app.py

from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool
import datetime
import requests
import pytz
import yaml
from tools.final_answer import FinalAnswerTool
from tools.web_search import DuckDuckGoSearchTool
from tools.visit_webpage import VisitWebpageTool

from Gradio_UI import GradioUI

from kokoro import KPipeline
import soundfile as sf
import os
import numpy as np


# Initialize the Kokoro pipeline
pipeline = KPipeline(lang_code='a')  # 'a' stands for American English

@tool
def text_to_speech_kokoro(text: str, voice: str = 'af_heart', speed: float = 1.0) -> str:
    """Convert text to speech using the Kokoro-82M model.
    
    Args:
        text: The text to be converted to speech.
        voice: The voice to use for speech synthesis (default is 'af_heart').
        speed: The speed of the speech (default is 1.0).
    
    Returns:
        An AgentAudio object with the relative URL to the generated audio file.
    """
    try:
        # Generate speech audio
        generator = pipeline(text, voice=voice, speed=speed, split_pattern=r'\n+')
        audio_segments = []
        for _, _, audio in generator:
            audio_segments.append(audio)
        if not audio_segments:
            raise ValueError("No audio generated.")
        # Concatenate segments into one audio array
        full_audio = np.concatenate(audio_segments)
        sample_rate = 24000  # Kokoro outputs at 24 kHz
        # Ensure the static folder exists and save the file there
        os.makedirs("tools", exist_ok=True)
        filename = os.path.join("tools", "output.wav")
        sf.write(filename, full_audio, sample_rate)
        # Return an AgentAudio object pointing to the relative URL of the audio file
        from smolagents.agent_types import AgentAudio
        return AgentAudio(f"tools/output.wav")
    except Exception as e:
        return f"Error generating speech: {str(e)}"

@tool
def search_dad_jokes(term: str) -> str:
    """A tool that searches for dad jokes containing a specific term.
    Args:
        term: The keyword to search for in dad jokes.
    """
    try:
        headers = {
            "Accept": "application/json",
            "User-Agent": "YourAppName (https://yourappurl.com)"
        }
        response = requests.get(f"https://icanhazdadjoke.com/search?term={term}", headers=headers)
        data = response.json()
        if data['results']:
            jokes = [joke['joke'] for joke in data['results']]
            return f"Found {len(jokes)} jokes:\n" + "\n\n".join(jokes)
        else:
            return f"No jokes found for the term '{term}'."
    except Exception as e:
        return f"Error searching for jokes: {str(e)}"




final_answer = FinalAnswerTool()
web_search_tool = DuckDuckGoSearchTool()
visit_webpage_tool = VisitWebpageTool()



# If the agent does not answer, the model is overloaded, please use another model or the following Hugging Face Endpoint that also contains qwen2.5 coder:
# model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud' 

model = HfApiModel(
max_tokens=2096,
temperature=0.5,
model_id='Qwen/Qwen2.5-Coder-32B-Instruct',# it is possible that this model may be overloaded
custom_role_conversions=None,
)


# Import tool from Hub
image_generation_tool = load_tool("agents-course/text-to-image", trust_remote_code=True)

with open("prompts.yaml", 'r') as stream:
    prompt_templates = yaml.safe_load(stream)
    
agent = CodeAgent(
    model=model,
    tools=[visit_webpage_tool, web_search_tool, final_answer, image_generation_tool, get_current_time_in_timezone, get_random_cocktail, search_dad_jokes, text_to_speech_kokoro], ## add your tools here (don't remove final answer)
    max_steps=6,
    verbosity_level=1,
    grammar=None,
    planning_interval=None,
    name=None,
    description=None,
    prompt_templates=prompt_templates
)


GradioUI(agent).launch()