kwargs)
+    def client(prompt:, openai_access_token: nil, **kwargs)
       access_token = Boxcars.configuration.openai_access_token(openai_access_token: openai_access_token)
       organization_id = Boxcars.configuration.organization_id
       clnt = ::OpenAI::Client.new(access_token: access_token, organization_id: organization_id)

data/lib/boxcars/ruby_repl.rb

@@ -6,14 +6,14 @@ module Boxcars
     # Execute ruby code
     # @param code [String] The code to run
     def call(code:)
-      puts "RubyREPL: #{code}".colorize(:red)
+      puts "RubyREPL: #{code}".colorize(:yellow)
       output = ""
       IO.popen("ruby", "r+") do |io|
         io.puts code
         io.close_write
         output = io.read
       end
-      puts "Answer: #{output}".colorize(:red, style: :bold)
+      puts "Answer: #{output}".colorize(:yellow, style: :bold)
       output
     end
 

data/lib/boxcars/train/zero_shot.rb

@@ -33,14 +33,15 @@ module Boxcars
     # @param engine [Boxcars::Engine] The engine to use for this train.
     # @param name [String] The name of the train. Defaults to 'Zero Shot'.
     # @param description [String] The description of the train. Defaults to 'Zero Shot Train'.
-    def initialize(boxcars:, engine: nil, name: 'Zero Shot', description: 'Zero Shot Train')
+    # @param prompt [Boxcars::Prompt] The prompt to use. Defaults to the built-in prompt.
+    def initialize(boxcars:, engine: nil, name: 'Zero Shot', description: 'Zero Shot Train', prompt: nil)
       @observation_prefix = 'Observation: '
       @engine_prefix = 'Thought:'
-      prompt = self.class.create_prompt(boxcars: boxcars)
+      prompt ||= self.class.create_prompt(boxcars: boxcars)
       super(engine: engine, boxcars: boxcars, prompt: prompt, name: name, description: description)
     end
 
-    # Create prompt in the style of the zero shot agent.
+    # Create prompt in the style of the zero shot agent. Without arguments, returns the default prompt.
     # @param boxcars [Array<Boxcars::Boxcar>] List of boxcars the agent will have access to, used to format the prompt.
     # @param prefix [String] String to put before the main prompt.
     # @param suffix [String] String to put after the main prompt.

data/lib/boxcars/train.rb

@@ -112,7 +112,7 @@ module Boxcars
     # @param intermediate_steps [Array<Hash>] The intermediate steps.
     # @return [Hash] The final output.
     def pre_return(output, intermediate_steps)
-      puts output.log.colorize(:yellow)
+      puts output.log.colorize(:yellow, style: :bold)
       final_output = output.return_values
       final_output["intermediate_steps"] = intermediate_steps if return_intermediate_steps
       final_output

data/lib/boxcars/version.rb

@@ -2,5 +2,5 @@
 
 module Boxcars
   # The current version of the gem.
-  VERSION = "0.1.4"
+  VERSION = "0.1.6"
 end

data/lib/boxcars.rb

@@ -16,6 +16,9 @@ module Boxcars
   # Error class for all Boxcars value errors.
   class ValueError < Error; end
 
+  # Error class for all Boxcars security errors.
+  class SecurityError < Error; end
+
   # simple string colorization
   class ::String
     # colorize a string
@@ -65,7 +68,7 @@ module Boxcars
     end
 
     def key_lookup(key, kwargs)
-      rv = if kwargs.key?(key) && kwargs[key] != "not set"
+      rv = if kwargs.key?(key) && !kwargs[key].nil?
              # override with kwargs if present
              kwargs[key]
            elsif (set_val = instance_variable_get("@#{key}"))
@@ -74,7 +77,6 @@ module Boxcars
            else
              # otherwise, dig out of the environment
              new_key = ENV.fetch(key.to_s.upcase, nil)
-             send("#{key}=", new_key) if new_key
              new_key
            end
       check_key(key, rv)
@@ -104,6 +106,15 @@ module Boxcars
   def self.engine
     configuration.default_engine || Boxcars::Openai
   end
+
+  # return a proc that will ask the user for input
+  def self.ask_user
+    proc do |changes, _code|
+      puts "This request will make #{changes} changes. Are you sure you want to run it? (y/[n])"
+      answer = gets.chomp
+      answer.downcase == 'y'
+    end
+  end
 end
 
 require "boxcars/version"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: boxcars
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.1.6
 platform: ruby
 authors:
 - Francis Sullivan
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2023-02-22 00:00:00.000000000 Z
+date: 2023-02-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: debug
@@ -81,7 +81,8 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '3.0'
-description: You simply give a number of boxcars to a train, and it does the magic.
+description: You simply set an OpenAI key, give a number of Boxcars to a Train, and
+  magic ensues when you run it.
 email:
 - hi@boxcars.ai
 executables: []
@@ -144,6 +145,6 @@ requirements: []
 rubygems_version: 3.2.32
 signing_key:
 specification_version: 4
-summary: Boxcars provide an API to connect together Boxcars and then conduct them.
+summary: Boxcars is a gem that enables you to create new systems with AI composability.
   Inspired by python langchain.
 test_files: []