langchainrb 0.5.6 → 0.5.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 045a900e70f73ac0c969ea0e7cb0130d12219ad869583a55d5f5857ceccac618
-  data.tar.gz: f6202d34280eeda69026add6cb0bcadb1625da58ed729e1b4ca02c2cfdbd76b4
+  metadata.gz: ba5e9e8257d18c0940fdaf4fe84c03d594d8f1151e40e1bb35de059f8e6e5094
+  data.tar.gz: 11310635819502b9bfbd66bc45dc7aa1ce500d4a874dcc5ab550d6c5edf7194f
 SHA512:
-  metadata.gz: b5cd3983b8a7389baace3befd24751d1c2974b94da29868fac6bfcd048681d2b6cc603d13f791d7ca4bffbc18b9278704c3db188112b51f1c71ac528c6c04f70
-  data.tar.gz: c061c1a877bc94488177ef79a46ed558540ba664a001a463a95fbe7f1f5f50c8895f359ab06fce26bb7dedf8cd246713b96e553fdfa55ca5b68c78f124e87a2a
+  metadata.gz: 4b97e21bcbc0c5f1d842271b64949c07d6d78190cd97c22fd0dab735d6b6ae2f2e6328ba2631dfc77ed0a5dd227573e3f84f064e8dd9332701848a798747ac9a
+  data.tar.gz: 267b2029de10acf45bb97a040d174102f666e048aaaf03ab76218cd5281574c1ae977ba8e975faf4b690e677611daba2fb0fc975801c0e41072f050ec2ac2e34

data/CHANGELOG.md

@@ -1,5 +1,9 @@
 ## [Unreleased]
 
+## [0.5.7] - 2023-06-19
+- Developer can modify models used when initiliazing `Langchain::LLM::*` clients
+- Improvements to the `SQLQueryAgent` and the database tool
+
 ## [0.5.6] - 2023-06-18
 - If used with OpenAI, Langchain::Conversation responses can now be streamed.
 - Improved logging

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    langchainrb (0.5.6)
+    langchainrb (0.5.7)
       baran (~> 0.1.6)
       colorize (~> 0.8.1)
       tiktoken_ruby (~> 0.0.5)

data/README.md

@@ -10,7 +10,7 @@
 [![Gem Version](https://badge.fury.io/rb/langchainrb.svg)](https://badge.fury.io/rb/langchainrb)
 [![Docs](http://img.shields.io/badge/yard-docs-blue.svg)](http://rubydoc.info/gems/langchainrb)
 [![License](https://img.shields.io/badge/license-MIT-green.svg)](https://github.com/andreibondarev/langchainrb/blob/main/LICENSE.txt)
-[![](https://dcbadge.vercel.app/api/server/WWqjwxMv?compact=true&style=flat)](https://discord.gg/WWqjwxMv)
+[![](https://dcbadge.vercel.app/api/server/WDARp7J2n8?compact=true&style=flat)](https://discord.gg/WDARp7J2n8)
 
 
 Langchain.rb is a library that's an abstraction layer on top many emergent AI, ML and other DS tools. The goal is to abstract complexity and difficult concepts to make building AI/ML-supercharged applications approachable for traditional software engineers.
@@ -367,7 +367,7 @@ Langchain.logger.level = :info
 5. Optionally, install lefthook git hooks for pre-commit to auto lint: `gem install lefthook && lefthook install -f`
 
 ## Discord
-Join us in the [Langchain.rb](https://discord.gg/hXutDWGDd) Discord server.
+Join us in the [Langchain.rb](https://discord.gg/WDARp7J2n8) Discord server.
 
 ## Core Contributors
 [<img style="border-radius:50%" alt="Andrei Bondarev" src="https://avatars.githubusercontent.com/u/541665?v=4" width="80" height="80" class="avatar">](https://github.com/andreibondarev)

data/lib/langchain/agent/sql_query_agent/sql_query_agent.rb

@@ -13,7 +13,7 @@ module Langchain::Agent
     def initialize(llm:, db:)
       @llm = llm
       @db = db
-      @schema = @db.schema
+      @schema = @db.dump_schema
     end
 
     #

data/lib/langchain/llm/ai21.rb

@@ -11,11 +11,17 @@ module Langchain::LLM
   #     ai21 = Langchain::LLM::AI21.new(api_key:)
   #
   class AI21 < Base
-    def initialize(api_key:)
+    DEFAULTS = {
+      temperature: 0.0,
+      model: "j2-large"
+    }.freeze
+
+    def initialize(api_key:, default_options: {})
       depends_on "ai21"
       require "ai21"
 
       @client = ::AI21::Client.new(api_key)
+      @defaults = DEFAULTS.merge(default_options)
     end
 
     #
@@ -26,7 +32,9 @@ module Langchain::LLM
     # @return [String] The completion
     #
     def complete(prompt:, **params)
-      response = client.complete(prompt, params)
+      parameters = complete_parameters params
+
+      response = client.complete(prompt, parameters)
       response.dig(:completions, 0, :data, :text)
     end
 
@@ -41,5 +49,11 @@ module Langchain::LLM
       response = client.summarize(text, "TEXT", params)
       response.dig(:summary)
     end
+
+    private
+
+    def complete_parameters(params)
+      @defaults.dup.merge(params)
+    end
   end
 end

data/lib/langchain/llm/cohere.rb

@@ -18,11 +18,12 @@ module Langchain::LLM
       dimension: 1024
     }.freeze
 
-    def initialize(api_key:)
+    def initialize(api_key:, default_options: {})
       depends_on "cohere-ruby"
       require "cohere"
 
       @client = ::Cohere::Client.new(api_key: api_key)
+      @defaults = DEFAULTS.merge(default_options)
     end
 
     #
@@ -34,7 +35,7 @@ module Langchain::LLM
     def embed(text:)
       response = client.embed(
         texts: [text],
-        model: DEFAULTS[:embeddings_model_name]
+        model: @defaults[:embeddings_model_name]
       )
       response.dig("embeddings").first
     end
@@ -49,8 +50,8 @@ module Langchain::LLM
     def complete(prompt:, **params)
       default_params = {
         prompt: prompt,
-        temperature: DEFAULTS[:temperature],
-        model: DEFAULTS[:completion_model_name]
+        temperature: @defaults[:temperature],
+        model: @defaults[:completion_model_name]
       }
 
       if params[:stop_sequences]

data/lib/langchain/llm/google_palm.rb

@@ -22,15 +22,19 @@ module Langchain::LLM
 
     DEFAULTS = {
       temperature: 0.0,
-      dimension: 768 # This is what the `embedding-gecko-001` model generates
+      dimension: 768, # This is what the `embedding-gecko-001` model generates
+      completion_model_name: "text-bison-001",
+      chat_completion_model_name: "chat-bison-001",
+      embeddings_model_name: "embedding-gecko-001"
     }.freeze
     LENGTH_VALIDATOR = Langchain::Utils::TokenLength::GooglePalmValidator
 
-    def initialize(api_key:)
+    def initialize(api_key:, default_options: {})
       depends_on "google_palm_api"
       require "google_palm_api"
 
       @client = ::GooglePalmApi::Client.new(api_key: api_key)
+      @defaults = DEFAULTS.merge(default_options)
     end
 
     #
@@ -56,7 +60,8 @@ module Langchain::LLM
     def complete(prompt:, **params)
       default_params = {
         prompt: prompt,
-        temperature: DEFAULTS[:temperature]
+        temperature: @defaults[:temperature],
+        completion_model_name: @defaults[:completion_model_name]
       }
 
       if params[:stop_sequences]
@@ -85,12 +90,14 @@ module Langchain::LLM
       raise ArgumentError.new(":prompt or :messages argument is expected") if prompt.empty? && messages.empty?
 
       default_params = {
-        temperature: DEFAULTS[:temperature],
+        temperature: @defaults[:temperature],
+        chat_completion_model_name: @defaults[:chat_completion_model_name],
         context: context,
         messages: compose_chat_messages(prompt: prompt, messages: messages),
         examples: compose_examples(examples)
       }
 
+      # chat-bison-001 is the only model that currently supports countMessageTokens functions
       LENGTH_VALIDATOR.validate_max_tokens!(default_params[:messages], "chat-bison-001", llm: self)
 
       if options[:stop_sequences]
@@ -123,7 +130,7 @@ module Langchain::LLM
 
       complete(
         prompt: prompt,
-        temperature: DEFAULTS[:temperature],
+        temperature: @defaults[:temperature],
         # Most models have a context length of 2048 tokens (except for the newest models, which support 4096).
         max_tokens: 2048
       )

data/lib/langchain/llm/openai.rb

@@ -19,11 +19,12 @@ module Langchain::LLM
     }.freeze
     LENGTH_VALIDATOR = Langchain::Utils::TokenLength::OpenAIValidator
 
-    def initialize(api_key:, llm_options: {})
+    def initialize(api_key:, llm_options: {}, default_options: {})
       depends_on "ruby-openai"
       require "openai"
 
       @client = ::OpenAI::Client.new(access_token: api_key, **llm_options)
+      @defaults = DEFAULTS.merge(default_options)
     end
 
     #
@@ -34,7 +35,7 @@ module Langchain::LLM
     # @return [Array] The embedding
     #
     def embed(text:, **params)
-      parameters = {model: DEFAULTS[:embeddings_model_name], input: text}
+      parameters = {model: @defaults[:embeddings_model_name], input: text}
 
       validate_max_tokens(text, parameters[:model])
 
@@ -50,7 +51,7 @@ module Langchain::LLM
     # @return [String] The completion
     #
     def complete(prompt:, **params)
-      parameters = compose_parameters DEFAULTS[:completion_model_name], params
+      parameters = compose_parameters @defaults[:completion_model_name], params
 
       parameters[:prompt] = prompt
       parameters[:max_tokens] = validate_max_tokens(prompt, parameters[:model])
@@ -60,20 +61,59 @@ module Langchain::LLM
     end
 
     #
-    # Generate a chat completion for a given prompt
+    # Generate a chat completion for a given prompt or messages.
+    #
+    # == Examples
+    #
+    #     # simplest case, just give a prompt
+    #     openai.chat prompt: "When was Ruby first released?"
+    #
+    #     # prompt plus some context about how to respond
+    #     openai.chat context: "You are RubyGPT, a helpful chat bot for helping people learn Ruby", prompt: "Does Ruby have a REPL like IPython?"
+    #
+    #     # full control over messages that get sent, equivilent to the above
+    #     openai.chat messages: [
+    #       {
+    #         role: "system",
+    #         content: "You are RubyGPT, a helpful chat bot for helping people learn Ruby", prompt: "Does Ruby have a REPL like IPython?"
+    #       },
+    #       {
+    #         role: "user",
+    #         content: "When was Ruby first released?"
+    #       }
+    #     ]
+    #
+    #     # few-short prompting with examples
+    #     openai.chat prompt: "When was factory_bot released?",
+    #       examples: [
+    #         {
+    #           role: "user",
+    #           content: "When was Ruby on Rails released?"
+    #         }
+    #         {
+    #           role: "assistant",
+    #           content: "2004"
+    #         },
+    #       ]
     #
     # @param prompt [String] The prompt to generate a chat completion for
-    # @param messages [Array] The messages that have been sent in the conversation
-    # @param context [String] The context of the conversation
-    # @param examples [Array] Examples of messages provide model with
-    # @param options extra parameters passed to OpenAI::Client#chat
-    # @param block [Block] Pass the block to stream the response
+    # @param messages [Array<Hash>] The messages that have been sent in the conversation
+    #   Each message should be a Hash with the following keys:
+    #   - :content [String] The content of the message
+    #   - :role [String] The role of the sender (system, user, assistant, or function)
+    # @param context [String] An initial context to provide as a system message, ie "You are RubyGPT, a helpful chat bot for helping people learn Ruby"
+    # @param examples [Array<Hash>] Examples of messages to provide to the model. Useful for Few-Shot Prompting
+    #   Each message should be a Hash with the following keys:
+    #   - :content [String] The content of the message
+    #   - :role [String] The role of the sender (system, user, assistant, or function)
+    # @param options <Hash> extra parameters passed to OpenAI::Client#chat
+    # @yield [String] Stream responses back one String at a time
     # @return [String] The chat completion
     #
     def chat(prompt: "", messages: [], context: "", examples: [], **options)
       raise ArgumentError.new(":prompt or :messages argument is expected") if prompt.empty? && messages.empty?
 
-      parameters = compose_parameters DEFAULTS[:chat_completion_model_name], options
+      parameters = compose_parameters @defaults[:chat_completion_model_name], options
       parameters[:messages] = compose_chat_messages(prompt: prompt, messages: messages, context: context, examples: examples)
       parameters[:max_tokens] = validate_max_tokens(parameters[:messages], parameters[:model])
 
@@ -104,13 +144,13 @@ module Langchain::LLM
       )
       prompt = prompt_template.format(text: text)
 
-      complete(prompt: prompt, temperature: DEFAULTS[:temperature])
+      complete(prompt: prompt, temperature: @defaults[:temperature])
     end
 
     private
 
     def compose_parameters(model, params)
-      default_params = {model: model, temperature: DEFAULTS[:temperature]}
+      default_params = {model: model, temperature: @defaults[:temperature]}
 
       default_params[:stop] = params.delete(:stop_sequences) if params[:stop_sequences]
 

data/lib/langchain/llm/replicate.rb

@@ -32,7 +32,7 @@ module Langchain::LLM
     #
     # @param api_key [String] The API key to use
     #
-    def initialize(api_key:)
+    def initialize(api_key:, default_options: {})
       depends_on "replicate-ruby"
       require "replicate"
 
@@ -41,6 +41,7 @@ module Langchain::LLM
       end
 
       @client = ::Replicate.client
+      @defaults = DEFAULTS.merge(default_options)
     end
 
     #
@@ -100,7 +101,7 @@ module Langchain::LLM
 
       complete(
         prompt: prompt,
-        temperature: DEFAULTS[:temperature],
+        temperature: @defaults[:temperature],
         # Most models have a context length of 2048 tokens (except for the newest models, which support 4096).
         max_tokens: 2048
       )
@@ -111,11 +112,11 @@ module Langchain::LLM
     private
 
     def completion_model
-      @completion_model ||= client.retrieve_model(DEFAULTS[:completion_model_name]).latest_version
+      @completion_model ||= client.retrieve_model(@defaults[:completion_model_name]).latest_version
     end
 
     def embeddings_model
-      @embeddings_model ||= client.retrieve_model(DEFAULTS[:embeddings_model_name]).latest_version
+      @embeddings_model ||= client.retrieve_model(@defaults[:embeddings_model_name]).latest_version
     end
   end
 end

data/lib/langchain/tool/calculator.rb

@@ -16,6 +16,11 @@ module Langchain::Tool
       Useful for getting the result of a math expression.
 
       The input to this tool should be a valid mathematical expression that could be executed by a simple calculator.
+      Usage:
+        Action Input: 1 + 1
+        Action Input: 3 * 2 / 4
+        Action Input: 9 - 7
+        Action Input: (4.1 + 2.3) / (2.0 - 5.6) * 3
     DESC
 
     def initialize
@@ -32,14 +37,7 @@ module Langchain::Tool
 
       Eqn::Calculator.calc(input)
     rescue Eqn::ParseError, Eqn::NoVariableValueError
-      # Sometimes the input is not a pure math expression, e.g: "12F in Celsius"
-      # We can use the google answer box to evaluate this expression
-      # TODO: Figure out to find a better way to evaluate these language expressions.
-      hash_results = Langchain::Tool::GoogleSearch
-        .new(api_key: ENV["SERPAPI_API_KEY"])
-        .execute_search(input: input)
-      hash_results.dig(:answer_box, :to) ||
-        hash_results.dig(:answer_box, :result)
+      "\"#{input}\" is an invalid mathematical expression"
     end
   end
 end

data/lib/langchain/tool/database.rb

@@ -14,15 +14,18 @@ module Langchain::Tool
       The input to this tool should be valid SQL.
     DESC
 
-    attr_reader :db
+    attr_reader :db, :requested_tables, :except_tables
 
     #
     # Establish a database connection
     #
     # @param connection_string [String] Database connection info, e.g. 'postgres://user:password@localhost:5432/db_name'
+    # @param tables [Array<Symbol>] The tables to use. Will use all if empty.
+    # @param except_tables [Array<Symbol>] The tables to exclude. Will exclude none if empty.
+
     # @return [Database] Database object
     #
-    def initialize(connection_string:)
+    def initialize(connection_string:, tables: [], except_tables: [])
       depends_on "sequel"
       require "sequel"
       require "sequel/extensions/schema_dumper"
@@ -30,7 +33,8 @@ module Langchain::Tool
       raise StandardError, "connection_string parameter cannot be blank" if connection_string.empty?
 
       @db = Sequel.connect(connection_string)
-      @db.extension :schema_dumper
+      @requested_tables = tables
+      @except_tables = except_tables
     end
 
     #
@@ -38,9 +42,26 @@ module Langchain::Tool
     #
     # @return [String] schema
     #
-    def schema
-      Langchain.logger.info("Dumping schema", for: self.class)
-      db.dump_schema_migration(same_db: true, indexes: false) unless db.adapter_scheme == :mock
+    def dump_schema
+      Langchain.logger.info("Dumping schema tables and keys", for: self.class)
+      schema = ""
+      db.tables.each do |table|
+        next if except_tables.include?(table)
+        next unless requested_tables.empty? || requested_tables.include?(table)
+
+        schema << "CREATE TABLE #{table}(\n"
+        db.schema(table).each do |column|
+          schema << "#{column[0]} #{column[1][:type]}"
+          schema << " PRIMARY KEY" if column[1][:primary_key] == true
+          schema << "," unless column == db.schema(table).last
+          schema << "\n"
+        end
+        schema << ");\n"
+        db.foreign_key_list(table).each do |fk|
+          schema << "ALTER TABLE #{table} ADD FOREIGN KEY (#{fk[:columns][0]}) REFERENCES #{fk[:table]}(#{fk[:key][0]});\n"
+        end
+      end
+      schema
     end
 
     #

data/lib/langchain/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Langchain
-  VERSION = "0.5.6"
+  VERSION = "0.5.7"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: langchainrb
 version: !ruby/object:Gem::Version
-  version: 0.5.6
+  version: 0.5.7
 platform: ruby
 authors:
 - Andrei Bondarev
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2023-06-18 00:00:00.000000000 Z
+date: 2023-06-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: baran