chatgpt.sh

#!/bin/bash

GLOBIGNORE="*"

CHAT_INIT_PROMPT="You are ChatGPT, a Large Language Model trained by OpenAI. You will be answering questions from users. You answer as concisely as possible for each response (e.g. don’t be verbose). If you are generating a list, do not have too many items. Keep the number of items short. Before each user prompt you will be given the chat history in Q&A form. Output your answer directly, with no labels in front. Do not start your answers with A or Anwser. You were trained on data up until 2021. Today's date is $(date +%d/%m/%Y)"

SYSTEM_PROMPT="You are ChatGPT, a large language model trained by OpenAI. Answer as concisely as possible. Current date: $(date +%d/%m/%Y). Knowledge cutoff: 9/1/2021."

COMMAND_GENERATION_PROMPT="You are a Command Line Interface expert and your task is to provide functioning shell commands. Return a CLI command and nothing else - do not send it in a code block, quotes, or anything else, just the pure text CONTAINING ONLY THE COMMAND. If possible, return a one-line bash command or chain many commands together. Return ONLY the command ready to run in the terminal. The command should do the following:"

CHATGPT_CYAN_LABEL="\033[36mchatgpt \033[0m"
PROCESSING_LABEL="\n\033[90mProcessing... \033[0m\033[0K\r"
OVERWRITE_PROCESSING_LINE="             \033[0K\r"


if [[ -z "$OPENAI_KEY" ]]; then
	echo "You need to set your OPENAI_KEY to use this script"
	echo "You can set it temporarily by running this on your terminal: export OPENAI_KEY=YOUR_KEY_HERE"
	exit 1
fi

usage() {
	cat <<EOF
A simple, lightweight shell script to use OpenAI's Language Models and DALL-E from the terminal without installing Python or Node.js. Open Source and written in 100% Shell (Bash) 

https://github.com/0xacx/chatGPT-shell-cli/

By default the script uses the "gpt-3.5-turbo" model. It will upgrade to "gpt-4" when the API is accessible to anyone.

Commands:
  image: - To generate images, start a prompt with image: If you are using iTerm, you can view the image directly in the terminal. Otherwise the script will ask to open the image in your browser.
  history - To view your chat history
  models - To get a list of the models available at OpenAI API
  model: - To view all the information on a specific model, start a prompt with model: and the model id as it appears in the list of models. For example: "model:text-babbage:001" will get you all the fields for text-babbage:001 model
  command: - To get a command with the specified functionality and run it, just type "command:" and explain what you want to achieve. The script will always ask you if you want to execute the command. i.e. 
  "command: show me all files in this directory that have more than 150 lines of code" 
  *If a command modifies your file system or dowloads external files the script will show a warning before executing.

Options:
  -i, --init-prompt - Provide initial chat prompt to use in context
  --init-prompt-from-file - Provide initial prompt from file
  -p, --prompt - Provide prompt instead of starting chat
  --prompt-from-file - Provide prompt from file
  -t, --temperature - Temperature
  --max-tokens - Max number of tokens
  -m, --model - Model
  -s, --size - Image size. (The sizes that are accepted by the OpenAI API are 256x256, 512x512, 1024x1024)
  -c, --chat-context - For models that do not support chat context by default (all models except gpt-3.5-turbo and gpt-4), you can enable chat context, for the model to remember your previous questions and its previous answers. It also makes models aware of todays date and what data it was trained on.

EOF
}

# error handling function
# $1 should be the response body
handle_error() {
	if echo "$1" | jq -e '.error' >/dev/null; then
		echo -e "Your request to Open AI API failed: \033[0;31m$(echo $1 | jq -r '.error.type')\033[0m"
		echo $1 | jq -r '.error.message'
		exit 1
	fi
}

# request to OpenAI API completions endpoint function
# $1 should be the request prompt
request_to_completions() {
	request_prompt="$1"

	response=$(curl https://api.openai.com/v1/completions \
		-sS \
		-H 'Content-Type: application/json' \
		-H "Authorization: Bearer $OPENAI_KEY" \
		-d '{
  			"model": "'"$MODEL"'",
  			"prompt": "'"${request_prompt}"'",
  			"max_tokens": '$MAX_TOKENS',
  			"temperature": '$TEMPERATURE'
			}')
}

# request to OpenAI API image generations endpoint function
# $1 should be the prompt
request_to_image() {
	prompt="$1"
	image_response=$(curl https://api.openai.com/v1/images/generations \
		-sS \
		-H 'Content-Type: application/json' \
		-H "Authorization: Bearer $OPENAI_KEY" \
		-d '{
    		"prompt": "'"${prompt#*image:}"'",
    		"n": 1,
    		"size": "'"$SIZE"'"
			}')
}

# request to OpenAPI API chat completion endpoint function
# $1 should be the message(s) formatted with role and content
request_to_chat() {
	message="$1"
	response=$(curl https://api.openai.com/v1/chat/completions \
		-sS \
		-H 'Content-Type: application/json' \
		-H "Authorization: Bearer $OPENAI_KEY" \
		-d '{
            "model": "'"$MODEL"'",
            "messages": [
                {"role": "system", "content": "'"$SYSTEM_PROMPT"'"},
                '"$message"'
                ],
            "max_tokens": '$MAX_TOKENS',
            "temperature": '$TEMPERATURE'
            }')
}

# build chat context before each request for /completions (all models except
# gpt turbo and gpt 4)
# $1 should be the chat context
# $2 should be the escaped prompt
build_chat_context() {
	chat_context="$1"
	escaped_prompt="$2"
	if [ -z "$chat_context" ]; then
		chat_context="$CHAT_INIT_PROMPT\nQ: $escaped_prompt"
	else
		chat_context="$chat_context\nQ: $escaped_prompt"
	fi
	request_prompt="${chat_context//$'\n'/\\n}"
}

# maintain chat context function for /completions (all models except
# gpt turbo and gpt 4)
# builds chat context from response,
# keeps chat context length under max token limit
# $1 should be the chat context
# $2 should be the response data (only the text)
maintain_chat_context() {
	chat_context="$1"
	response_data="$2"
	# add response to chat context as answer
	chat_context="$chat_context${chat_context:+\n}\nA: ${response_data//$'\n'/\\n}"
	# check prompt length, 1 word =~ 1.3 tokens
	# reserving 100 tokens for next user prompt
	while (($(echo "$chat_context" | wc -c) * 1, 3 > (MAX_TOKENS - 100))); do
		# remove first/oldest QnA from prompt
		chat_context=$(echo "$chat_context" | sed -n '/Q:/,$p' | tail -n +2)
		# add init prompt so it is always on top
		chat_context="$CHAT_INIT_PROMPT $chat_context"
	done
}

# build user chat message function for /chat/completions (gpt models)
# builds chat message before request,
# $1 should be the chat message
# $2 should be the escaped prompt
build_user_chat_message() {
	chat_message="$1"
	escaped_prompt="$2"
	if [ -z "$chat_message" ]; then
		chat_message="{\"role\": \"user\", \"content\": \"$escaped_prompt\"}"
	else
		chat_message="$chat_message, {\"role\": \"user\", \"content\": \"$escaped_prompt\"}"
	fi

	request_prompt="$chat_message"
}

# adds the assistant response to the message in (chatml) format
# for /chat/completions (gpt models)
# keeps messages length under max token limit
# $1 should be the chat message
# $2 should be the response data (only the text)
add_assistant_response_to_chat_message() {
	chat_message="$1"
	local local_response_data="$2"

	# replace new line characters from response with space
	local_response_data=$(echo "$local_response_data" | tr '\n' ' ')
	# add response to chat context as answer
	chat_message="$chat_message, {\"role\": \"assistant\", \"content\": \"$local_response_data\"}"

	# transform to json array to parse with jq
	chat_message_json="[ $chat_message ]"
	# check prompt length, 1 word =~ 1.3 tokens
	# reserving 100 tokens for next user prompt
	while (($(echo "$chat_message" | wc -c) * 1, 3 > (MAX_TOKENS - 100))); do
		# remove first/oldest QnA from prompt
		chat_message=$(echo "$chat_message_json" | jq -c '.[2:] | .[] | {role, content}')
	done
}

# parse command line arguments
while [[ "$#" -gt 0 ]]; do
	case $1 in
	-i | --init-prompt)
		CHAT_INIT_PROMPT="$2"
		SYSTEM_PROMPT="$2"
		CONTEXT=true
		shift
		shift
		;;
	--init-prompt-from-file)
		CHAT_INIT_PROMPT=$(cat "$2")
		SYSTEM_PROMPT=$(cat "$2")
		CONTEXT=true
		shift
		shift
		;;
	-p | --prompt)
		prompt="$2"
		shift
		shift
		;;
	--prompt-from-file)
		prompt=$(cat "$2")
		shift
		shift
		;;
	-t | --temperature)
		TEMPERATURE="$2"
		shift
		shift
		;;
	--max-tokens)
		MAX_TOKENS="$2"
		shift
		shift
		;;
	-m | --model)
		MODEL="$2"
		shift
		shift
		;;
	-s | --size)
		SIZE="$2"
		shift
		shift
		;;
	-c | --chat-context)
		CONTEXT=true
		shift
		;;
	-h | --help)
		usage
		exit 0
		;;
	*)
		echo "Unknown parameter: $1"
		exit 1
		;;
	esac
done

# set defaults
TEMPERATURE=${TEMPERATURE:-0.7}
MAX_TOKENS=${MAX_TOKENS:-1024}
MODEL=${MODEL:-gpt-3.5-turbo}
SIZE=${SIZE:-512x512}
CONTEXT=${CONTEXT:-false}

# create history file
if [ ! -f ~/.chatgpt_history ]; then
	touch ~/.chatgpt_history
	chmod 600 ~/.chatgpt_history
fi

running=true
# check input source and determine run mode

# prompt from argument, run on pipe mode (run once, no chat)
if [ -n "$prompt" ]; then
	pipe_mode_prompt=${prompt}
# if input file_descriptor is a terminal, run on chat mode
elif [ -t 0 ]; then
	echo -e "Welcome to chatgpt. You can quit with '\033[36mexit\033[0m' or '\033[36mq\033[0m'."
# prompt from pipe or redirected stdin, run on pipe mode
else
	pipe_mode_prompt+=$(cat -)
fi

while $running; do

	if [ -z "$pipe_mode_prompt" ]; then
		echo -e "\nEnter a prompt:"
		read -e prompt
		if [ "$prompt" != "exit" ] && [ "$prompt" != "q" ]; then
			echo -ne $PROCESSING_LABEL
		fi
	else
		# set vars for pipe mode
		prompt=${pipe_mode_prompt}
		running=false
		CHATGPT_CYAN_LABEL=""
	fi

	if [ "$prompt" == "exit" ] || [ "$prompt" == "q" ]; then
		running=false
	elif [[ "$prompt" =~ ^image: ]]; then
		request_to_image "$prompt"
		handle_error "$image_response"
		image_url=$(echo $image_response | jq -r '.data[0].url')
		echo -e "$OVERWRITE_PROCESSING_LINE"
		echo -e "${CHATGPT_CYAN_LABEL}Your image was created. \n\nLink: ${image_url}\n"

		if [[ "$TERM_PROGRAM" == "iTerm.app" ]]; then
			curl -sS $image_url -o temp_image.png
			imgcat temp_image.png
			rm temp_image.png
		elif [[ "$TERM" == "xterm-kitty" ]]; then
			curl -sS $image_url -o temp_image.png
			kitty +kitten icat temp_image.png
			rm temp_image.png
		else
			echo "Would you like to open it? (Yes/No)"
			read -e answer
			if [ "$answer" == "Yes" ] || [ "$answer" == "yes" ] || [ "$answer" == "y" ] || [ "$answer" == "Y" ] || [ "$answer" == "ok" ]; then
				open "${image_url}"
			fi
		fi
	elif [[ "$prompt" == "history" ]]; then
		echo -e "\n$(cat ~/.chatgpt_history)"
	elif [[ "$prompt" == "models" ]]; then
		models_response=$(curl https://api.openai.com/v1/models \
			-sS \
			-H "Authorization: Bearer $OPENAI_KEY")
		handle_error "$models_response"
		models_data=$(echo $models_response | jq -r -C '.data[] | {id, owned_by, created}')
		echo -e "$OVERWRITE_PROCESSING_LINE"
		echo -e "${CHATGPT_CYAN_LABEL}This is a list of models currently available at OpenAI API:\n ${models_data}"
	elif [[ "$prompt" =~ ^model: ]]; then
		models_response=$(curl https://api.openai.com/v1/models \
			-sS \
			-H "Authorization: Bearer $OPENAI_KEY")
		handle_error "$models_response"
		model_data=$(echo $models_response | jq -r -C '.data[] | select(.id=="'"${prompt#*model:}"'")')
		echo -e "$OVERWRITE_PROCESSING_LINE"
		echo -e "${CHATGPT_CYAN_LABEL}Complete details for model: ${prompt#*model:}\n ${model_data}"
	elif [[ "$prompt" =~ ^command: ]]; then
		# escape quotation marks
		escaped_prompt=$(echo "$prompt" | sed 's/"/\\"/g')
		# escape new lines
		if [[ "$prompt" =~ ^command: ]]; then
			escaped_prompt=${prompt#command:}
			request_prompt=$COMMAND_GENERATION_PROMPT${escaped_prompt//$'\n'/' '}
		fi
		build_user_chat_message "$chat_message" "$request_prompt"
		request_to_chat "$request_prompt"
		handle_error "$response"
		response_data=$(echo $response | jq -r '.choices[].message.content')

		if [[ "$prompt" =~ ^command: ]]; then
			echo -e "$OVERWRITE_PROCESSING_LINE"
			echo -e "${CHATGPT_CYAN_LABEL} ${response_data}" | fold -s -w $COLUMNS
			dangerous_commands=("rm" ">" "mv" "mkfs" ":(){:|:&};" "dd" "chmod" "wget" "curl")

			for dangerous_command in "${dangerous_commands[@]}"; do
				if [[ "$response_data" == *"$dangerous_command"* ]]; then
					echo "Warning! This command can change your file system or download external scripts & data. Please do not execute code that you don't understand completely."
				fi
			done
			echo "Would you like to execute it? (Yes/No)"
			read run_answer
			if [ "$run_answer" == "Yes" ] || [ "$run_answer" == "yes" ] || [ "$run_answer" == "y" ] || [ "$run_answer" == "Y" ]; then
				echo -e "\nExecuting command: $response_data\n"
				eval $response_data
			fi
		fi
		escaped_response_data=$(echo "$response_data" | sed 's/"/\\"/g')
		add_assistant_response_to_chat_message "$chat_message" "$escaped_response_data"

		timestamp=$(date +"%d/%m/%Y %H:%M")
		echo -e "$timestamp $prompt \n$response_data \n" >>~/.chatgpt_history

	elif [[ "$MODEL" =~ ^gpt- ]]; then
		# escape quotation marks
		escaped_prompt=$(echo "$prompt" | sed 's/"/\\"/g')
		# escape new lines
		request_prompt=${escaped_prompt//$'\n'/' '}

		build_user_chat_message "$chat_message" "$request_prompt"
		request_to_chat "$request_prompt"
		handle_error "$response"
		response_data=$(echo "$response" | jq -r '.choices[].message.content')

		echo -e "$OVERWRITE_PROCESSING_LINE"
		# if glow installed, print parsed markdown
		if command -v glow &>/dev/null; then
			echo -e "${CHATGPT_CYAN_LABEL}"
			echo "${response_data}" | glow -
			#echo -e "${formatted_text}" 
		else
			echo -e "${CHATGPT_CYAN_LABEL}${response_data}" | fold -s -w $COLUMNS
		fi
		escaped_response_data=$(echo "$response_data" | sed 's/"/\\"/g')
		add_assistant_response_to_chat_message "$chat_message" "$escaped_response_data"

		timestamp=$(date +"%d/%m/%Y %H:%M")
		echo -e "$timestamp $prompt \n$response_data \n" >>~/.chatgpt_history
	else
		# escape quotation marks
		escaped_prompt=$(echo "$prompt" | sed 's/"/\\"/g')
		# escape new lines
		request_prompt=${escaped_prompt//$'\n'/' '}

		if [ "$CONTEXT" = true ]; then
			build_chat_context "$chat_context" "$escaped_prompt"
		fi

		request_to_completions "$request_prompt"
		handle_error "$response"
		response_data=$(echo "$response" | jq -r '.choices[].text')

		echo -e "$OVERWRITE_PROCESSING_LINE"
		# if glow installed, print parsed markdown
		if command -v glow &>/dev/null; then
			echo -e "${CHATGPT_CYAN_LABEL}"
			echo "${response_data}" | glow -
		else
		# else remove empty lines and print
			formatted_text=$(echo "${response_data}" | sed '1,2d; s/^A://g')
			echo -e "${CHATGPT_CYAN_LABEL}${formatted_text}" | fold -s -w $COLUMNS
		fi

		if [ "$CONTEXT" = true ]; then
			escaped_response_data=$(echo "$response_data" | sed 's/"/\\"/g')
			maintain_chat_context "$chat_context" "$escaped_response_data"
		fi

		timestamp=$(date +"%d/%m/%Y %H:%M")
		echo -e "$timestamp $prompt \n$response_data \n" >>~/.chatgpt_history
	fi
done