STELLA / predefined_tools.py
asfarasimconcerned's picture
Upload 23 files
69596ec verified
Raw
History Blame Contribute Delete
38 kB
import requests
from markdownify import markdownify
import re
from pathlib import Path
import subprocess
from requests.exceptions import RequestException
from smolagents import tool
from bs4 import BeautifulSoup
from io import BytesIO
import PyPDF2
import os
from typing import Optional, Dict, Any
# Import googlesearch-python for reliable Google search
try:
from googlesearch import search
GOOGLE_SEARCH_AVAILABLE = True
except ImportError:
GOOGLE_SEARCH_AVAILABLE = False
print("⚠️ googlesearch-python not available. Install with: pip install googlesearch-python")
# === Reliable Search Tools for STELLA ===
# Simplified and optimized for maximum reliability
@tool
def enhanced_google_search(query: str, num_results: int = 5, include_snippets: bool = True) -> str:
"""Enhanced Google search with reliable implementation (most reliable search method).
Args:
query: Search query
num_results: Number of results to return (default: 5)
include_snippets: Whether to include result snippets (default: True)
Returns:
Well-formatted search results with titles, URLs, and descriptions
"""
try:
if GOOGLE_SEARCH_AVAILABLE:
# Use googlesearch-python for reliable results
results = []
search_results = list(search(query, num_results=num_results, advanced=True))
for i, result in enumerate(search_results, 1):
if i > num_results:
break
title = getattr(result, 'title', f'Search Result {i}')
url = getattr(result, 'url', 'No URL')
description = getattr(result, 'description', 'No description available')
if include_snippets:
results.append(f"**{i}. {title}**\n🔗 {url}\n📄 {description}\n")
else:
results.append(f"**{i}. {title}**\n🔗 {url}\n")
if results:
formatted_results = "\n".join(results)
return f"🔍 Enhanced Google Search Results for '{query}':\n\n{formatted_results}"
else:
return f"No search results found for query: '{query}'"
else:
# Fallback to basic search implementation
return search_google_basic(query, num_results)
except Exception as e:
return f"❌ Enhanced Google search failed: {str(e)}"
@tool
def search_google_basic(query: str, num_results: int = 3) -> str:
"""Basic Google search fallback implementation.
Args:
query: Search query
num_results: Number of results to return
Returns:
Basic search results
"""
try:
if GOOGLE_SEARCH_AVAILABLE:
results_string = ""
search_results = list(search(query, num_results=num_results))
for i, url in enumerate(search_results, 1):
results_string += f"{i}. {url}\n"
return f"Google Search Results for '{query}':\n\n{results_string}" if results_string else f"No results found for: {query}"
else:
return "❌ Google search not available. Please install googlesearch-python: pip install googlesearch-python"
except Exception as e:
return f"❌ Google search failed: {str(e)}"
@tool
def multi_source_search(query: str, sources: str = "google,serpapi") -> str:
"""Unified search tool with flexible source combinations for different needs.
Args:
query: Search query
sources: Comma-separated sources. Options:
• "google" - Basic Google search (~0.3s)
• "google,serpapi" - Enhanced Google search (~1-2s, DEFAULT)
• "google,knowledge" - Google + AI knowledge base (~30s)
• "google,knowledge,serpapi" - All sources (~45s, most comprehensive)
Quick Guide:
• Simple queries → "google"
• Most queries → "google,serpapi" (DEFAULT - enhanced results, fast)
• Deep research → "google,knowledge"
• Comprehensive research → "google,knowledge,serpapi"
Returns:
Consolidated search results with clear source attribution
"""
source_list = [s.strip().lower() for s in sources.split(",")]
results = []
# Enhanced Google search (most reliable)
if "google" in source_list:
google_result = enhanced_google_search(query, num_results=3)
if google_result and not google_result.startswith("❌"):
results.append(f"## 🌐 Enhanced Google Search\n{google_result}")
# SerpAPI search (if API key available and requested)
if "serpapi" in source_list:
serpapi_result = search_with_serpapi(query)
if serpapi_result and not serpapi_result.startswith("⚠️") and not serpapi_result.startswith("❌"):
results.append(f"## {serpapi_result}")
# Knowledge search (using OpenRouter)
if "knowledge" in source_list:
knowledge_result = enhanced_knowledge_search(query)
if knowledge_result and not knowledge_result.startswith("⚠️") and not knowledge_result.startswith("❌"):
results.append(f"## {knowledge_result}")
if not results:
return f"❌ No successful searches completed for query: '{query}'"
consolidated = f"# 🔍 Multi-Source Search Results for: '{query}'\n\n" + "\n\n---\n\n".join(results)
return consolidated
@tool
def smart_search_router(query: str, domain: str = "auto") -> str:
"""Intelligently route search queries to the most appropriate reliable method.
Args:
query: Search query
domain: Domain hint (auto, scientific, general, technical, research)
Returns:
Results from the most appropriate search method
"""
query_lower = query.lower()
# Auto-detect domain if not specified
if domain == "auto":
scientific_keywords = ["research", "study", "paper", "journal", "publication", "experiment",
"clinical", "trial", "hypothesis", "methodology", "analysis"]
technical_keywords = ["algorithm", "code", "programming", "software", "framework",
"implementation", "technical", "engineering", "tutorial", "install"]
if any(keyword in query_lower for keyword in scientific_keywords):
domain = "scientific"
elif any(keyword in query_lower for keyword in technical_keywords):
domain = "technical"
else:
domain = "general"
# Route to appropriate reliable search method
if domain == "scientific" or domain == "research":
# For scientific queries, try enhanced knowledge search first
knowledge_result = enhanced_knowledge_search(query)
if knowledge_result and not knowledge_result.startswith("⚠️") and not knowledge_result.startswith("❌"):
return knowledge_result
else:
# Fallback to enhanced Google search
return enhanced_google_search(query, num_results=5)
elif domain == "technical":
# Use multi-source for technical queries
return multi_source_search(query, "google,knowledge")
else: # general queries
# Use enhanced Google search for general queries (most reliable)
return enhanced_google_search(query, num_results=5)
# === Supplementary Search Tools (kept for completeness but not guaranteed reliable) ===
@tool
def search_with_serpapi(query: str, num_results: int = 5) -> str:
"""Advanced Google search using SerpAPI (requires API key).
Args:
query: Search query
num_results: Number of results to return (default: 5)
Returns:
Formatted search results with titles, URLs, and descriptions
"""
serpapi_key = os.getenv("SERPAPI_API_KEY")
if not serpapi_key:
return "⚠️ SerpAPI key not found. Please set SERPAPI_API_KEY in your .env file"
try:
url = "https://serpapi.com/search"
params = {
"engine": "google",
"q": query,
"api_key": serpapi_key,
"num": min(num_results, 10),
"format": "json"
}
response = requests.get(url, params=params, timeout=10)
response.raise_for_status()
data = response.json()
if "organic_results" not in data:
return f"No results found for query: '{query}'"
results = []
for result in data["organic_results"][:num_results]:
title = result.get("title", "No title")
link = result.get("link", "No link")
snippet = result.get("snippet", "No description")
results.append(f"**{title}**\n{link}\n{snippet}\n")
formatted_results = "\n".join(results)
return f"🔍 SerpAPI Results for '{query}':\n\n{formatted_results}"
except Exception as e:
return f"❌ SerpAPI search failed: {str(e)}"
@tool
def enhanced_knowledge_search(query: str, model_name: str = "gemini-2.5-pro") -> str:
"""Use LLM's internal knowledge to provide detailed information.
Args:
query: Knowledge query or research question
model_name: LLM model to use for knowledge expansion
Returns:
Detailed information based on LLM's training knowledge
"""
openrouter_key = os.getenv("OPENROUTER_API_KEY")
if not openrouter_key:
return "⚠️ OpenRouter API key not found for knowledge search"
try:
url = "https://openrouter.ai/api/v1/chat/completions"
prompt = f"""You are an assistant to researchers and scientists. Provide detailed, accurate information for this query.
Query: {query}
Please provide:
1. Clear explanation of the topic
2. Key concepts and principles
3. Current understanding and developments
4. Practical applications
5. Important considerations
Be comprehensive and accurate."""
payload = {
"model": f"google/{model_name}",
"messages": [
{
"role": "system",
"content": "You are a knowledgeable research assistant providing accurate information."
},
{
"role": "user",
"content": prompt
}
],
"temperature": 0.1,
"max_tokens": 3000
}
headers = {
"Authorization": f"Bearer {openrouter_key}",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers, timeout=30)
response.raise_for_status()
result = response.json()
if "choices" in result and len(result["choices"]) > 0:
content = result["choices"][0]["message"]["content"]
return f"🧠 Knowledge Base Response for '{query}':\n\n{content}"
return "Unable to generate knowledge response."
except Exception as e:
return f"❌ Knowledge search failed: {str(e)}"
# === Legacy Search Function (for compatibility) ===
@tool
def search_google(query: str, num_results: int = 3, language: str = 'en') -> str:
"""
Legacy Google search function (redirects to reliable implementation).
Args:
query: The search query
num_results: Number of results to return (default: 3)
language: Language code for search results (default: 'en')
Returns:
Formatted string containing search results
"""
# Redirect to reliable enhanced Google search
return enhanced_google_search(query, num_results, include_snippets=True)
# --- Custom Tools ---
@tool
def visit_webpage(url: str) -> str:
"""Visits a webpage at the given URL and returns its content as a markdown string.
Args:
url: The URL of the webpage to visit.
Returns:
The content of the webpage converted to Markdown, or an error message if the request fails.
"""
try:
# Send a GET request to the URL
response = requests.get(url)
response.raise_for_status() # Raise an exception for bad status codes
# Convert the HTML content to Markdown
markdown_content = markdownify(response.text).strip()
# Remove multiple line breaks
markdown_content = re.sub(r"\n{3,}", "\n\n", markdown_content)
return markdown_content
except RequestException as e:
return f"Error fetching the webpage: {str(e)}"
except Exception as e:
return f"An unexpected error occurred: {str(e)}"
@tool
def search_github_repositories(query: str, language: str = "", sort: str = "stars", order: str = "desc", per_page: int = 10) -> str:
"""Search GitHub repositories for packages, models, or projects.
Args:
query: Search query (e.g., "transformer model", "pytorch CNN", "machine learning")
language: Programming language filter (e.g., "Python", "JavaScript")
sort: Sort results by "stars", "forks", "updated", or "created"
order: Order results "desc" or "asc"
per_page: Number of results to return (max 100)
Returns:
Formatted list of repository information including name, description, stars, and URL
"""
try:
# GitHub API endpoint for repository search
url = "https://api.github.com/search/repositories"
# Build search query
search_query = query
if language:
search_query += f" language:{language}"
params = {
"q": search_query,
"sort": sort,
"order": order,
"per_page": min(per_page, 100) # GitHub API limit
}
# Make request to GitHub API
response = requests.get(url, params=params)
response.raise_for_status()
data = response.json()
repositories = data.get("items", [])
if not repositories:
return f"No repositories found for query: {query}"
# Format results
result = f"GitHub搜索结果 (查询: '{query}'):\n\n"
for i, repo in enumerate(repositories, 1):
name = repo.get("name", "N/A")
full_name = repo.get("full_name", "N/A")
description = repo.get("description", "No description available")
stars = repo.get("stargazers_count", 0)
forks = repo.get("forks_count", 0)
language_used = repo.get("language", "N/A")
html_url = repo.get("html_url", "N/A")
updated_at = repo.get("updated_at", "N/A")
result += f"{i}. **{name}** ({full_name})\n"
result += f" 描述: {description}\n"
result += f" 语言: {language_used} | ⭐ {stars} | 🍴 {forks}\n"
result += f" 更新时间: {updated_at[:10]}\n"
result += f" 链接: {html_url}\n\n"
result += f"总共找到 {data.get('total_count', 0)} 个仓库"
return result
except RequestException as e:
return f"GitHub API请求失败: {str(e)}"
except Exception as e:
return f"搜索GitHub仓库时发生错误: {str(e)}"
@tool
def search_github_code(query: str, language: str = "", filename: str = "", extension: str = "", per_page: int = 10) -> str:
"""Search for code snippets in GitHub repositories.
Args:
query: Code search query (e.g., "def train_model", "class CNN")
language: Programming language filter
filename: Specific filename to search in
extension: File extension filter (e.g., "py", "js")
per_page: Number of results to return (max 100)
Returns:
Code search results with file paths, repository info, and code snippets
"""
try:
url = "https://api.github.com/search/code"
# Build search query
search_query = query
if language:
search_query += f" language:{language}"
if filename:
search_query += f" filename:{filename}"
if extension:
search_query += f" extension:{extension}"
params = {
"q": search_query,
"per_page": min(per_page, 100)
}
response = requests.get(url, params=params)
response.raise_for_status()
data = response.json()
code_items = data.get("items", [])
if not code_items:
return f"未找到相关代码: {query}"
result = f"GitHub代码搜索结果 (查询: '{query}'):\n\n"
for i, item in enumerate(code_items, 1):
name = item.get("name", "N/A")
path = item.get("path", "N/A")
repo_name = item.get("repository", {}).get("full_name", "N/A")
html_url = item.get("html_url", "N/A")
result += f"{i}. **{name}**\n"
result += f" 仓库: {repo_name}\n"
result += f" 路径: {path}\n"
result += f" 链接: {html_url}\n\n"
result += f"总共找到 {data.get('total_count', 0)} 个代码文件"
return result
except RequestException as e:
return f"GitHub代码搜索失败: {str(e)}"
except Exception as e:
return f"搜索GitHub代码时发生错误: {str(e)}"
@tool
def get_github_repository_info(repo_owner: str, repo_name: str) -> str:
"""Get detailed information about a specific GitHub repository.
Args:
repo_owner: Repository owner/organization name
repo_name: Repository name
Returns:
Detailed repository information including README, releases, and installation instructions
"""
try:
# Get repository information
repo_url = f"https://api.github.com/repos/{repo_owner}/{repo_name}"
response = requests.get(repo_url)
response.raise_for_status()
repo_data = response.json()
# Get README content
readme_url = f"https://api.github.com/repos/{repo_owner}/{repo_name}/readme"
try:
readme_response = requests.get(readme_url)
readme_response.raise_for_status()
readme_data = readme_response.json()
readme_content = requests.get(readme_data["download_url"]).text
except:
readme_content = "README不可用"
# Get latest release
releases_url = f"https://api.github.com/repos/{repo_owner}/{repo_name}/releases/latest"
try:
release_response = requests.get(releases_url)
release_response.raise_for_status()
latest_release = release_response.json()
release_info = f"最新版本: {latest_release.get('tag_name', 'N/A')} ({latest_release.get('published_at', 'N/A')[:10]})"
except:
release_info = "无可用版本信息"
# Format result
result = f"GitHub仓库详细信息: {repo_owner}/{repo_name}\n\n"
result += f"描述: {repo_data.get('description', 'N/A')}\n"
result += f"语言: {repo_data.get('language', 'N/A')}\n"
result += f"⭐ Stars: {repo_data.get('stargazers_count', 0)}\n"
result += f"🍴 Forks: {repo_data.get('forks_count', 0)}\n"
result += f"📁 Size: {repo_data.get('size', 0)} KB\n"
result += f"📅 创建时间: {repo_data.get('created_at', 'N/A')[:10]}\n"
result += f"🔄 更新时间: {repo_data.get('updated_at', 'N/A')[:10]}\n"
result += f"{release_info}\n"
result += f"🔗 链接: {repo_data.get('html_url', 'N/A')}\n"
result += f"📄 Clone URL: {repo_data.get('clone_url', 'N/A')}\n\n"
# Add topics/tags if available
topics = repo_data.get('topics', [])
if topics:
result += f"🏷️ 标签: {', '.join(topics)}\n\n"
# Add README content (first 1000 characters)
if readme_content and readme_content != "README不可用":
result += "📖 README预览:\n"
result += "=" * 50 + "\n"
result += readme_content[:1000]
if len(readme_content) > 1000:
result += "\n... (截取前1000字符)"
result += "\n" + "=" * 50 + "\n"
return result
except RequestException as e:
return f"获取GitHub仓库信息失败: {str(e)}"
except Exception as e:
return f"处理GitHub仓库信息时发生错误: {str(e)}"
@tool
def run_shell_command(command: str, working_directory: str = None) -> str:
"""Execute a shell command and return the output.
Args:
command: The shell command to execute
working_directory: Optional working directory for the command
Returns:
Command output or error message
"""
try:
result = subprocess.run(
command,
shell=True,
capture_output=True,
text=True,
cwd=working_directory,
timeout=300 # 5 minute timeout
)
output = f"Command: {command}\n"
output += f"Return code: {result.returncode}\n"
output += f"STDOUT:\n{result.stdout}\n"
if result.stderr:
output += f"STDERR:\n{result.stderr}\n"
return output
except subprocess.TimeoutExpired:
return f"Command timed out after 5 minutes: {command}"
except Exception as e:
return f"Error executing command '{command}': {str(e)}"
@tool
def create_conda_environment(env_name: str, python_version: str = "3.9") -> str:
"""Create a new conda environment.
Args:
env_name: Name of the conda environment
python_version: Python version for the environment (default: 3.9)
Returns:
Result of the conda environment creation
"""
command = f"conda create -n {env_name} python={python_version} -y"
return run_shell_command(command)
@tool
def install_packages_conda(env_name: str, packages: str) -> str:
"""Install packages in a conda environment.
Args:
env_name: Name of the conda environment
packages: Space-separated list of packages to install
Returns:
Result of the package installation
"""
command = f"conda activate {env_name} && conda install {packages} -y"
return run_shell_command(command)
@tool
def install_packages_pip(env_name: str, packages: str) -> str:
"""Install pip packages in a conda environment.
Args:
env_name: Name of the conda environment
packages: Space-separated list of pip packages to install
Returns:
Result of the pip installation
"""
command = f"conda activate {env_name} && pip install {packages}"
return run_shell_command(command)
@tool
def check_gpu_status(dummy_param: str = "") -> str:
"""Check GPU status and availability using nvidia-smi.
Args:
dummy_param: Unused parameter to handle smolagents' automatic parameter passing
Returns:
GPU status information or error if no GPUs available
"""
try:
result = subprocess.run(
["nvidia-smi", "--query-gpu=index,name,memory.total,memory.used,memory.free,utilization.gpu", "--format=csv"],
capture_output=True,
text=True,
timeout=30
)
if result.returncode == 0:
return f"GPU Status:\n{result.stdout}"
else:
return f"Error checking GPU status: {result.stderr}"
except FileNotFoundError:
return "nvidia-smi not found. No NVIDIA GPUs available or drivers not installed."
except Exception as e:
return f"Error checking GPU status: {str(e)}"
@tool
def create_script(script_name: str, script_content: str, directory: str = ".", script_type: str = "python") -> str:
"""Create a script file.
Args:
script_name: Name of the script file (should include appropriate extension)
script_content: Content of the script
directory: Directory to create the script in (default: current directory)
script_type: Type of script (python, bash, etc.) for informational purposes
Returns:
Result of script creation
"""
try:
script_path = Path(directory) / script_name
script_path.parent.mkdir(parents=True, exist_ok=True)
with open(script_path, 'w', encoding='utf-8') as f:
f.write(script_content)
# Make script executable if it's a bash script
if script_name.endswith('.sh') or script_type.lower() in ['bash', 'shell']:
import os
os.chmod(script_path, 0o755)
return f"Successfully created {script_type} script: {script_path}"
except Exception as e:
return f"Error creating script '{script_name}': {str(e)}"
@tool
def run_script(script_path: str, env_name: str = None, working_directory: str = None, interpreter: str = "python") -> str:
"""Run a script with optional conda environment activation.
Args:
script_path: Path to the script to run
env_name: Name of the conda environment (optional)
working_directory: Working directory for the script execution (optional)
interpreter: Script interpreter (python, bash, etc.) - default: python
Returns:
Output from the script execution
"""
# Determine the appropriate command based on file extension and interpreter
script_name = Path(script_path).name
if script_name.endswith('.sh') or interpreter.lower() in ['bash', 'shell']:
base_command = f"bash {script_path}"
elif script_name.endswith('.py') or interpreter.lower() == 'python':
base_command = f"python {script_path}"
else:
# Try to run with specified interpreter
base_command = f"{interpreter} {script_path}"
# Add conda environment activation if specified
if env_name:
command = f"conda activate {env_name} && {base_command}"
else:
command = base_command
return run_shell_command(command, working_directory)
@tool
def create_requirements_file(requirements: str, directory: str = ".") -> str:
"""Create a requirements.txt file.
Args:
requirements: Content of the requirements file (one package per line)
directory: Directory to create the file in
Returns:
Result of file creation
"""
try:
req_path = Path(directory) / "requirements.txt"
req_path.parent.mkdir(parents=True, exist_ok=True)
with open(req_path, 'w', encoding='utf-8') as f:
f.write(requirements)
return f"Successfully created requirements.txt: {req_path}"
except Exception as e:
return f"Error creating requirements.txt: {str(e)}"
@tool
def monitor_training_logs(log_file_path: str, lines: int = 50) -> str:
"""Monitor training logs by reading the last N lines of a log file.
Args:
log_file_path: Path to the log file
lines: Number of lines to read from the end (default: 50)
Returns:
Last N lines of the log file
"""
try:
command = f"tail -n {lines} {log_file_path}"
return run_shell_command(command)
except Exception as e:
return f"Error reading log file '{log_file_path}': {str(e)}"
@tool
def fetch_supplementary_info_from_doi(doi: str, output_dir: str = "supplementary_info") -> str:
"""
Fetches supplementary information for a paper given its DOI and returns a research log.
Args:
doi: The paper DOI
output_dir: Directory to save supplementary files (default: "supplementary_info")
Returns:
A formatted research log string containing the download process and results
"""
research_log = []
research_log.append(f"Starting process for DOI: {doi}")
# CrossRef API to resolve DOI to a publisher page
crossref_url = f"https://doi.org/{doi}"
headers = {"User-Agent": "Mozilla/5.0"}
response = requests.get(crossref_url, headers=headers)
if response.status_code != 200:
log_message = f"Failed to resolve DOI: {doi}. Status Code: {response.status_code}"
research_log.append(log_message)
return '\n'.join(research_log)
publisher_url = response.url
research_log.append(f"Resolved DOI to publisher page: {publisher_url}")
# Fetch publisher page
response = requests.get(publisher_url, headers=headers)
if response.status_code != 200:
log_message = f"Failed to access publisher page for DOI {doi}."
research_log.append(log_message)
return '\n'.join(research_log)
# Parse page content
soup = BeautifulSoup(response.content, "html.parser")
supplementary_links = []
# Look for supplementary materials by keywords or links
for link in soup.find_all("a", href=True):
href = link.get("href")
text = link.get_text().lower()
if "supplementary" in text or "supplemental" in text or "appendix" in text:
full_url = urljoin(publisher_url, href)
supplementary_links.append(full_url)
research_log.append(f"Found supplementary material link: {full_url}")
if not supplementary_links:
log_message = f"No supplementary materials found for DOI {doi}."
research_log.append(log_message)
return '\n'.join(research_log)
# Create output directory
os.makedirs(output_dir, exist_ok=True)
research_log.append(f"Created output directory: {output_dir}")
# Download supplementary materials
downloaded_files = []
for link in supplementary_links:
file_name = os.path.join(output_dir, link.split("/")[-1])
file_response = requests.get(link, headers=headers)
if file_response.status_code == 200:
with open(file_name, "wb") as f:
f.write(file_response.content)
downloaded_files.append(file_name)
research_log.append(f"Downloaded file: {file_name}")
else:
research_log.append(f"Failed to download file from {link}")
if downloaded_files:
research_log.append(f"Successfully downloaded {len(downloaded_files)} file(s).")
else:
research_log.append(f"No files could be downloaded for DOI {doi}.")
return '\n'.join(research_log)
@tool
def query_arxiv(query: str, max_papers: int = 10) -> str:
"""
Query arXiv for papers based on the provided search query.
Args:
query: The search query string
max_papers: The maximum number of papers to retrieve (default: 10)
Returns:
The formatted search results or an error message
"""
import arxiv
try:
client = arxiv.Client()
search = arxiv.Search(query=query, max_results=max_papers, sort_by=arxiv.SortCriterion.Relevance)
results = "\n\n".join([f"Title: {paper.title}\nSummary: {paper.summary}" for paper in client.results(search)])
return results if results else "No papers found on arXiv."
except Exception as e:
return f"Error querying arXiv: {e}"
@tool
def query_scholar(query: str) -> str:
"""
Query Google Scholar for papers based on the provided search query.
Args:
query: The search query string
Returns:
The first search result formatted or an error message
"""
from scholarly import scholarly
try:
search_query = scholarly.search_pubs(query)
result = next(search_query, None)
if result:
return f"Title: {result['bib']['title']}\nYear: {result['bib']['pub_year']}\nVenue: {result['bib']['venue']}\nAbstract: {result['bib']['abstract']}"
else:
return "No results found on Google Scholar."
except Exception as e:
return f"Error querying Google Scholar: {e}"
@tool
def query_pubmed(query: str, max_papers: int = 10, max_retries: int = 3) -> str:
"""
Query PubMed for papers based on the provided search query.
Args:
query: The search query string
max_papers: The maximum number of papers to retrieve (default: 10)
max_retries: Maximum number of retry attempts with modified queries (default: 3)
Returns:
The formatted search results or an error message
"""
from pymed import PubMed
import time
try:
pubmed = PubMed(tool="MyTool", email="your-email@example.com") # Update with a valid email address
# Initial attempt
papers = list(pubmed.query(query, max_results=max_papers))
# Retry with modified queries if no results
retries = 0
while not papers and retries < max_retries:
retries += 1
# Simplify query with each retry by removing the last word
simplified_query = ' '.join(query.split()[:-retries]) if len(query.split()) > retries else query
time.sleep(1) # Add delay between requests
papers = list(pubmed.query(simplified_query, max_results=max_papers))
if papers:
results = "\n\n".join([f"Title: {paper.title}\nAbstract: {paper.abstract}\nJournal: {paper.journal}" for paper in papers])
return results
else:
return "No papers found on PubMed after multiple query attempts."
except Exception as e:
return f"Error querying PubMed: {e}"
@tool
def extract_url_content(url: str) -> str:
"""
Extract the text content of a webpage using requests and BeautifulSoup.
Args:
url: Webpage URL to extract content from
Returns:
Text content of the webpage
"""
try:
response = requests.get(url, headers={'User-Agent': 'Mozilla/5.0'})
# Check if the response is in text format
if 'text/plain' in response.headers.get('Content-Type', '') or 'application/json' in response.headers.get('Content-Type', ''):
return response.text.strip() # Return plain text or JSON response directly
# If it's HTML, use BeautifulSoup to parse
soup = BeautifulSoup(response.text, 'html.parser')
# Try to find main content first, fallback to body
content = soup.find('main') or soup.find('article') or soup.body
# Remove unwanted elements
for element in content(['script', 'style', 'nav', 'header', 'footer', 'aside', 'iframe']):
element.decompose()
# Extract text with better formatting
paragraphs = content.find_all(['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6'])
cleaned_text = []
for p in paragraphs:
text = p.get_text().strip()
if text: # Only add non-empty paragraphs
cleaned_text.append(text)
return '\n\n'.join(cleaned_text)
except Exception as e:
return f"Error extracting content from URL: {str(e)}"
@tool
def extract_pdf_content(url: str) -> str:
"""
Extract the text content of a PDF file given its URL.
Args:
url: URL of the PDF file to extract text from
Returns:
The extracted text content from the PDF
"""
try:
# Check if the URL ends with .pdf
if not url.lower().endswith('.pdf'):
# If not, try to find a PDF link on the page
response = requests.get(url, timeout=30)
if response.status_code == 200:
# Look for PDF links in the HTML content
pdf_links = re.findall(r'href=[\'"]([^\'"]+\.pdf)[\'"]', response.text)
if pdf_links:
# Use the first PDF link found
if not pdf_links[0].startswith('http'):
# Handle relative URLs
base_url = '/'.join(url.split('/')[:3])
url = base_url + pdf_links[0] if pdf_links[0].startswith('/') else base_url + '/' + pdf_links[0]
else:
url = pdf_links[0]
else:
return f"No PDF file found at {url}. Please provide a direct link to a PDF file."
# Download the PDF
response = requests.get(url, timeout=30)
# Check if we actually got a PDF file (by checking content type or magic bytes)
content_type = response.headers.get('Content-Type', '').lower()
if 'application/pdf' not in content_type and not response.content.startswith(b'%PDF'):
return f"The URL did not return a valid PDF file. Content type: {content_type}"
pdf_file = BytesIO(response.content)
# Try with PyPDF2 first
try:
text = ""
pdf_reader = PyPDF2.PdfReader(pdf_file)
for page_num in range(len(pdf_reader.pages)):
page = pdf_reader.pages[page_num]
text += page.extract_text() + "\n\n"
except Exception as e:
print(f"Error extracting text from PDF: {str(e)}")
# Clean up the text
text = re.sub(r'\s+', ' ', text).strip()
if not text:
return "The PDF file did not contain any extractable text. It may be an image-based PDF requiring OCR."
return text
except requests.exceptions.RequestException as e:
return f"Error downloading PDF: {str(e)}"
except Exception as e:
return f"Error extracting text from PDF: {str(e)}"