app.py

from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool
import xml.etree.ElementTree as ET
import matplotlib.pyplot as plt
from wordcloud import WordCloud
from collections import Counter
import re
import datetime
import requests
import pytz
import yaml
from tools.final_answer import FinalAnswerTool

from Gradio_UI import GradioUI

# Below is an example of a tool that does nothing. Amaze us with your creativity !
@tool
def search_arxiv(query: str):
    """Searches arXiv for academic papers and returns structured results.

    Args:
        query (str): The topic or keywords to search for.

    Returns:
        list: A list of tuples containing titles, summaries, and links.
    """
    max_results = 5
    url = f"http://export.arxiv.org/api/query?search_query={query}&max_results={max_results}"
    response = requests.get(url)

    if response.status_code == 200:
        papers = []
        root = ET.fromstring(response.text)
        for entry in root.findall("{http://www.w3.org/2005/Atom}entry"):
            title = entry.find("{http://www.w3.org/2005/Atom}title").text
            summary = entry.find("{http://www.w3.org/2005/Atom}summary").text
            link = entry.find("{http://www.w3.org/2005/Atom}id").text
            papers.append((title, summary, link))

        return papers

    return []

def generate_visuals(query):
    results = search_arxiv(query)
    if not results:
        return "No papers found.", None, None

    # Extract text data
    titles = [title for title, _, _ in results]
    summaries = " ".join(summary for _, summary, _ in results)

    # Generate Bar Chart for Keyword Frequency in Titles
    words = [word.lower() for title in titles for word in re.findall(r'\b\w+\b', title) if len(word) > 3]
    word_counts = Counter(words).most_common(10)

    plt.figure(figsize=(8, 5))
    plt.bar(*zip(*word_counts), color='skyblue')
    plt.xticks(rotation=45)
    plt.title("Top Keywords in Titles")
    plt.xlabel("Keywords")
    plt.ylabel("Frequency")
    plt.tight_layout()
    bar_chart_path = "bar_chart.png"
    plt.savefig(bar_chart_path)
    plt.close()

    # Generate Word Cloud for Summary Text
    wordcloud = WordCloud(width=500, height=300, background_color="white").generate(summaries)
    plt.figure(figsize=(8, 5))
    plt.imshow(wordcloud, interpolation="bilinear")
    plt.axis("off")
    wordcloud_path = "wordcloud.png"
    plt.savefig(wordcloud_path)
    plt.close()

    # Display Search Results as Clickable Links
    markdown_text = "\n\n".join(
        [f"**[{title}]({link})**\n\n{summary}" for title, summary, link in results]
    )

    return markdown_text, bar_chart_path, wordcloud_path


@tool
def summarize_text(text: str) -> str:
    """Summarizes long academic papers or articles.
    Args:
        text: The text to summarize.
    """
    model = HfApiModel(
        max_tokens=512,
        temperature=0.5,
        model_id='Qwen/Qwen2.5-Coder-32B-Instruct',
    )
    return model.generate(f"Summarize this research paper: {text}")

@tool
def get_current_time_in_timezone(timezone: str) -> str:
    """A tool that fetches the current local time in a specified timezone.
    Args:
        timezone: A string representing a valid timezone (e.g., 'America/New_York').
    """
    try:
        # Create timezone object
        tz = pytz.timezone(timezone)
        # Get current time in that timezone
        local_time = datetime.datetime.now(tz).strftime("%Y-%m-%d %H:%M:%S")
        return f"The current local time in {timezone} is: {local_time}"
    except Exception as e:
        return f"Error fetching time for timezone '{timezone}': {str(e)}"


final_answer = FinalAnswerTool()

# 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=[final_answer, search_arxiv, summarize_text], ## add your tools here (don't remove final answer)
    max_steps=6,
    verbosity_level=1,
    grammar=None,
    planning_interval=None,
    name="Research Assistant",
    description=None,
    prompt_templates=prompt_templates
)

iface = gr.Interface(
    fn=generate_visuals,
    inputs="text",
    outputs=["markdown", "image", "image"],
    title="🔎 arXiv Research Paper Search",
    description="Enter a topic or keywords to search for academic papers on arXiv. Get a list of papers with visual analysis.",
    examples=[["Machine Learning"], ["Quantum Computing"], ["Climate Change"]]
)

iface.launch()

GradioUI(agent).launch()