aljaz/Auto-GPT
1
0
Fork 0
mirror of https://github.com/aljazceru/Auto-GPT.git synced 2025-12-18 14:34:23 +01:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into add_website_memory

Browse Source
This commit is contained in:
Maiko Bossuyt
2023-04-14 00:45:41 +02:00
committed by GitHub
parent 1b49c1d37a a3024ca80d
commit c4a45eb406
37 changed files with 789 additions and 270 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
scripts/agent_manager.py
View File

@@ -6,6 +6,7 @@ agents = {} # key, (task, full_message_history, model)
# Create new GPT agent
# TODO: Centralise use of create_chat_completion() to globally enforce token limit
def create_agent(task, prompt, model):
"""Create a new agent and return its key"""
global next_key

7
scripts/ai_config.py
View File

@@ -1,6 +1,7 @@
import yaml
import data
import os
from prompt import get_prompt
class AIConfig:
"""
@@ -46,7 +47,7 @@ class AIConfig:
"""
try:
with open(config_file) as file:
with open(config_file, encoding='utf-8') as file:
config_params = yaml.load(file, Loader=yaml.FullLoader)
except FileNotFoundError:
config_params = {}
@@ -90,5 +91,5 @@ class AIConfig:
for i, goal in enumerate(self.ai_goals):
full_prompt += f"{i+1}. {goal}\n"
full_prompt += f"\n\n{data.load_prompt()}"
full_prompt += f"\n\n{get_prompt()}"
return full_prompt

3
scripts/ai_functions.py
View File

@@ -1,8 +1,7 @@
from typing import List, Optional
from typing import List
import json
from config import Config
from call_ai_function import call_ai_function
from json_parser import fix_and_parse_json
cfg = Config()

2
scripts/call_ai_function.py
View File

@@ -3,6 +3,8 @@ from config import Config
cfg = Config()
from llm_utils import create_chat_completion
# This is a magic function that can do anything with no-code. See
# https://github.com/Torantulino/AI-Functions for more info.
def call_ai_function(function, args, description, model=None):

3
scripts/chat.py
View File

@@ -9,6 +9,7 @@ import logging
cfg = Config()
def create_chat_message(role, content):
"""
Create a chat message with the given role and content.
@@ -69,7 +70,7 @@ def chat_with_ai(
logger.debug(f"Token limit: {token_limit}")
send_token_limit = token_limit - 1000
relevant_memory = permanent_memory.get_relevant(str(full_message_history[-9:]), 10)
relevant_memory = '' if len(full_message_history) ==0 else permanent_memory.get_relevant(str(full_message_history[-9:]), 10)
logger.debug(f'Memory Stats: {permanent_memory.get_stats()}')

3
scripts/commands.py
View File

@@ -24,6 +24,7 @@ def is_valid_int(value):
except ValueError:
return False
def get_command(response):
"""Parse the response and return the command name and arguments"""
try:
@@ -135,6 +136,7 @@ def google_search(query, num_results=8):
return json.dumps(search_results, ensure_ascii=False, indent=4)
def google_official_search(query, num_results=8):
"""Return the results of a google search using the official Google API"""
from googleapiclient.discovery import build
@@ -171,6 +173,7 @@ def google_official_search(query, num_results=8):
# Return the list of search result URLs
return search_results_links
def browse_website(url, question):
"""Browse a website and return the summary and links"""
summary = get_text_summary(url, question)

16
scripts/config.py
View File

@@ -36,6 +36,7 @@ class Config(metaclass=Singleton):
"""Initialize the Config class"""
self.debug_mode = False
self.continuous_mode = False
self.continuous_limit = 0
self.speak_mode = False
self.fast_llm_model = os.getenv("FAST_LLM_MODEL", "gpt-3.5-turbo")
@@ -46,14 +47,13 @@ class Config(metaclass=Singleton):
self.browse_summary_max_token = int(os.getenv("BROWSE_SUMMARY_MAX_TOKEN", 300))
self.openai_api_key = os.getenv("OPENAI_API_KEY")
self.temperature = int(os.getenv("TEMPERATURE", "1"))
self.use_azure = False
self.temperature = float(os.getenv("TEMPERATURE", "1"))
self.use_azure = os.getenv("USE_AZURE") == 'True'
self.execute_local_commands = os.getenv('EXECUTE_LOCAL_COMMANDS', 'False') == 'True'
if self.use_azure:
self.load_azure_config()
openai.api_type = "azure"
openai.api_type = self.openai_api_type
openai.api_base = self.openai_api_base
openai.api_version = self.openai_api_version
@@ -76,7 +76,6 @@ class Config(metaclass=Singleton):
# User agent headers to use when browsing web
# Some websites might just completely deny request with an error code if no user agent was found.
self.user_agent = os.getenv("USER_AGENT", "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36")
self.redis_host = os.getenv("REDIS_HOST", "localhost")
self.redis_port = os.getenv("REDIS_PORT", "6379")
self.redis_password = os.getenv("REDIS_PASSWORD", "")
@@ -124,14 +123,19 @@ class Config(metaclass=Singleton):
config_params = yaml.load(file, Loader=yaml.FullLoader)
except FileNotFoundError:
config_params = {}
self.openai_api_base = config_params.get("azure_api_base", "")
self.openai_api_version = config_params.get("azure_api_version", "")
self.openai_api_type = os.getenv("OPENAI_API_TYPE", config_params.get("azure_api_type", "azure"))
self.openai_api_base = os.getenv("OPENAI_AZURE_API_BASE", config_params.get("azure_api_base", ""))
self.openai_api_version = os.getenv("OPENAI_AZURE_API_VERSION", config_params.get("azure_api_version", ""))
self.azure_model_to_deployment_id_map = config_params.get("azure_model_map", [])
def set_continuous_mode(self, value: bool):
"""Set the continuous mode value."""
self.continuous_mode = value
def set_continuous_limit(self, value: int):
"""Set the continuous limit value."""
self.continuous_limit = value
def set_speak_mode(self, value: bool):
"""Set the speak mode value."""
self.speak_mode = value

18
scripts/data.py
View File

@@ -1,18 +0,0 @@
import os
from pathlib import Path
def load_prompt():
"""Load the prompt from data/prompt.txt"""
try:
# get directory of this file:
file_dir = Path(__file__).parent
prompt_file_path = file_dir / "data" / "prompt.txt"
# Load the prompt from data/prompt.txt
with open(prompt_file_path, "r") as prompt_file:
prompt = prompt_file.read()
return prompt
except FileNotFoundError:
print("Error: Prompt file not found", flush=True)
return ""

64
scripts/data/prompt.txt
View File

@@ -1,64 +0,0 @@
CONSTRAINTS:
1. ~4000 word limit for short term memory. Your short term memory is short, so immediately save important information to files.
2. If you are unsure how you previously did something or want to recall past events, thinking about similar events will help you remember.
3. No user assistance
4. Exclusively use the commands listed in double quotes e.g. "command name"
COMMANDS:
1. Google Search: "google", args: "input": "<search>"
5. Browse Website: "browse_website", args: "url": "<url>", "question": "<what_you_want_to_find_on_website>"
6. Start GPT Agent: "start_agent", args: "name": "<name>", "task": "<short_task_desc>", "prompt": "<prompt>"
7. Message GPT Agent: "message_agent", args: "key": "<key>", "message": "<message>"
8. List GPT Agents: "list_agents", args: ""
9. Delete GPT Agent: "delete_agent", args: "key": "<key>"
10. Write to file: "write_to_file", args: "file": "<file>", "text": "<text>"
11. Read file: "read_file", args: "file": "<file>"
12. Append to file: "append_to_file", args: "file": "<file>", "text": "<text>"
13. Delete file: "delete_file", args: "file": "<file>"
14. Search Files: "search_files", args: "directory": "<directory>"
15. Evaluate Code: "evaluate_code", args: "code": "<full_code_string>"
16. Get Improved Code: "improve_code", args: "suggestions": "<list_of_suggestions>", "code": "<full_code_string>"
17. Write Tests: "write_tests", args: "code": "<full_code_string>", "focus": "<list_of_focus_areas>"
18. Execute Python File: "execute_python_file", args: "file": "<file>"
19. Execute Shell Command, non-interactive commands only: "execute_shell", args: "command_line": "<command_line>".
20. Task Complete (Shutdown): "task_complete", args: "reason": "<reason>"
21. Generate Image: "generate_image", args: "prompt": "<prompt>"
22. Do Nothing: "do_nothing", args: ""
RESOURCES:
1. Internet access for searches and information gathering.
2. Long Term memory management.
3. GPT-3.5 powered Agents for delegation of simple tasks.
4. File output.
PERFORMANCE EVALUATION:
1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.
2. Constructively self-criticize your big-picture behavior constantly.
3. Reflect on past decisions and strategies to refine your approach.
4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.
You should only respond in JSON format as described below
RESPONSE FORMAT:
{
"thoughts":
{
"text": "thought",
"reasoning": "reasoning",
"plan": "- short bulleted\n- list that conveys\n- long-term plan",
"criticism": "constructive self-criticism",
"speak": "thoughts summary to say to user"
},
"command": {
"name": "command name",
"args":{
"arg name": "value"
}
}
}
Ensure the response can be parsed by Python json.loads

1
scripts/execute_code.py
View File

@@ -67,6 +67,7 @@ def execute_python_file(file):
except Exception as e:
return f"Error: {str(e)}"
def execute_shell(command_line):
current_dir = os.getcwd()

3
scripts/file_operations.py
View File

@@ -38,7 +38,7 @@ def write_to_file(filename, text):
directory = os.path.dirname(filepath)
if not os.path.exists(directory):
os.makedirs(directory)
with open(filepath, "w") as f:
with open(filepath, "w", encoding='utf-8') as f:
f.write(text)
return "File written to successfully."
except Exception as e:
@@ -65,6 +65,7 @@ def delete_file(filename):
except Exception as e:
return "Error: " + str(e)
def search_files(directory):
found_files = []

1
scripts/image_gen.py
View File

@@ -11,6 +11,7 @@ cfg = Config()
working_directory = "auto_gpt_workspace"
def generate_image(prompt):
filename = str(uuid.uuid4()) + ".jpg"

56
scripts/llm_utils.py
View File

@@ -1,26 +1,52 @@
import time
import openai
from colorama import Fore
from config import Config
cfg = Config()
openai.api_key = cfg.openai_api_key
# Overly simple abstraction until we create something better
# simple retry mechanism when getting a rate error or a bad gateway
def create_chat_completion(messages, model=None, temperature=cfg.temperature, max_tokens=None)->str:
"""Create a chat completion using the OpenAI API"""
if cfg.use_azure:
response = openai.ChatCompletion.create(
deployment_id=cfg.get_azure_deployment_id_for_model(model),
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens
)
else:
response = openai.ChatCompletion.create(
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens
)
response = None
num_retries = 5
for attempt in range(num_retries):
try:
if cfg.use_azure:
response = openai.ChatCompletion.create(
deployment_id=cfg.get_azure_deployment_id_for_model(model),
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens
)
else:
response = openai.ChatCompletion.create(
model=model,
messages=messages,
temperature=temperature,
max_tokens=max_tokens
)
break
except openai.error.RateLimitError:
if cfg.debug_mode:
print(Fore.RED + "Error: ", "API Rate Limit Reached. Waiting 20 seconds..." + Fore.RESET)
time.sleep(20)
except openai.error.APIError as e:
if e.http_status == 502:
if cfg.debug_mode:
print(Fore.RED + "Error: ", "API Bad gateway. Waiting 20 seconds..." + Fore.RESET)
time.sleep(20)
else:
raise
if attempt == num_retries - 1:
raise
if response is None:
raise RuntimeError("Failed to get response after 5 retries")
return response.choices[0].message["content"]

15
scripts/logger.py
View File

@@ -124,6 +124,12 @@ class Logger(metaclass=Singleton):
self.logger.setLevel(level)
self.typing_logger.setLevel(level)
def double_check(self, additionalText=None):
if not additionalText:
additionalText = "Please ensure you've setup and configured everything correctly. Read https://github.com/Torantulino/Auto-GPT#readme to double check. You can also create a github issue or join the discord and ask there!"
self.typewriter_log("DOUBLE CHECK CONFIGURATION", Fore.YELLOW, additionalText)
'''
Output stream to console using simulated typing
@@ -151,6 +157,7 @@ class TypingConsoleHandler(logging.StreamHandler):
except Exception:
self.handleError(record)
class ConsoleHandler(logging.StreamHandler):
def emit(self, record):
msg = self.format(record)
@@ -160,11 +167,11 @@ class ConsoleHandler(logging.StreamHandler):
self.handleError(record)
'''
Allows to handle custom placeholders 'title_color' and 'message_no_color'.
To use this formatter, make sure to pass 'color', 'title' as log extras.
'''
class AutoGptFormatter(logging.Formatter):
"""
Allows to handle custom placeholders 'title_color' and 'message_no_color'.
To use this formatter, make sure to pass 'color', 'title' as log extras.
"""
def format(self, record: LogRecord) -> str:
if (hasattr(record, 'color')):
record.title_color = getattr(record, 'color') + getattr(record, 'title') + " " + Style.RESET_ALL

260
scripts/main.py
View File

@@ -3,7 +3,6 @@ import random
import commands as cmd
import utils
from memory import get_memory, get_supported_memory_backends
import data
import chat
from colorama import Fore, Style
from spinner import Spinner
@@ -17,19 +16,22 @@ import yaml
import argparse
from logger import logger
import logging
from prompt import get_prompt
cfg = Config()
def check_openai_api_key():
"""Check if the OpenAI API key is set in config.py or as an environment variable."""
if not cfg.openai_api_key:
print(
Fore.RED +
"Please set your OpenAI API key in config.py or as an environment variable."
"Please set your OpenAI API key in .env or as an environment variable."
)
print("You can get your key from https://beta.openai.com/account/api-keys")
exit(1)
def attempt_to_fix_json_by_finding_outermost_brackets(json_string):
if cfg.speak_mode and cfg.debug_mode:
speak.say_text("I have received an invalid JSON response from the OpenAI API. Trying to fix it now.")
@@ -58,6 +60,7 @@ def attempt_to_fix_json_by_finding_outermost_brackets(json_string):
return json_string
def print_assistant_thoughts(assistant_reply):
"""Prints the assistant's thoughts to the console"""
global ai_name
@@ -168,8 +171,8 @@ def load_variables(config_file="config.yaml"):
with open(config_file, "w") as file:
documents = yaml.dump(config, file)
prompt = data.load_prompt()
prompt_start = """Your decisions must always be made independently without seeking user assistance. Play to your strengths as a LLM and pursue simple strategies with no legal complications."""
prompt = get_prompt()
prompt_start = """Your decisions must always be made independently without seeking user assistance. Play to your strengths as an LLM and pursue simple strategies with no legal complications."""
# Construct full prompt
full_prompt = f"You are {ai_name}, {ai_role}\n{prompt_start}\n\nGOALS:\n\n"
@@ -262,6 +265,7 @@ def prompt_user():
config = AIConfig(ai_name, ai_role, ai_goals)
return config
def parse_arguments():
"""Parses the arguments passed to the script"""
global cfg
@@ -271,6 +275,7 @@ def parse_arguments():
parser = argparse.ArgumentParser(description='Process arguments.')
parser.add_argument('--continuous', action='store_true', help='Enable Continuous Mode')
parser.add_argument('--continuous-limit', '-l', type=int, dest="continuous_limit", help='Defines the number of times to run in continuous mode')
parser.add_argument('--speak', action='store_true', help='Enable Speak Mode')
parser.add_argument('--debug', action='store_true', help='Enable Debug Mode')
parser.add_argument('--gpt3only', action='store_true', help='Enable GPT3.5 Only Mode')
@@ -290,6 +295,17 @@ def parse_arguments():
"Continuous mode is not recommended. It is potentially dangerous and may cause your AI to run forever or carry out actions you would not usually authorise. Use at your own risk.")
cfg.set_continuous_mode(True)
if args.continuous_limit:
logger.typewriter_log(
"Continuous Limit: ",
Fore.GREEN,
f"{args.continuous_limit}")
cfg.set_continuous_limit(args.continuous_limit)
# Check if continuous limit is used without continuous mode
if args.continuous_limit and not args.continuous:
parser.error("--continuous-limit can only be used with --continuous")
if args.speak:
logger.typewriter_log("Speak Mode: ", Fore.GREEN, "ENABLED")
cfg.set_speak_mode(True)
@@ -310,124 +326,136 @@ def parse_arguments():
supported_memory = get_supported_memory_backends()
chosen = args.memory_type
if not chosen in supported_memory:
print_to_console("ONLY THE FOLLOWING MEMORY BACKENDS ARE SUPPORTED: ", Fore.RED, f'{supported_memory}')
print_to_console(f"Defaulting to: ", Fore.YELLOW, cfg.memory_backend)
logger.typewriter_log("ONLY THE FOLLOWING MEMORY BACKENDS ARE SUPPORTED: ", Fore.RED, f'{supported_memory}')
logger.typewriter_log(f"Defaulting to: ", Fore.YELLOW, cfg.memory_backend)
else:
cfg.memory_backend = chosen
# TODO: fill in llm values here
check_openai_api_key()
parse_arguments()
logger.set_level(logging.DEBUG if cfg.debug_mode else logging.INFO)
ai_name = ""
prompt = construct_prompt()
# print(prompt)
# Initialize variables
full_message_history = []
result = None
next_action_count = 0
# Make a constant:
user_input = "Determine which next command to use, and respond using the format specified above:"
# Initialize memory and make sure it is empty.
# this is particularly important for indexing and referencing pinecone memory
memory = get_memory(cfg, init=True)
print('Using memory of type: ' + memory.__class__.__name__)
# Interaction Loop
while True:
# Send message to AI, get response
with Spinner("Thinking... "):
assistant_reply = chat.chat_with_ai(
prompt,
user_input,
full_message_history,
memory,
cfg.fast_token_limit) # TODO: This hardcodes the model to use GPT3.5. Make this an argument
# Print Assistant thoughts
print_assistant_thoughts(assistant_reply)
# Get command name and arguments
try:
command_name, arguments = cmd.get_command(attempt_to_fix_json_by_finding_outermost_brackets(assistant_reply))
if cfg.speak_mode:
speak.say_text(f"I want to execute {command_name}")
except Exception as e:
logger.error("Error: \n", str(e))
if not cfg.continuous_mode and next_action_count == 0:
### GET USER AUTHORIZATION TO EXECUTE COMMAND ###
# Get key press: Prompt the user to press enter to continue or escape
# to exit
user_input = ""
logger.typewriter_log(
"NEXT ACTION: ",
Fore.CYAN,
f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}")
print(
f"Enter 'y' to authorise command, 'y -N' to run N continuous commands, 'n' to exit program, or enter feedback for {ai_name}...",
flush=True)
while True:
console_input = utils.clean_input(Fore.MAGENTA + "Input:" + Style.RESET_ALL)
if console_input.lower().rstrip() == "y":
user_input = "GENERATE NEXT COMMAND JSON"
break
elif console_input.lower().startswith("y -"):
try:
next_action_count = abs(int(console_input.split(" ")[1]))
user_input = "GENERATE NEXT COMMAND JSON"
except ValueError:
print("Invalid input format. Please enter 'y -n' where n is the number of continuous tasks.")
continue
break
elif console_input.lower() == "n":
user_input = "EXIT"
break
else:
user_input = console_input
command_name = "human_feedback"
break
if user_input == "GENERATE NEXT COMMAND JSON":
logger.typewriter_log(
"-=-=-=-=-=-=-= COMMAND AUTHORISED BY USER -=-=-=-=-=-=-=",
Fore.MAGENTA,
"")
elif user_input == "EXIT":
print("Exiting...", flush=True)
def main():
global ai_name, memory
# TODO: fill in llm values here
check_openai_api_key()
parse_arguments()
logger.set_level(logging.DEBUG if cfg.debug_mode else logging.INFO)
ai_name = ""
prompt = construct_prompt()
# print(prompt)
# Initialize variables
full_message_history = []
result = None
next_action_count = 0
# Make a constant:
user_input = "Determine which next command to use, and respond using the format specified above:"
# Initialize memory and make sure it is empty.
# this is particularly important for indexing and referencing pinecone memory
memory = get_memory(cfg, init=True)
print('Using memory of type: ' + memory.__class__.__name__)
# Interaction Loop
loop_count = 0
while True:
# Discontinue if continuous limit is reached
loop_count += 1
if cfg.continuous_mode and cfg.continuous_limit > 0 and loop_count > cfg.continuous_limit:
logger.typewriter_log("Continuous Limit Reached: ", Fore.YELLOW, f"{cfg.continuous_limit}")
break
else:
# Print command
logger.typewriter_log(
"NEXT ACTION: ",
Fore.CYAN,
f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}")
# Execute command
if command_name is not None and command_name.lower().startswith( "error" ):
result = f"Command {command_name} threw the following error: " + arguments
elif command_name == "human_feedback":
result = f"Human feedback: {user_input}"
else:
result = f"Command {command_name} returned: {cmd.execute_command(command_name, arguments)}"
if next_action_count > 0:
next_action_count -= 1
# Send message to AI, get response
with Spinner("Thinking... "):
assistant_reply = chat.chat_with_ai(
prompt,
user_input,
full_message_history,
memory,
cfg.fast_token_limit) # TODO: This hardcodes the model to use GPT3.5. Make this an argument
memory_to_add = f"Assistant Reply: {assistant_reply} " \
f"\nResult: {result} " \
f"\nHuman Feedback: {user_input} "
# Print Assistant thoughts
print_assistant_thoughts(assistant_reply)
memory.add(memory_to_add)
# Get command name and arguments
try:
command_name, arguments = cmd.get_command(
attempt_to_fix_json_by_finding_outermost_brackets(assistant_reply))
if cfg.speak_mode:
speak.say_text(f"I want to execute {command_name}")
except Exception as e:
logger.error("Error: \n", str(e))
# Check if there's a result from the command append it to the message
# history
if result is not None:
full_message_history.append(chat.create_chat_message("system", result))
logger.typewriter_log("SYSTEM: ", Fore.YELLOW, result)
else:
full_message_history.append(
chat.create_chat_message(
"system", "Unable to execute command"))
logger.typewriter_log("SYSTEM: ", Fore.YELLOW, "Unable to execute command")
if not cfg.continuous_mode and next_action_count == 0:
### GET USER AUTHORIZATION TO EXECUTE COMMAND ###
# Get key press: Prompt the user to press enter to continue or escape
# to exit
user_input = ""
logger.typewriter_log(
"NEXT ACTION: ",
Fore.CYAN,
f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}")
print(
f"Enter 'y' to authorise command, 'y -N' to run N continuous commands, 'n' to exit program, or enter feedback for {ai_name}...",
flush=True)
while True:
console_input = utils.clean_input(Fore.MAGENTA + "Input:" + Style.RESET_ALL)
if console_input.lower().rstrip() == "y":
user_input = "GENERATE NEXT COMMAND JSON"
break
elif console_input.lower().startswith("y -"):
try:
next_action_count = abs(int(console_input.split(" ")[1]))
user_input = "GENERATE NEXT COMMAND JSON"
except ValueError:
print("Invalid input format. Please enter 'y -n' where n is the number of continuous tasks.")
continue
break
elif console_input.lower() == "n":
user_input = "EXIT"
break
else:
user_input = console_input
command_name = "human_feedback"
break
if user_input == "GENERATE NEXT COMMAND JSON":
logger.typewriter_log(
"-=-=-=-=-=-=-= COMMAND AUTHORISED BY USER -=-=-=-=-=-=-=",
Fore.MAGENTA,
"")
elif user_input == "EXIT":
print("Exiting...", flush=True)
break
else:
# Print command
logger.typewriter_log(
"NEXT ACTION: ",
Fore.CYAN,
f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}")
# Execute command
if command_name is not None and command_name.lower().startswith("error"):
result = f"Command {command_name} threw the following error: " + arguments
elif command_name == "human_feedback":
result = f"Human feedback: {user_input}"
else:
result = f"Command {command_name} returned: {cmd.execute_command(command_name, arguments)}"
if next_action_count > 0:
next_action_count -= 1
memory_to_add = f"Assistant Reply: {assistant_reply} " \
f"\nResult: {result} " \
f"\nHuman Feedback: {user_input} "
memory.add(memory_to_add)
# Check if there's a result from the command append it to the message
# history
if result is not None:
full_message_history.append(chat.create_chat_message("system", result))
logger.typewriter_log("SYSTEM: ", Fore.YELLOW, result)
else:
full_message_history.append(
chat.create_chat_message(
"system", "Unable to execute command"))
logger.typewriter_log("SYSTEM: ", Fore.YELLOW, "Unable to execute command")
if __name__ == "__main__":
main()

6
scripts/memory/__init__.py
View File

@@ -1,4 +1,5 @@
from memory.local import LocalCache
from memory.no_memory import NoMemory
# List of supported memory backends
# Add a backend to this list if the import attempt is successful
@@ -18,6 +19,7 @@ except ImportError:
print("Pinecone not installed. Skipping import.")
PineconeMemory = None
def get_memory(cfg, init=False):
memory = None
if cfg.memory_backend == "pinecone":
@@ -34,6 +36,8 @@ def get_memory(cfg, init=False):
" use Redis as a memory backend.")
else:
memory = RedisMemory(cfg)
elif cfg.memory_backend == "no_memory":
memory = NoMemory(cfg)
if memory is None:
memory = LocalCache(cfg)
@@ -41,6 +45,7 @@ def get_memory(cfg, init=False):
memory.clear()
return memory
def get_supported_memory_backends():
return supported_memory
@@ -50,4 +55,5 @@ __all__ = [
"LocalCache",
"RedisMemory",
"PineconeMemory",
"NoMemory"
]

2
scripts/memory/base.py
View File

@@ -2,10 +2,10 @@
import abc
from config import AbstractSingleton, Config
import openai
cfg = Config()
cfg = Config()
def get_ada_embedding(text):
text = text.replace("\n", " ")
if cfg.use_azure:

66
scripts/memory/no_memory.py Normal file
View File

@@ -0,0 +1,66 @@
from typing import Optional, List, Any
from memory.base import MemoryProviderSingleton
class NoMemory(MemoryProviderSingleton):
def __init__(self, cfg):
"""
Initializes the NoMemory provider.
Args:
cfg: The config object.
Returns: None
"""
pass
def add(self, data: str) -> str:
"""
Adds a data point to the memory. No action is taken in NoMemory.
Args:
data: The data to add.
Returns: An empty string.
"""
return ""
def get(self, data: str) -> Optional[List[Any]]:
"""
Gets the data from the memory that is most relevant to the given data.
NoMemory always returns None.
Args:
data: The data to compare to.
Returns: None
"""
return None
def clear(self) -> str:
"""
Clears the memory. No action is taken in NoMemory.
Returns: An empty string.
"""
return ""
def get_relevant(self, data: str, num_relevant: int = 5) -> Optional[List[Any]]:
"""
Returns all the data in the memory that is relevant to the given data.
NoMemory always returns None.
Args:
data: The data to compare to.
num_relevant: The number of relevant data to return.
Returns: None
"""
return None
def get_stats(self):
"""
Returns: An empty dictionary as there are no stats in NoMemory.
"""
return {}

11
scripts/memory/pinecone.py
View File

@@ -2,6 +2,8 @@
import pinecone
from memory.base import MemoryProviderSingleton, get_ada_embedding
from logger import logger
from colorama import Fore, Style
class PineconeMemory(MemoryProviderSingleton):
@@ -17,6 +19,15 @@ class PineconeMemory(MemoryProviderSingleton):
# for now this works.
# we'll need a more complicated and robust system if we want to start with memory.
self.vec_num = 0
try:
pinecone.whoami()
except Exception as e:
logger.typewriter_log("FAILED TO CONNECT TO PINECONE", Fore.RED, Style.BRIGHT + str(e) + Style.RESET_ALL)
logger.double_check("Please ensure you have setup and configured Pinecone properly for use. " +
f"You can check out {Fore.CYAN + Style.BRIGHT}https://github.com/Torantulino/Auto-GPT#-pinecone-api-key-setup{Style.RESET_ALL} to ensure you've set up everything correctly.")
exit(1)
if table_name not in pinecone.list_indexes():
pinecone.create_index(table_name, dimension=dimension, metric=metric, pod_type=pod_type)
self.index = pinecone.Index(table_name)

12
scripts/memory/redismem.py
View File

@@ -7,6 +7,8 @@ from redis.commands.search.indexDefinition import IndexDefinition, IndexType
import numpy as np
from memory.base import MemoryProviderSingleton, get_ada_embedding
from logger import logger
from colorama import Fore, Style
SCHEMA = [
@@ -44,6 +46,16 @@ class RedisMemory(MemoryProviderSingleton):
db=0 # Cannot be changed
)
self.cfg = cfg
# Check redis connection
try:
self.redis.ping()
except redis.ConnectionError as e:
logger.typewriter_log("FAILED TO CONNECT TO REDIS", Fore.RED, Style.BRIGHT + str(e) + Style.RESET_ALL)
logger.double_check("Please ensure you have setup and configured Redis properly for use. " +
f"You can check out {Fore.CYAN + Style.BRIGHT}https://github.com/Torantulino/Auto-GPT#redis-setup{Style.RESET_ALL} to ensure you've set up everything correctly.")
exit(1)
if cfg.wipe_redis_on_start:
self.redis.flushall()
try:

63
scripts/prompt.py Normal file
View File

@@ -0,0 +1,63 @@
from promptgenerator import PromptGenerator
def get_prompt():
"""
This function generates a prompt string that includes various constraints, commands, resources, and performance evaluations.
Returns:
str: The generated prompt string.
"""
# Initialize the PromptGenerator object
prompt_generator = PromptGenerator()
# Add constraints to the PromptGenerator object
prompt_generator.add_constraint("~4000 word limit for short term memory. Your short term memory is short, so immediately save important information to files.")
prompt_generator.add_constraint("If you are unsure how you previously did something or want to recall past events, thinking about similar events will help you remember.")
prompt_generator.add_constraint("No user assistance")
prompt_generator.add_constraint('Exclusively use the commands listed in double quotes e.g. "command name"')
# Define the command list
commands = [
("Google Search", "google", {"input": "<search>"}),
("Browse Website", "browse_website", {"url": "<url>", "question": "<what_you_want_to_find_on_website>"}),
("Start GPT Agent", "start_agent", {"name": "<name>", "task": "<short_task_desc>", "prompt": "<prompt>"}),
("Message GPT Agent", "message_agent", {"key": "<key>", "message": "<message>"}),
("List GPT Agents", "list_agents", {}),
("Delete GPT Agent", "delete_agent", {"key": "<key>"}),
("Write to file", "write_to_file", {"file": "<file>", "text": "<text>"}),
("Read file", "read_file", {"file": "<file>"}),
("Append to file", "append_to_file", {"file": "<file>", "text": "<text>"}),
("Delete file", "delete_file", {"file": "<file>"}),
("Search Files", "search_files", {"directory": "<directory>"}),
("Evaluate Code", "evaluate_code", {"code": "<full_code_string>"}),
("Get Improved Code", "improve_code", {"suggestions": "<list_of_suggestions>", "code": "<full_code_string>"}),
("Write Tests", "write_tests", {"code": "<full_code_string>", "focus": "<list_of_focus_areas>"}),
("Execute Python File", "execute_python_file", {"file": "<file>"}),
("Execute Shell Command, non-interactive commands only", "execute_shell", { "command_line": "<command_line>"}),
("Task Complete (Shutdown)", "task_complete", {"reason": "<reason>"}),
("Generate Image", "generate_image", {"prompt": "<prompt>"}),
("Do Nothing", "do_nothing", {}),
]
# Add commands to the PromptGenerator object
for command_label, command_name, args in commands:
prompt_generator.add_command(command_label, command_name, args)
# Add resources to the PromptGenerator object
prompt_generator.add_resource("Internet access for searches and information gathering.")
prompt_generator.add_resource("Long Term memory management.")
prompt_generator.add_resource("GPT-3.5 powered Agents for delegation of simple tasks.")
prompt_generator.add_resource("File output.")
# Add performance evaluations to the PromptGenerator object
prompt_generator.add_performance_evaluation("Continuously review and analyze your actions to ensure you are performing to the best of your abilities.")
prompt_generator.add_performance_evaluation("Constructively self-criticize your big-picture behavior constantly.")
prompt_generator.add_performance_evaluation("Reflect on past decisions and strategies to refine your approach.")
prompt_generator.add_performance_evaluation("Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.")
# Generate the prompt string
prompt_string = prompt_generator.generate_prompt_string()
return prompt_string

129
scripts/promptgenerator.py Normal file
View File

@@ -0,0 +1,129 @@
import json
class PromptGenerator:
"""
A class for generating custom prompt strings based on constraints, commands, resources, and performance evaluations.
"""
def __init__(self):
"""
Initialize the PromptGenerator object with empty lists of constraints, commands, resources, and performance evaluations.
"""
self.constraints = []
self.commands = []
self.resources = []
self.performance_evaluation = []
self.response_format = {
"thoughts": {
"text": "thought",
"reasoning": "reasoning",
"plan": "- short bulleted\n- list that conveys\n- long-term plan",
"criticism": "constructive self-criticism",
"speak": "thoughts summary to say to user"
},
"command": {
"name": "command name",
"args": {
"arg name": "value"
}
}
}
def add_constraint(self, constraint):
"""
Add a constraint to the constraints list.
Args:
constraint (str): The constraint to be added.
"""
self.constraints.append(constraint)
def add_command(self, command_label, command_name, args=None):
"""
Add a command to the commands list with a label, name, and optional arguments.
Args:
command_label (str): The label of the command.
command_name (str): The name of the command.
args (dict, optional): A dictionary containing argument names and their values. Defaults to None.
"""
if args is None:
args = {}
command_args = {arg_key: arg_value for arg_key,
arg_value in args.items()}
command = {
"label": command_label,
"name": command_name,
"args": command_args,
}
self.commands.append(command)
def _generate_command_string(self, command):
"""
Generate a formatted string representation of a command.
Args:
command (dict): A dictionary containing command information.
Returns:
str: The formatted command string.
"""
args_string = ', '.join(
f'"{key}": "{value}"' for key, value in command['args'].items())
return f'{command["label"]}: "{command["name"]}", args: {args_string}'
def add_resource(self, resource):
"""
Add a resource to the resources list.
Args:
resource (str): The resource to be added.
"""
self.resources.append(resource)
def add_performance_evaluation(self, evaluation):
"""
Add a performance evaluation item to the performance_evaluation list.
Args:
evaluation (str): The evaluation item to be added.
"""
self.performance_evaluation.append(evaluation)
def _generate_numbered_list(self, items, item_type='list'):
"""
Generate a numbered list from given items based on the item_type.
Args:
items (list): A list of items to be numbered.
item_type (str, optional): The type of items in the list. Defaults to 'list'.
Returns:
str: The formatted numbered list.
"""
if item_type == 'command':
return "\n".join(f"{i+1}. {self._generate_command_string(item)}" for i, item in enumerate(items))
else:
return "\n".join(f"{i+1}. {item}" for i, item in enumerate(items))
def generate_prompt_string(self):
"""
Generate a prompt string based on the constraints, commands, resources, and performance evaluations.
Returns:
str: The generated prompt string.
"""
formatted_response_format = json.dumps(self.response_format, indent=4)
prompt_string = (
f"Constraints:\n{self._generate_numbered_list(self.constraints)}\n\n"
f"Commands:\n{self._generate_numbered_list(self.commands, item_type='command')}\n\n"
f"Resources:\n{self._generate_numbered_list(self.resources)}\n\n"
f"Performance Evaluation:\n{self._generate_numbered_list(self.performance_evaluation)}\n\n"
f"You should only respond in JSON format as described below \nResponse Format: \n{formatted_response_format} \nEnsure the response can be parsed by Python json.loads"
)
return prompt_string

4
scripts/speak.py
View File

@@ -31,6 +31,7 @@ tts_headers = {
mutex_lock = Lock() # Ensure only one sound is played at a time
queue_semaphore = Semaphore(1) # The amount of sounds to queue before blocking the main thread
def eleven_labs_speech(text, voice_index=0):
"""Speak text using elevenlabs.io's API"""
tts_url = "https://api.elevenlabs.io/v1/text-to-speech/{voice_id}".format(
@@ -51,6 +52,7 @@ def eleven_labs_speech(text, voice_index=0):
print("Response content:", response.content)
return False
def gtts_speech(text):
tts = gtts.gTTS(text)
with mutex_lock:
@@ -58,6 +60,7 @@ def gtts_speech(text):
playsound("speech.mp3", True)
os.remove("speech.mp3")
def macos_tts_speech(text, voice_index=0):
if voice_index == 0:
os.system(f'say "{text}"')
@@ -67,6 +70,7 @@ def macos_tts_speech(text, voice_index=0):
else:
os.system(f'say -v Samantha "{text}"')
def say_text(text, voice_index=0):
def speak():

2
scripts/token_counter.py
View File

@@ -1,6 +1,7 @@
import tiktoken
from typing import List, Dict
def count_message_tokens(messages : List[Dict[str, str]], model : str = "gpt-3.5-turbo-0301") -> int:
"""
Returns the number of tokens used by a list of messages.
@@ -41,6 +42,7 @@ def count_message_tokens(messages : List[Dict[str, str]], model : str = "gpt-3.5
num_tokens += 3 # every reply is primed with <|start|>assistant<|message|>
return num_tokens
def count_string_tokens(string: str, model_name: str) -> int:
"""
Returns the number of tokens in a text string.