promptcraft 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ccfea1d708de15e5bc3e3d8b2bf179813014ba45f84d4a2b950fcb91109c1698
-  data.tar.gz: 36acdfa7e390788b0cc62a3f88bd6259cbc8cab0140d8d1b0c68b9eb197f9839
+  metadata.gz: 7759927da72f5b771061dbe0b7a465e1509728f9ced3b6ad5efc7b341c754576
+  data.tar.gz: e4a89b721bdae7fde071f8c13aea70259bc909c35ab5c77ce82ea9091130441d
 SHA512:
-  metadata.gz: 23b0f0febcf5574f183b6f35a0452c75ba4f4be43e1edb35dfa96c0b957d0d9db42e4bf9a7edaf212d080e5e41b91a959d81d5a56e8addca7819a85552bc43ba
-  data.tar.gz: be1c072618fe888cae89f2e1ef2a5e781b9ce65c3bc38b7359d507487ee95d114a7c76286920fba2cc170adfc7a19f2099581668dde80518e213ac5c83672282
+  metadata.gz: 5a8eaaac272a535c1f9e8b704b05a3871fa0ca681f42657257cdc66778ad69b84a4cbd74b9ecdd0317f7b80b074d6f9b88dded966b43ec1c8a9a43d0198ee289
+  data.tar.gz: 8fc6e1a05a7a9a16a29b48fc47dd9006a06cd487fb476e53a456d39982aad2df8242a66393c3fa7865583c315d2af17a2be569e6125f21451d01ab66c0a6ec46

data/README.md

@@ -19,7 +19,7 @@ messages:
 Let's replay this single conversation with a new system prompt:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --prompt "I'm terrible at maths. If I'm asked a maths question, I reply with a question." \
     --conversation examples/maths/start/already_answered.yml
 ```
@@ -69,7 +69,7 @@ messages:
 The CLI will replay each conversation with the new system prompt.
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/already_answered_multiple.yml \
     --prompt "I like cats. Answer any questions using cats."
 ```
@@ -118,7 +118,7 @@ When you're getting started, you don't even need to know the conversation file f
 
 ```plain
 echo "---\nWhat is 2+2?\n---\nWhat is 6 divided by 2?" | \
-  bundle exec exe/promptcraft --prompt "I solve maths using pizza metaphors."
+  promptcraft --prompt "I solve maths using pizza metaphors."
 ```
 
 The output will be our conversation YAML format, with the system prompt, the incoming user messages as separate conversations, and the assistant replies within each conversation:
@@ -164,7 +164,7 @@ You'll notice, the LLM used (which defaults to Groq's `llama3-70b-8192` because
 Of course, you could pass each plain text user message using the `--conversation` argument too:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation "What is 2+2?"
     --conversation "What is 6 divided by 2?" \
     --prompt "I solve maths using pizza metaphors."
@@ -173,13 +173,13 @@ bundle exec exe/promptcraft \
 Why does it output YAML? (or JSON if you pass `--json` flag) So that you can save it to a file; and then replay (or rechat) this new set of conversations in a minute with a new system prompt.
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation "What is 2+2?" \
     --conversation "What is 6 divided by 2?" \
     --prompt "I am happy person". \
   > tmp/maths-as-happy-person.yml
 
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation tmp/maths-as-happy-person.yml \
     --prompt "I solve maths using pizza metaphors." \
 > tmp/maths-with-pizza.yml
@@ -190,7 +190,7 @@ echo "I am an excellent maths tutor.
 When I'm asked a maths question, I will first
 ask a question in return to help the student." > tmp/prompt-maths-tutor.txt
 
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation tmp/maths-with-pizza.yml \
     --prompt tmp/prompt-maths-tutor.txt
 ```
@@ -210,7 +210,7 @@ messages:
 With each one separated new line. Say nothing else except producing YAML.
 " > tmp/prompt-list-20-hellos.txt
 
-bundle exec exe/promptcraft \
+promptcraft \
   -c "Generate a list of 20 things a customer might say when they first ring into a hair salon phone service" \
   -p tmp/prompt-list-20-hellos.txt \
   --format json > tmp/hair-salon-20-hellos.json
@@ -222,7 +222,7 @@ cat tmp/hair-salon-20-hellos.json | jq -r ".messages[1].content" \
 The file `tmp/hair-salon-20-0000.txt` now contains 20 user messages that you can use to initiate a conversation with your AI assistant system prompt.
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
   -p "I'm a hair salon phone service. I sell haircuts" \
   -c tmp/hair-salon-20-0000.txt \
   > tmp/hair-salon-20-replies-0001.yml
@@ -263,7 +263,27 @@ Tools you might want to use in conjunction with `promptcraft`:
 
 ## Installation
 
-Right now, you need to run the CLI from the source code.
+Right now, you can either install with:
+
+* Homebrew
+* Or, you need to run the CLI from the source code
+
+Whilst this is a RubyGem, it currently requires some Git branches that are not yet released. The Homebrew recipe takes a big tarball of all source code and dependencies and installs it. The tarball is also available via [Github Releases](https://github.com/drnic/promptcraft/releases).
+
+Once the Git branches are released, then the `promptcraft` gem will also be installable via RubyGems.
+
+It requires Ruby 3.3 or later. Mostly because I like the new syntax features.
+
+### Homebrew
+
+The project is currently distributed by the Homebrew tap [`drnic/ai`](https://github.com/drnic/homebrew-ai).
+
+```plain
+brew tap drnic/ai
+brew install promptcraft
+```
+
+### Run from Source
 
 ```plain
 git clone https://github.com/drnic/promptcraft
@@ -275,7 +295,9 @@ bundle exec exe/promptcraft \
     --provider groq
 ```
 
-It defaults to `--provider groq --model llama3-70b-8192` and assumes you have `$GROQ_API_KEY` set in your environment.
+### Configuration
+
+The `promptcraft` CLI defaults to `--provider groq --model llama3-70b-8192` and assumes you have `$GROQ_API_KEY` set in your environment.
 
 You can also use [OpenAI](https://openai.com/) with `--provider openai`, which defaults to `--model gpt-3.5-turbo`. It assumes you have `$OPENAI_API_KEY` set in your environment.
 
@@ -286,7 +308,7 @@ You can also use [Ollama](https://ollama.com/) locally with `--provider ollama`,
 If the conversation file has an `llm` key with `provider` and `model` keys, then those will be used instead of the defaults.
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/already_answered_gpt4.yml \
     --prompt "I always reply with a question"
 
@@ -311,7 +333,7 @@ The following example commands assume you have `$GROQ_API_KEY` and will use Groq
 Run `promptcraft` with no arguments to get a default prompt and an initial assistant message.
 
 ```plain
-bundle exec exe/promptcraft
+promptcraft
 ```
 
 The output might be:
@@ -327,13 +349,13 @@ messages:
 Provide a different provider, such as `openai` (which assumes you have `$OPENAI_API_KEY` set in your environment).
 
 ```plain
-bundle exec exe/promptcraft --provider openai
+promptcraft --provider openai
 ```
 
 Or you could provide your own system prompt, and it will generate an initial assistant message.
 
 ```plaim
-bundle exec exe/promptcraft --prompt "I like to solve maths problems."
+promptcraft --prompt "I like to solve maths problems."
 ```
 
 The output might be:
@@ -367,26 +389,26 @@ The primary point of `promptcraft` is to replay conversations with a new system
 An example of the `--conversation` option:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/basic.yml
 ```
 
 You can also pipe a stream of conversation YAML into `promptcraft` via STDIN
 
 ```plain
-echo "---\nsystem_prompt: I like to solve maths problems.\nmessages:\n- role: \"user\"\n  content: \"What is 2+2?\"" | bundle exec exe/promptcraft
+echo "---\nsystem_prompt: I like to solve maths problems.\nmessages:\n- role: \"user\"\n  content: \"What is 2+2?\"" | promptcraft
 ```
 
 JSON is valid YAML, so you can also use JSON:
 
 ```plain
-echo "{\"system_prompt\": \"I like to solve maths problems.\", \"messages\": [{\"role\": \"user\", \"content\": \"What is 2+2?\"}]}" | bundle exec exe/promptcraft
+echo "{\"system_prompt\": \"I like to solve maths problems.\", \"messages\": [{\"role\": \"user\", \"content\": \"What is 2+2?\"}]}" | promptcraft
 ```
 
 Or pipe one or more files into `promptcraft`:
 
 ```plain
-( cat examples/maths/start/basic.yml ; cat examples/maths/start/already_answered.yml ) | bundle exec exe/promptcraft
+( cat examples/maths/start/basic.yml ; cat examples/maths/start/already_answered.yml ) | promptcraft
 ```
 
 As long as the input is a stream of YAML documents (separated by `---`), it will be processed.
@@ -421,7 +443,7 @@ messages:
 When we replay the conversation with the same system prompt (by omitting the `--prompt` option), it will add the missing assistant reply:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/basic.yml
 ```
 
@@ -440,12 +462,38 @@ messages:
   content: That's an easy one! The answer is... 4!
 ```
 
+### Set the temperature
+
+The `--temperature` option controls the randomness of the assistant's responses. The default for each provider is typically `0.0`. Lower values will produce more deterministic responses, while higher values will produce more creative responses.
+
+```plain
+promptcraft \
+    --conversation examples/maths/start/basic.yml \
+    --temperature 0.5
+```
+
+The output YAML for each conversation will store the `temperature` value used in the `llm:` section:
+
+```yaml
+---
+system_prompt: I like to solve maths problems.
+llm:
+  provider: groq
+  model: llama3-70b-8192
+  temperature: 0.5
+messages:
+- role: user
+  content: What is 2+2?
+- role: assistant
+  content: That's an easy one! The answer is... 4!
+```
+
 ### Limericks
 
 Here are some previously [generated limericks](examples/maths/start/many_limericks.yml). To regenerate them to start with letter "E" on each line:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/many_limericks.yml \
     --prompt "I am excellent at limericks. I always start each line with the letter E."
 ```
@@ -453,7 +501,7 @@ bundle exec exe/promptcraft \
 It might still include some preamble in each response. To try to encourage the LLM to remove it:
 
 ```plain
-bundle exec exe/promptcraft \
+promptcraft \
     --conversation examples/maths/start/many_limericks.yml \
     --prompt "I am excellent at limericks. I always start each line with the letter E. This is very important. Only return the limerick without any other comments."
 ```
@@ -472,12 +520,22 @@ To set new version number:
 gem bump --version [patch|minor|major]
 ```
 
-To tag and release:
+To tag and release Rubygem to <Rubygems.org>:
 
 ```plain
 rake release
 ```
 
+To update Homebrew formula:
+
+```plain
+rake release:build_package
+rake release:upload_package
+rake release:generate_homebrew_formula
+```
+
+Now copy `tmp/promptcraft.rb` formula into <https://github.com/drnic/homebrew-ai> and push.
+
 ## Contributing
 
 Bug reports and pull requests are welcome on GitHub at https://github.com/drnic/promptcraft. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the [code of conduct](https://github.com/drnic/promptcraft/blob/develop/CODE_OF_CONDUCT.md).

data/Rakefile

@@ -2,6 +2,7 @@
 
 require "bundler/gem_tasks"
 require "minitest/test_task"
+Dir["#{File.dirname(__FILE__)}/lib/tasks/**/*.rake"].each { |file| load file }
 
 Minitest::TestTask.create
 

data/lib/promptcraft/cli/run_command.rb

@@ -33,6 +33,13 @@ class Promptcraft::Cli::RunCommand
     desc "String or filename containing system prompt"
   end
 
+  option :temperature do
+    short "-t"
+    long "--temperature temperature"
+    desc "Temperature for chat completion"
+    convert :float
+  end
+
   flag :help do
     short "-h"
     long "--help"
@@ -72,14 +79,16 @@ class Promptcraft::Cli::RunCommand
     desc "Enable debug mode"
   end
 
-  def run(stdin: nil)
+  # Arguments are for the benefit of test suite
+  def run(stdin: nil, threads: nil)
     if params[:help]
       warn help
     elsif params.errors.any?
       warn params.errors.summary
     else
       # Load files in threads
-      pool = Concurrent::FixedThreadPool.new(params[:threads])
+      threads ||= params[:threads]
+      pool = Concurrent::FixedThreadPool.new(threads)
       conversations = Concurrent::Array.new
       # TODO: load in thread pool
       (params[:conversation] || []).each do |filename|
@@ -123,7 +132,7 @@ class Promptcraft::Cli::RunCommand
       end
 
       # Process each conversation in a concurrent thread via a thread pool
-      pool = Concurrent::FixedThreadPool.new(params[:threads])
+      pool = Concurrent::FixedThreadPool.new(threads)
       mutex = Mutex.new
 
       updated_conversations = Concurrent::Array.new
@@ -138,6 +147,7 @@ class Promptcraft::Cli::RunCommand
             Promptcraft::Llm.new
           end
           llm.model = params[:model] if params[:model]
+          llm.temperature = params[:temperature] if params[:temperature]
 
           system_prompt = new_system_prompt || conversation.system_prompt
 

data/lib/promptcraft/conversation.rb

@@ -4,13 +4,14 @@ class Promptcraft::Conversation
   include Promptcraft::Helpers
   extend Promptcraft::Helpers
 
-  attr_accessor :system_prompt, :messages
+  attr_accessor :system_prompt, :messages, :temperature
   attr_accessor :llm
 
-  def initialize(system_prompt:, messages: [], llm: nil)
+  def initialize(system_prompt:, messages: [], llm: nil, temperature: 0)
     @system_prompt = system_prompt
     @messages = messages
     @llm = llm
+    @temperature = temperature
   end
 
   def add_message(role:, content:)
@@ -85,9 +86,9 @@ class Promptcraft::Conversation
   #   content: 2 + 2 = 4
   def to_yaml
     YAML.dump(deep_stringify_keys({
-      system_prompt: @system_prompt&.strip,
-      llm: @llm&.to_h,
-      messages: @messages
+      system_prompt: system_prompt&.strip,
+      llm: llm&.to_h,
+      messages:
     }.compact))
   end
 

data/lib/promptcraft/llm.rb

@@ -4,12 +4,13 @@ class Promptcraft::Llm
   DEFAULT_PROVIDER = "groq"
 
   attr_reader :langchain
-  attr_accessor :provider, :model
+  attr_accessor :provider, :model, :temperature
 
   delegate_missing_to :langchain
 
-  def initialize(provider: DEFAULT_PROVIDER, model: nil, api_key: nil)
+  def initialize(provider: DEFAULT_PROVIDER, model: nil, temperature: nil, api_key: nil)
     @provider = provider
+    @temperature = temperature
     @langchain = case provider
     when "groq"
       @model = model || "llama3-70b-8192"
@@ -17,14 +18,20 @@ class Promptcraft::Llm
       Langchain::LLM::OpenAI.new(
         api_key: api_key || ENV.fetch("GROQ_API_KEY"),
         llm_options: {uri_base: "https://api.groq.com/openai/"},
-        default_options: {chat_completion_model_name: @model}
+        default_options: {
+          temperature: temperature,
+          chat_completion_model_name: @model
+        }.compact
       )
     when "openai"
       @model = model || "gpt-3.5-turbo"
       require "openai"
       Langchain::LLM::OpenAI.new(
         api_key: api_key || ENV.fetch("OPENAI_API_KEY"),
-        default_options: {chat_completion_model_name: @model}
+        default_options: {
+          temperature: temperature,
+          chat_completion_model_name: @model
+        }.compact
       )
     when "openrouter"
       @model = model || "meta-llama/llama-3-8b-instruct:free"
@@ -32,7 +39,10 @@ class Promptcraft::Llm
       Langchain::LLM::OpenAI.new(
         api_key: api_key || ENV.fetch("OPENROUTER_API_KEY"),
         llm_options: {uri_base: "https://openrouter.ai/api/"},
-        default_options: {chat_completion_model_name: @model}
+        default_options: {
+          temperature: temperature,
+          chat_completion_model_name: @model
+        }.compact
       )
     when "ollama"
       @model = model || "llama3"
@@ -46,10 +56,10 @@ class Promptcraft::Llm
   end
 
   def to_h
-    {provider: provider, model: model}
+    {provider:, model:, temperature:}.compact
   end
 
   def self.from_h(hash)
-    new(provider: hash[:provider], model: hash[:model])
+    new(provider: hash[:provider], model: hash[:model], temperature: hash[:temperature])
   end
 end

data/lib/promptcraft/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Promptcraft
-  VERSION = "0.1.0"
+  VERSION = "0.1.1"
 end

data/lib/tasks/release.rake

@@ -0,0 +1,69 @@
+# lib/tasks/homebrew_formula.rake
+
+require "rubygems"
+require "rake"
+
+namespace :release do
+  task :build_package do
+    system "bundle config set cache_all true"
+    system "bundle package"
+    name = Gem::Specification.load(Dir.glob("*.gemspec").first).name
+    version = Gem::Specification.load(Dir.glob("*.gemspec").first).version
+    license = Gem::Specification.load(Dir.glob("*.gemspec").first).license
+    system "rm -f #{name}*.tar* #{name}*.sha256"
+    system "fpm -s dir -t tar --name #{name} --version #{version} --license #{license} --exclude .git --exclude test --exclude spec ."
+    system "mv #{name}.tar #{name}-#{version}.tar"
+    system "xz -z #{name}-#{version}.tar"
+    sha = `shasum -a 256 #{name}-#{version}.tar.xz`.split(" ").first
+    File.write("#{name}-#{version}.sha256", sha)
+  end
+
+  task :upload_package do
+    name = Gem::Specification.load(Dir.glob("*.gemspec").first).name
+    version = Gem::Specification.load(Dir.glob("*.gemspec").first).version
+    file = "#{name}-#{version}.tar.xz"
+    file_sha256 = "#{name}-#{version}.sha256"
+    system "gh release create v#{version} #{file} #{file_sha256} --title 'v#{version}' --notes ''"
+  end
+
+  desc "Generate Homebrew formula"
+  task :generate_homebrew_formula do
+    spec = Gem::Specification.load(Dir.glob("*.gemspec").first)
+    name = spec.name
+    version = spec.version
+    sha256sum = File.read("#{name}-#{version}.sha256").strip
+    url = `gh release view v#{version} --json assets | jq -r '.assets[] | select(.name == "#{name}-#{version}.tar.xz") | .url'`.strip
+
+    formula_name = spec.name.capitalize
+    class_name = formula_name.gsub(/_\w/) { |match| match[1].upcase }.to_s
+
+    formula = <<~RUBY
+      class #{class_name} < Formula
+        desc "#{spec.summary}"
+        homepage "#{spec.homepage}"
+        version "#{spec.version}"
+        url "#{url}"
+        sha256 "#{sha256sum}"
+
+        depends_on "ruby@3.3"
+
+        def install
+          # Extract all files to libexec, which is a common Homebrew practice for third-party tools
+          libexec.install Dir["*"]
+          # Create a symbolic link for the executable in Homebrew's bin directory
+          bin.install_symlink "\#{libexec}/exe/promptcraft" => "promptcraft"
+        end
+
+        test do
+          # Simple test to check the version or a help command
+          system "\#{bin}/promptcraft", "--version"
+        end
+      end
+    RUBY
+
+    Dir.mkdir("tmp") unless Dir.exist?("tmp")
+    path = "tmp/#{spec.name.downcase}.rb"
+    File.write(path, formula)
+    puts "Generated Homebrew formula for #{spec.name} version #{spec.version} at #{path}"
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: promptcraft
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Dr Nic Williams
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-05-13 00:00:00.000000000 Z
+date: 2024-05-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.12.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.12.1
 - !ruby/object:Gem::Dependency
   name: faraday
   requirement: !ruby/object:Gem::Requirement
@@ -185,6 +185,7 @@ files:
 - lib/promptcraft/helpers.rb
 - lib/promptcraft/llm.rb
 - lib/promptcraft/version.rb
+- lib/tasks/release.rake
 - sig/promptcraft.rbs
 homepage: https://github.com/drnic/promptcraft
 licenses: