CodeMixt / app.py
acecalisto3's picture
Update app.py
b605b14 verified
raw
history blame
13.4 kB
from huggingface_hub import InferenceClient
import gradio as gr
import random
import os
import subprocess
from typing import Dict, Tuple, List
import dotenv
import json
from rich import print as rprint
from rich.panel import Panel
from rich.progress import track
from rich.table import Table
from rich.prompt import Prompt, Confirm
from rich.markdown import Markdown
from rich.traceback import install
install() # Enable rich tracebacks for easier debugging
from huggingface_hub import RepositoryInfo
model_info = RepositoryInfo.fetch('huggingface/models/{}/{}'.format(" Meta-Llama-3.1-8B ", MODEL_NAME))
client = InferenceClient(model_info, token=os.environ.get("HF_TOKEN"))
# Chat Interface Parameters
DEFAULT_TEMPERATURE = 0.9
DEFAULT_MAX_NEW_TOKENS = 2048
DEFAULT_TOP_P = 0.95
DEFAULT_REPETITION_PENALTY = 1.2
# Local Server
LOCAL_HOST_PORT = 7860
# --- Agent Roles ---
agent_roles: Dict[str, Dict[str, bool]] = {
"Web Developer": {"description": "A master of front-end and back-end web development.", "active": False},
"Prompt Engineer": {"description": "An expert in crafting effective prompts for AI models.", "active": False},
"Python Code Developer": {"description": "A skilled Python programmer who can write clean and efficient code.", "active": False},
"Hugging Face Hub Expert": {"description": "A specialist in navigating and utilizing the Hugging Face Hub.", "active": False},
"AI-Powered Code Assistant": {"description": "An AI assistant that can help with coding tasks and provide code snippets.", "active": False},
}
# --- Initial Prompt ---
initial_prompt = """
You are an expert agent cluster, consisting of a Web Developer, a Prompt Engineer, a Python Code Developer, a Hugging Face Hub Expert, and an AI-Powered Code Assistant.
Respond with complete program coding to client requests.
Use your combined expertise to research information and explain it clearly.
Don't answer solely based on what you already know. Always perform a search before providing a response.
In special cases, like when the user specifies a page to read, there's no need to search.
Read the provided page and answer the user's question accordingly.
If you find limited information from search results, try these options:
- Click on the links of the search results to access and read the content of each page.
- Change your search query and perform a new search.
Users are busy, so provide direct answers.
BAD ANSWER EXAMPLE
- Please refer to these pages.
- You can write code referring to these pages.
- The following page will be helpful.
GOOD ANSWER EXAMPLE
- This is the complete code: -- complete code here --
- The answer to your question is -- answer here --
List the URLs of the pages you referenced at the end of your answer for verification.
Answer in the language used by the user. If the user asks in Japanese, answer in Japanese. If the user asks in Spanish, answer in Spanish.
Search in English, especially for programming-related questions. ALWAYS SEARCH IN ENGLISH FOR THOSE.
"""
# --- Custom CSS ---
customCSS = """
#component-7 {
height: 1600px;
flex-grow: 4;
}
.gradio-container {
display: flex;
flex-direction: column;
height: 100vh;
}
.gradio-interface {
flex-grow: 1;
display: flex;
flex-direction: column;
}
"""
# --- Functions ---
# Function to toggle the active state of an agent
def toggle_agent(agent_name: str) -> str:
"""Toggles the active state of an agent."""
global agent_roles
agent_roles[agent_name]["active"] = not agent_roles[agent_name]["active"]
return f"{agent_name} is now {'active' if agent_roles[agent_name]['active'] else 'inactive'}"
# Function to get the active agent cluster
def get_active_agents() -> List[str]:
"""Returns a list of active agents."""
return [agent for agent, is_active in agent_roles.items() if is_active]
# Function to execute code
def run_code(code: str) -> str:
"""Executes the provided code and returns the output."""
try:
output = subprocess.check_output(
['python', '-c', code],
stderr=subprocess.STDOUT,
universal_newlines=True,
)
return output
except subprocess.CalledProcessError as e:
return f"Error: {e.output}"
# Function to format the prompt
def format_prompt(message: str, history: list[Tuple[str, str]], agent_roles: list[str]) -> str:
"""Formats the prompt with the selected agent roles and conversation history."""
prompt = initial_prompt # Use the global initial prompt
for user_prompt, bot_response in history:
prompt += f"[INST] {user_prompt} [/INST]"
prompt += f" {bot_response}</s> "
prompt += f"[INST] {message} [/INST]"
return prompt
# Function to generate a response
def generate(prompt: str, history: list[Tuple[str, str]], agent_roles: list[str], temperature: float = DEFAULT_TEMPERATURE, max_new_tokens: int = DEFAULT_MAX_NEW_TOKENS, top_p: float = DEFAULT_TOP_P, repetition_penalty: float = DEFAULT_REPETITION_PENALTY) -> str:
"""Generates a response using the selected agent roles and parameters."""
temperature = float(temperature)
if temperature < 1e-2:
temperature = 1e-2
top_p = float(top_p)
generate_kwargs = dict(
temperature=temperature,
max_new_tokens=max_new_tokens,
top_p=top_p,
repetition_penalty=repetition_penalty,
do_sample=True,
seed=random.randint(0, 10**7),
)
formatted_prompt = format_prompt(prompt, history, agent_roles)
stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
output = ""
for response in stream:
output += response.token.text
yield output
return output
# Function to handle user input and generate responses
def chat_interface(message: str, history: list[Tuple[str, str]], temperature: float, max_new_tokens: int, top_p: float, repetition_penalty: float) -> Tuple[str, str]:
"""Handles user input and generates responses."""
if message.startswith("python"):
# User entered code, execute it
code = message[9:-3]
output = run_code(code)
return (message, output)
else:
# User entered a normal message, generate a response
active_agents = get_active_agents()
response = generate(message, history, active_agents, temperature, max_new_tokens, top_p, repetition_penalty)
return (message, response)
# Function to create a new web app instance
def create_web_app(app_name: str, code: str) -> None:
"""Creates a new web app instance with the given name and code."""
# Create a new directory for the app
os.makedirs(app_name, exist_ok=True)
# Create the app.py file
with open(os.path.join(app_name, 'app.py'), 'w') as f:
f.write(code)
# Create the requirements.txt file
with open(os.path.join(app_name, 'requirements.txt'), 'w') as f:
f.write("gradio\nhuggingface_hub\nrich")
# Print a success message
print(f"Web app '{app_name}' created successfully!")
# Function to handle the "Create Web App" button click
def create_web_app_button_click(app_name: str, code: str) -> str:
"""Handles the "Create Web App" button click."""
# Validate the app name
if not app_name:
return "Please enter a valid app name."
# Create the web app instance
create_web_app(app_name, code)
# Return a success message
return f"Web app '{app_name}' created successfully!"
# Function to handle the "Deploy" button click
def deploy_button_click(app_name: str, code: str) -> str:
"""Handles the "Deploy" button click."""
# Validate the app name
if not app_name:
return "Please enter a valid app name."
# Deploy the web app instance
# ... (Implement deployment logic here)
# Return a success message
return f"Web app '{app_name}' deployed successfully!"
# Function to handle the "Local Host" button click
def local_host_button_click(app_name: str, code: str) -> str:
"""Handles the "Local Host" button click."""
# Validate the app name
if not app_name:
return "Please enter a valid app name."
# Start the local server
os.chdir(app_name)
subprocess.Popen(['gradio', 'run', 'app.py', '--share', '--server_port', str(LOCAL_HOST_PORT)])
# Return a success message
return f"Web app '{app_name}' running locally on port {LOCAL_HOST_PORT}!"
# Function to handle the "Ship" button click
def ship_button_click(app_name: str, code: str) -> str:
"""Handles the "Ship" button click."""
# Validate the app name
if not app_name:
return "Please enter a valid app name."
# Ship the web app instance
# ... (Implement shipping logic here)
# Return a success message
return f"Web app '{app_name}' shipped successfully!"
# --- Gradio Interface ---
with gr.Blocks(css=customCSS, theme='ParityError/Interstellar') as demo:
gr.Markdown(
"""
# AI-Powered Code Generation and Web App Creation
This application allows you to interact with an AI agent cluster to generate code and create web apps.
"""
)
# --- Agent Selection ---
with gr.Row():
gr.Markdown("## Select Your Agent Cluster")
for agent_name, agent_data in agent_roles.items():
button = gr.Button(agent_name, variant="secondary")
textbox = gr.Textbox(agent_data["description"], interactive=False)
button.click(toggle_agent, inputs=[button], outputs=[textbox])
# --- Chat Interface ---
with gr.Row():
gr.Markdown("## Chat with the AI")
chatbot = gr.Chatbot()
chat_interface_input = gr.Textbox(label="Enter your message", placeholder="Ask me anything!")
# Parameters
with gr.Accordion("Advanced Parameters", open=False):
temperature_slider = gr.Slider(
label="Temperature",
value=DEFAULT_TEMPERATURE,
minimum=0.0,
maximum=1.0,
step=0.05,
interactive=True,
info="Higher values generate more diverse outputs",
)
max_new_tokens_slider = gr.Slider(
label="Maximum New Tokens",
value=DEFAULT_MAX_NEW_TOKENS,
minimum=64,
maximum=4096,
step=64,
interactive=True,
info="The maximum number of new tokens",
)
top_p_slider = gr.Slider(
label="Top-p (Nucleus Sampling)",
value=DEFAULT_TOP_P,
minimum=0.0,
maximum=1,
step=0.05,
interactive=True,
info="Higher values sample more low-probability tokens",
)
repetition_penalty_slider = gr.Slider(
label="Repetition Penalty",
value=DEFAULT_REPETITION_PENALTY,
minimum=1.0,
maximum=2.0,
step=0.05,
interactive=True,
info="Penalize repeated tokens",
)
# Submit Button
submit_button = gr.Button("Submit")
# Chat Interface Logic
submit_button.click(
chat_interface,
inputs=[
chat_interface_input,
chatbot,
temperature_slider,
max_new_tokens_slider,
top_p_slider,
repetition_penalty_slider,
],
outputs=[
chatbot,
],
)
# --- Web App Creation ---
with gr.Row():
gr.Markdown("## Create Your Web App")
app_name_input = gr.Textbox(label="App Name", placeholder="Enter your app name")
code_output = gr.Textbox(label="Code", interactive=False)
create_web_app_button = gr.Button("Create Web App")
deploy_button = gr.Button("Deploy")
local_host_button = gr.Button("Local Host")
ship_button = gr.Button("Ship")
# Web App Creation Logic
create_web_app_button.click(
create_web_app_button_click,
inputs=[app_name_input, code_output],
outputs=[gr.Textbox(label="Status", interactive=False)],
)
# Deploy the web app
deploy_button.click(
deploy_button_click,
inputs=[app_name_input, code_output],
outputs=[gr.Textbox(label="Status", interactive=False)],
)
# Local host the web app
local_host_button.click(
local_host_button_click,
inputs=[app_name_input, code_output],
outputs=[gr.Textbox(label="Status", interactive=False)],
)
# Ship the web app
ship_button.click(
ship_button_click,
inputs=[app_name_input, code_output],
outputs=[gr.Textbox(label="Status", interactive=False)],
)
# --- Connect Chat Output to Code Output ---
chatbot.change(
lambda x: x[-1][1] if x else "",
inputs=[chatbot],
outputs=[code_output],
)
# --- Initialize Hugging Face Client ---
client = InferenceClient(repo_id="MODEL_NAME", token=os.environ.get("HF_TOKEN"))
# --- Launch Gradio ---
demo.queue().launch(debug=True)