gpt-cli 0.1.3 → 0.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e970f253447aaac501f343585e284d0b7254b8fc696f4a9ad74430b38861c833
-  data.tar.gz: f1f40e893b47c5bd66d47b8a7156b855156bcb8bda8334340e0a0659746f4c01
+  metadata.gz: 50f343ce40c9b6b1a320305d13c63f53c27565e78b3eba260e24812ca4a59922
+  data.tar.gz: df9ed612ec576b66b925835a5ec8399104937261c5e86f4dac43573b7c842d30
 SHA512:
-  metadata.gz: 3f0e9f58ee5f1f5a03fb20c217d25d3d6da505d3481e9fad56d53dcb465e759da3a48d524aff9e093ca3c69d3dac72fa688e6c9e4301c66d603e67a5bfd4d099
-  data.tar.gz: 20d1bc99c963c155bbd396890cf216d06a665983709a76d2bdeb5179842cef1b5f923b024c5f7ebe794477362156666bc60f67ba2b663564313de6736008e451
+  metadata.gz: a8631cd80e9b5dd4609103b06309c578d06edbb2535b3ca9aa0d6a255126273e34fc84aadc06a13ef0cc51ac8b3ff8897855934508ea6bb9c1acac09eead7e48
+  data.tar.gz: ef833c19b1d265b4b41e7227f3bacadb2a7eba0546cef4598170d1cec7424665416bdd265def89fa6f36565d663cc497d3fd6d226acaebd2a0f68ec7c007d70c

data/CHANGELOG.md

@@ -9,6 +9,12 @@
 - Define open ai model in env var
 - Input context pompts
 
-## [0.1.3] - 2023-03-23
+## [0.1.3] - 2023-03-25
 
-- Custom contexts prompts support
+- Custom contexts prompts support
+
+## [0.1.4] - 2023-03-25
+
+- Added messages history
+- Added more context prompts
+- Added Dall-E support

data/README.md

@@ -1,17 +1,33 @@
 # GPTCLI
 
-This library provides a UNIX-ey interface to OpenAI.
-It is a fork of [openai_pipe](https://github.com/Aesthetikx/openai_pipe), created by [Aesthetikx](https://github.com/Aesthetikx/), kudos to him for his awesome work!  
-The goal of this fork was to add the ability to choose which OpenAI model to use, as well as adding context prompts support. Thanks to these new features, you can have your own highly specialized GPT4 personal assistant, directly in your terminal! Productivity stonks 📈📈📈   
+This library provides a UNIX-ey interface to OpenAI.  
+Say hello to your own highly specialized GPT4 personal assistant, directly in your terminal!  
+Productivity stonks 📈📈📈   
+
+Consider 🌟 this repo if you find this tool useful 🔥  
+
+## Features
+- Use the pipe to pass the output of commands to GPT or Dall-E.  
+- Choose your GPT model, GPT4 supported if you have access.  
+- Dall-E support: generate images directly in your terminal.  
+- Contexts prompts tailored for your needs: add/update predefined context prompts or write your own when running the command.  
+- Infinite conversation history: have chat sessions with GPT just like [chat.openai.com](https://chat.openai.com/) but in your terminal.  
 
 See [Installation](#installation) and [Setup](#setup) below, but first, some examples.
 
 ## Examples
 
 ```console
-$ gpt what is two plus two
-Two plus two is equal to four.
+$ gpt "what is your role?"
+As your personal assistant, my role is to assist you in various tasks and answer your questions related to my areas of expertise, which include UNIX systems, bash, Python, Django, SQL, Javascript, ReactJS. I can help you with programming and development, server administration, debugging your code or scripts, optimizing performance, code review, providing recommendations for best practices, and more.
+```
+
+Use Dall-E:
+```console
+$ gpt give me a very short description of the moon landscape | gpt --dalle
+Image saved in current directory to 7db7ab2e8914175d4f4819e033226563.png
 ```
+<img src="https://raw.githubusercontent.com/FlorianMgs/gpt-cli/master/.github/7db7ab2e8914175d4f4819e033226563.png" height=256 width=256></img>
 
 ```console
 $ uptime | gpt convert this to json
@@ -27,26 +43,6 @@ $ uptime | gpt convert this to json
 }
 ```
 
-```console
-$ gpt -c you are Vitalik Buterin, the creator of Ethereum. You know very well the whole EVM ecosystem and how to write perfectly optimized Solidity smart contracts -p Write a simple ERC20 token smart contract using OpenZeppelin library, respond only by the smart contract, do no write explanations > erc20.sol
-pragma solidity ^0.8.0;
-
-import "@openzeppelin/contracts/token/ERC20/ERC20.sol";
-
-contract MyToken is ERC20 {
-    constructor(uint256 initialSupply) ERC20("MyToken", "MTK") {
-        _mint(msg.sender, initialSupply);
-    }
-}
-```
-
-Setting default context to `fullstack`:
-```console
-$ gpt "what is your role?"
-As your personal assistant, my role is to assist you in various tasks and answer your questions related to my areas of expertise, which include UNIX systems, bash, Python, Django, SQL, Javascript, ReactJS, and NextJS. I can help you with programming and development, server administration, debugging your code or scripts, optimizing performance, code review, providing recommendations for best practices, and more.
-
-```
-
 ```console
 $ gpt list the nine planets as JSON | gpt convert this to XML but in French | tee planets.fr.xml
 <Planètes>
@@ -62,6 +58,19 @@ $ gpt list the nine planets as JSON | gpt convert this to XML but in French | te
 </Planètes>
 ```
 
+```console
+$ gpt -c "you are Vitalik Buterin, the creator of Ethereum. You know very well the whole EVM ecosystem and how to write perfectly optimized Solidity smart contracts" -p "Write a simple ERC20 token smart contract using OpenZeppelin library, respond only by the smart contract, do no write explanations" > erc20.sol
+pragma solidity ^0.8.0;
+
+import "@openzeppelin/contracts/token/ERC20/ERC20.sol";
+
+contract MyToken is ERC20 {
+    constructor(uint256 initialSupply) ERC20("MyToken", "MTK") {
+        _mint(msg.sender, initialSupply);
+    }
+}
+```
+
 ```console
 $ curl -sL "https://en.wikipedia.org/wiki/cats" | head -n 5 | gpt extract just the title of this webpage | figlet
   ____      _            __        ___ _    _                _ _
@@ -185,11 +194,11 @@ Set the model you want to use in ENV:
 ```bash
 export OPENAI_MODEL="gpt-3.5-turbo"
 ```
-Copy `gpt_contexts.sample.json` somewhere, for example `~/Documents/gpt_contexts.json`, then put the file path in ENV:
+Copy `gpt_contexts.sample.json` somewhere, for example `~/Documents/gpt_contexts.json`, then put the file path in ENV (don't forget to rename the file to `gpt_contexts.json`):
 ```bash
-export OPENAI_CONTEXTS_PATH="path to gpt_contexts.json"
+export OPENAI_CONTEXTS_PATH="path/to/gpt_contexts.json"
 ```
-(Optional) set the default context prompt you want to use (for now, there's only 3 basic prompts available `python`, `fullstack` and `blockchain` in `gpt_contexts.sample.json`):
+(Optional) set the default context prompt you want to use, see `gpt_contexts.sample.json` for examples:
 ```bash
 export OPENAI_DEFAULT_CONTEXT="python"
 ```
@@ -200,26 +209,31 @@ alias gpt="gpt-cli"
 
 ## Usage
 
-There's two optional parameters you can set when running `gpt`:  
+There's optional parameters you can set when running `gpt`:  
 `gpt -c <custom_context_prompt> -p <your prompt>`  
 
-`--context -c`: this will be the context prompt, see basic contexts in `gpt_contexts.sample.json`. You can put a key from `gpt_contexts.json` or a custom context. Default to ENV `OPENAI_DEFAULT_CONTEXT` if not set. Add your own context prompts in `gpt_contexts.json`.      
+`--context -c`: this will be the context prompt, see basic contexts in `gpt_contexts.sample.json`. You can put a key from `gpt_contexts.json` or a custom context. Default to ENV `OPENAI_DEFAULT_CONTEXT` if not set. Add your own context prompts in `gpt_contexts.json`.  
 
 `--prompt -p`: your actual prompt.  
 
-You can also run gpt without any arguments, just your question. In this case, the context prompt will default to the one defined by ENV var `OPENAI_DEFAULT_CONTEXT` if it exists. If not, no context will be added.  
-See examples above for an overview of some usecases. Possibilities are endless.
+`--history -h`: print conversation history.  
+
+`--dalle -d`: Generate an image from your prompt with Dall-E.  
+
+`--clear`: clear conversation history.  
+
+You can also run gpt without any arguments, just with your prompt. In this case, the context prompt will default to the one defined by ENV var `OPENAI_DEFAULT_CONTEXT` if it exists. If not, no context will be added.  
+See examples above for an overview of some usecases. Possibilities are endless.  
 
 ## Notes
 
-Be aware that there is a cost associated every time GPT3 is invoked, so be mindful of your account usage. Also be wary of sending sensitive data to OpenAI, and also wary of arbitrarily executing scripts or programs that GPT3 generates.
+This is made on top of [openai_pipe](https://github.com/Aesthetikx/openai_pipe), created by [Aesthetikx](https://github.com/Aesthetikx/), kudos to him for his awesome work!  
+Be aware that there is a cost associated every time GPT is invoked, so be mindful of your account usage. Also be wary of sending sensitive data to OpenAI, and also wary of arbitrarily executing scripts or programs that GPT generates.
 Also, this is my very first time working in Ruby. So please be indulgent 🙏  
 
 ## TODO
 
-- Add conversation history support  
 - Add internet search support to feed prompt more accurately  
-- Add dall-e support  
 - Add proper documentation  
 
 ## Development

data/gpt_contexts.sample.json

@@ -1,5 +1,12 @@
 {
-    "python": "Imagine you are a senior software developer specialized in Python, Django, SQL and Machine Learning with 20 years of experience. You are very proficient at writing high quality, clean, concise and highly optimized code that follows the PEP recommendations. You are my personal assistant. You are here to help me in various tasks and answer my questions.",
-    "fullstack": "Imagine you are an Senior Fullstack Engineer with 20 of experience in UNIX systems, bash, Python, Django, SQL, Javascript, ReactJS. You are very proficient at writing high quality, clean, concise and well optimized code that follows the PEP recommendations. You are my personal assistant. You are here to help me in various tasks and answer my questions.",
-    "blockchain": "Imagine you are Vitalik Buterin. You created Ethereum and you are the most famous blockchain developer in the world.You have a very deep knowledge about EVM, Solidity, and the whole blockchain ecosystem.You know everything about the latest trends in the blockchain space and you are very proficient at writing high quality, clean, concise and well optimized code that follows the good practices.You are my personal assistant. You are here to help me in various tasks and answer my questions."
+    "python": "Imagine you are a senior software developer specialized in Python, Django, and SQL with 20 years of experience. You are very proficient at writing high quality, clean, concise and highly optimized code that follows the PEP recommendations. You are my personal assistant. You are here to help me in various tasks and answer my questions.",
+    "fullstack": "Imagine you are an Senior Fullstack Engineer with 20 of experience. You have a very deep knowledge about various backend tech stacks and frontend frameworks. You are very proficient at writing high quality, clean, concise and well optimized code. You are my personal assistant. You are here to help me in various tasks and answer my questions.",
+    "blockchain": "Imagine you are Vitalik Buterin. You created Ethereum and you are the most famous blockchain developer in the world. You have a very deep knowledge about EVM, Solidity, and the whole blockchain ecosystem.You know everything about the latest trends in the blockchain space and you are very proficient at writing high quality, clean, concise and well optimized code that follows the good practices.You are my personal assistant. You are here to help me in various tasks and answer my questions.",
+    "it_expert": "I want you to act as an IT Expert. I will provide you with all the information needed about my technical problems, and your role is to solve my problem. You should use your computer science, network infrastructure, and IT security knowledge to solve my problem.",
+    "ml_expert": "I want you to act as a machine learning engineer. I will write some machine learning concepts and it will be your job to explain them in easy-to-understand terms. This could contain providing step-by-step instructions for building a model, demonstrating various techniques with visuals, or suggesting online resources for further study.",
+    "hacker": "I want you to act as a cyber security specialist. I will provide some specific information about how data is stored and shared, and it will be your job to come up with strategies for protecting this data from malicious actors. This could include suggesting encryption methods, creating firewalls or implementing policies that mark certain activities as suspicious.",
+    "math": "I want you to act as a math teacher. I will provide some mathematical equations or concepts, and it will be your job to explain them in easy-to-understand terms. This could include providing step-by-step instructions for solving a problem, demonstrating various techniques with visuals or suggesting online resources for further study.",
+    "tech_transferer": "I want you to act as a Technology Transferer, I will provide resume bullet points and you will map each bullet point from one technology to a different technology. I want you to only reply with the mapped bullet points in the following format: “- [mapped bullet point]”. Do not write explanations. Do not provide additional actions unless instructed. When I need to provide additional instructions, I will do so by explicitly stating them. The technology in the original resume bullet point is {Android} and the technology I want to map to is {ReactJS}",
+    "prompt_engineer": "I want you to act as a ChatGPT prompt generator, I will send a topic, you have to generate a ChatGPT prompt based on the content of the topic, the prompt should start with “I want you to act as “, and guess what I might do, and expand the prompt accordingly Describe the content to make it useful.",
+    "linux_terminal": "I want you to act as a linux terminal. I will type commands and you will reply with what the terminal should show. I want you to only reply with the terminal output inside one unique code block, and nothing else. do not write explanations. do not type commands unless I instruct you to do so. When I need to tell you something in English, I will do so by putting text inside curly brackets {like this}."
 }

data/lib/gpt-cli/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module GPTCLI
-  VERSION = "0.1.3"
+  VERSION = "0.1.4"
 end

data/lib/gpt-cli.rb

@@ -2,11 +2,32 @@
 
 require_relative "gpt-cli/version"
 
+require 'digest/md5'
 require "json"
 require "quick_openai"
 require "optparse"
 
 class ChatGPT
+  attr_accessor :messages
+  MESSAGES_FILE = 'message_history.json'
+
+  def initialize
+    @messages = load_messages
+  end
+
+  def save_messages
+    File.open(MESSAGES_FILE, 'w') do |file|
+      file.write(JSON.dump(@messages))
+    end
+  end
+
+  def load_messages
+    return [] unless File.exist?(MESSAGES_FILE)
+
+    file_contents = File.read(MESSAGES_FILE)
+    JSON.parse(file_contents, symbolize_names: true)
+  end
+
   def contexts
     file_contents = File.read(ENV["OPENAI_CONTEXTS_PATH"])
     contexts = JSON.parse(file_contents)
@@ -15,15 +36,14 @@ class ChatGPT
 
   def gpt3(prompt, options)
     context = options[:context] || contexts[ENV["OPENAI_DEFAULT_CONTEXT"].to_sym]
-    messages = [
-      {role: "user", content: prompt}
-    ]
-    messages.unshift({role: "system", content: context.gsub("\n", ' ').squeeze(' ')}) if context
+    context_message = {role: "system", content: context.gsub("\n", ' ').squeeze(' ')}
+    @messages << context_message if context && !@messages.include?(context_message)
+    @messages << {role: "user", content: prompt}
 
     parameters = {
       model: ENV["OPENAI_MODEL"],
       max_tokens: 2048,
-      messages: messages,
+      messages: @messages,
     }
     
     response = QuickOpenAI.fetch_response_from_client do |client|
@@ -33,7 +53,8 @@ class ChatGPT
     text = response.dig("choices", 0, "message", "content")
 
     raise QuickOpenAI::Error, "Unable to parse response." if text.nil? || text.empty?
-
+    
+    @messages << {role: "assistant", content: text.chomp.strip}
     text.chomp.strip
   end
 end
@@ -51,20 +72,51 @@ module GPTCLI
       opts.on('-p', '--prompt PROMPT_TEXT', 'Prompt text to be passed to GPT-3') do |prompt_text|
         options[:prompt] = prompt_text
       end
+      opts.on('-h', '--history', 'Print the message history') do
+        options[:history] = true
+      end
+      opts.on('--clear', 'Clear the message history') do
+        options[:clear] = true
+      end
+      opts.on('-d', '--dalle', 'Use DALL-E instead of GPT. Prompt should be no more than 1000 characters.') do
+        options[:dalle] = true
+      end
     end
 
     remaining_args = parser.parse!(ARGV)
 
-    input_from_pipe = $stdin.read if $stdin.stat.pipe?
-    input_from_remaining_args = remaining_args.join(" ") if remaining_args.any?
+    if options[:history]
+      puts chatgpt.messages
+    elsif options[:clear]
+      File.delete(ChatGPT::MESSAGES_FILE) if File.exist?(ChatGPT::MESSAGES_FILE)
+      puts "Message history cleared."
+    else
+      input_from_pipe = $stdin.read if $stdin.stat.pipe?
+      input_from_remaining_args = remaining_args.join(" ") if remaining_args.any?
 
-    prompt = options[:prompt] || input_from_remaining_args || ""
-    full_prompt = [prompt, input_from_pipe].compact.join("\n\n")
-    full_prompt.strip!
+      prompt = options[:prompt] || input_from_remaining_args || ""
+      full_prompt = [prompt, input_from_pipe].compact.join("\n\n")
+      full_prompt.strip!
 
-    puts chatgpt.gpt3(full_prompt, options)
+      if options[:dalle]
+        if full_prompt.length > 1000
+          puts "Prompt is too long. Truncating to 1000 characters."
+          full_prompt = full_prompt[0...1000]
+        end
+        full_prompt.dalle2.then { |tempfile|
+          current_working_directory = Dir.pwd
+          filename = Digest::MD5.hexdigest(full_prompt)
+          output_file_path = File.join(current_working_directory, "#{filename}.png")
+          File.write(output_file_path, tempfile.read)
+          puts "Image saved in current directory to #{filename}.png"
+        }
+      else
+        puts chatgpt.gpt3(full_prompt, options)
+        chatgpt.save_messages
+      end
+    end
   rescue QuickOpenAI::Error => e
     warn e.message
     exit 1
   end
-end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gpt-cli
 version: !ruby/object:Gem::Version
-  version: 0.1.3
+  version: 0.1.4
 platform: ruby
 authors:
 - John DeSilva