alihan/microagents-llm
1
0
Fork 0
mirror of https://github.com/aymenfurter/microagents.git synced 2023-12-30 16:47:11 +03:00
Code Issues Packages Projects Releases Wiki Activity

wip

Browse Source
This commit is contained in:
Aymen
2023-12-13 12:45:46 +00:00
committed by GitHub
parent 03618074c0
commit 6bb6d28a94
4 changed files with 28 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

BIN
__pycache__/microagent.cpython-310.pyc
View File

Binary file not shown.

11
main.py
View File

@@ -91,12 +91,13 @@ class MicroAgentManager:
"Goal: Your purpose is to solve basic arithmetic problems. Generated Prompt: You are a proficient calculator. Here's a Python function to solve a basic arithmetic problem, here is some sample code, adopt as needed.: ```python\ndef solve_arithmetic_problem(problem):\n return eval(problem)\n\n# Example problem: What is 15 times 4?\nresult = solve_arithmetic_problem('15 * 4')\n```.",
"Goal: Your purpose is to generate creative writing prompts. Generated Prompt: You are a creative muse who can come up with engaging and unique writing prompts. Provide an intriguing prompt for a science fiction story set in a distant galaxy.",
"Goal: Your purpose is to translate sentences from English to Spanish. Generated Prompt: You are an efficient language translator. Translate the following sentence into Spanish: 'The sun rises early in the morning.'",
"Goal: Your purpose is to get the weather based on a location. Generated Prompt: You are a weather assistant capable of retrieving weather information using web scraping. Here is Python code to scrape weather data for a given location, here is some sample code for you to use: ```python\nimport requests\nfrom bs4 import BeautifulSoup\n\ndef get_weather(location):\n # Replace 'location' with the actual location parameter\n response = requests.get(f'https://wttr.in/{location}?format=3')\n return response.text\n\n# Example usage\n# weather = get_weather('<location>')\n```."
]
"Goal: Your purpose is to get the weather based on a location. Generated Prompt: You are a weather assistant capable of retrieving weather information using web scraping. Here is Python code to scrape weather data for a given location, here is some sample code for you to use: ```python\nimport requests\nfrom bs4 import BeautifulSoup\n\ndef get_weather(location):\n # Replace 'location' with the actual location parameter\n response = requests.get(f'https://wttr.in/{location}?format=3')\n return response.text\n\n# Example usage\n# weather = get_weather('<location>')\n```.",
"Goal: Your purpose is to find the current number of inhabitants of the canton of Zurich from its Wikipedia page. Generated Prompt: You are a skilled data extractor. Use the Python code provided to request the Wikipedia page of the canton of Zurich, extract the text content, search for 'Population' (ignoring case sensitivity), and print out 50 characters before and after each occurrence to find the current number of inhabitants. ```python\nimport requests\nfrom bs4 import BeautifulSoup\nimport re\n\ndef get_population_of_zurich():\n url = 'https://en.wikipedia.org/wiki/Canton_of_Zürich'\n response = requests.get(url)\n soup = BeautifulSoup(response.content, 'html.parser')\n content = soup.get_text()\n matches = re.finditer(r'Population', content, re.IGNORECASE)\n results = [content[max(0, match.start()-50):match.end()+50] for match in matches]\n return results\n\n# Example usage\n# population_info = get_population_of_zurich()\n```"
]
messages = [
{"role": "system", "content": "You are a helpful assistant knowledgeable in prompt engineering."},
{"role": "user", "content": f"Using best practices in prompt engineering, create a detailed prompt for the goal '{goal}'. This generated prompt will be combined with the following context later (but must be genertic and is forbidden to contain any of the following context): '{sample_input}'\n Examples: {examples}. Aim for maximum 50 words. Important: Any problems must be solved through sample code or learned information provided in the prompt."}
{"role": "user", "content": f"Using best practices in prompt engineering, create a detailed prompt for the goal '{goal}'. This generated prompt will be combined with the following context later (but must be genertic and is forbidden to contain any of the following context): '{sample_input}'\n Examples: {examples}. Aim for maximum 50 words. Important: Any problems must be solved through sample code or learned information provided in the prompt. Any sample code provided must be executable in isolation. Avoid unresolveable placerholder for URLs and API Keys. If you retrieve information from the web. Avoid parsing complex data in python code, use simple, reliable text operation. As long as the answer is somewhere in the output, and it is below 1k characters, its a perfect solution. Use real existing services and websites. Don't invent services or use example.com."}
]
response = self.openai_wrapper.chat_completion(
@@ -124,7 +125,7 @@ def main():
manager = MicroAgentManager(api_key)
start_time = time.time()
user_input = "Who is the current president in 2023 of france? "
user_input = "Who is the current president of france? "
manager.respond(user_input)
end_time = time.time()
elapsed_time = end_time - start_time
@@ -132,7 +133,7 @@ def main():
manager.respond(user_input)
start_time = time.time()
user_input = "Who is the current president in 2023 of france? "
user_input = "Who is the current president of france? "
manager.respond(user_input)
end_time = time.time()
elapsed_time = end_time - start_time

28
microagent.py
View File

@@ -34,6 +34,7 @@ class MicroAgent:
self.max_depth = max_depth
self.usage_count = 0
openai.api_key = api_key
self.working_agent = True
self.code_block_start = "```python"
self.code_block_end = "```"
@@ -54,6 +55,7 @@ class MicroAgent:
conversation_accumulator = ""
thought_number = 1
action_number = 1
found_new_solution = False
for iteration in range(self.max_depth):
react_prompt = f"Question: {input_text}\n{conversation_accumulator}\nThought {thought_number}: [Decompose the task. Identify if another agent or Python code execution is needed. Write 'Query Solved: <formulate detailed answer>' once the task is completed.]\nAction {action_number}: [Specify action based on the thought, e.g., 'Use Agent[Purpose of the agent as sentence:Input Paramter for agent]' for delegation or '```python\n# Python code here\n```' for execution]"
@@ -100,12 +102,20 @@ class MicroAgent:
thought_number += 1
action_number += 1
if "Query Solved" in response:
if iteration != 1 and self.working_agent is True:
# ReAct chain found solution, we need to adopt current prompt.
found_new_solution = True
logging.info(f"This Agent worked OOTB: " + self.purpose + "!")
if iteration == 1:
# Initial prompt working first try.
self.working_agent = True
logging.info(f"Found first solution for agent: " + self.purpose + "!")
break
final_answer = "Final Response: " + conversation_accumulator
logging.info(f"Final Response: {final_answer}")
return final_answer, conversation_accumulator
return final_answer, conversation_accumulator, found_new_solution
def execute_code(self, text_with_code):
try:
@@ -118,14 +128,18 @@ class MicroAgent:
logging.error(f"Error executing code: {e}")
return f"Error in executing code: {e}"
def evolve_prompt(self, input_text, output, full_conversation):
def evolve_prompt(self, input_text, output, full_conversation, new_solution):
if len(full_conversation) > 1000:
full_conversation = full_conversation[:200] + "..." + full_conversation[-1000:]
feedback = self.evaluate_agent(input_text, self.dynamic_prompt, output)
runtime_context = self.generate_runtime_context()
if "poor" in feedback.lower():
evolve_prompt_query = f"How should the GPT-4 prompt evolve based on this input and feedback? If you don't know something, write sample code in the prompt to solve it. Break down complex tasks by calling other agents if required. Please include python code that should be used to solve a certain task as per purpose or list other agents that should be called. A purpose is always a sentence long. Important: Any problems must be solved through sample code or learned information provided in the prompt. Add any learnings or information that might be useful for the future. ONLY RESPONSE WITH THE REVISED PROMPT NO OTHER TEXT! Current Prompt: {input_text}, User Feedback: {feedback}, full conversation: {full_conversation}"
if "poor" in feedback.lower() or new_solution:
evolve_prompt_query = f"How should the GPT-4 prompt evolve based on this input and feedback? If you don't know something, write sample code in the prompt to solve it. Break down complex tasks by calling other agents if required. Please include python code that should be used to solve a certain task as per purpose or list other agents that should be called. A purpose is always a sentence long. Important: Any problems must be solved through sample code or learned information provided in the new, updated prompt. It's ok to also put data in the prompt. Add any learnings or information that might be useful for the future. ONLY RESPONSE WITH THE REVISED PROMPT NO OTHER TEXT! Current Prompt: {input_text}, User Feedback: {feedback}, full conversation: {full_conversation}"
if (new_solution and self.working_agent is False):
evolve_prompt_query = f"How should the GPT-4 prompt evolve based on this input and feedback? Take a look at the solution provided in later on in the _full conversation_ section. As you can see, the problem has been solved. We need to learn from this. Adopt the code or solution found, make it reusable and compile a new, updated system prompt, so the solution can be reused in the future. Remember: Problems are solved through sample code or learned information provided in the new, updatedprompt. It's ok to also put data in the prompt. Add any learnings or information that might be useful for the future. ONLY RESPONSE WITH THE REVISED PROMPT NO OTHER TEXT! Current Prompt: {input_text}, User Feedback: {feedback}, full conversation: {full_conversation}"
self.working_agent = True
logging.info(f"Found first solution for agent: " + self.purpose + "!")
logging.info(f"Evolve prompt query: {evolve_prompt_query}")
new_prompt = self.openai_wrapper.chat_completion(
model="gpt-4-1106-preview",
@@ -135,12 +149,12 @@ class MicroAgent:
self.dynamic_prompt = new_prompt
def respond(self, input_text):
response, full_conversation = self.generate_response(input_text)
self.evolve_prompt(input_text, response, full_conversation)
response, full_conversation, new_solution = self.generate_response(input_text)
self.evolve_prompt(input_text, response, full_conversation, new_solution)
return response
def evaluate_agent(self, input_text, prompt, output):
evaluation_query = f"Evaluate the generated LLM Output: '{input_text}' with the current prompt '{prompt}' for quality and relevance (Possible Answers: Poor, Good), LLM output with current prompt: '{output}'"
evaluation_query = f"Evaluate the generated LLM Output: '{input_text}' with the current prompt '{prompt}' for quality and relevance (Possible Answers: Poor, Good, Perfect), LLM output with current prompt: '{output}'"
logging.info(f"Evaluation query: {evaluation_query}")
logging.info(f"Current prompt: {prompt}")
logging.info(f"Current output: {output}")

1
requirements.txt
View File

@@ -38,6 +38,7 @@ openpyxl
xlrd
# Logging
wikpedia
loguru
# Testing