tools.py

from langchain_community.document_loaders.wikipedia import WikipediaLoader
from langchain_community.document_loaders import YoutubeLoader
from langchain_community.document_loaders import ArxivLoader
from langchain_community.tools.tavily_search import TavilySearchResults
from langchain_core.tools import tool


@tool
def youtube_video_loader(url: str) -> str:
    """Load a Youtube video and return the transcript.
    Args:
        url (str): The Youtube video URL."""
    loader = YoutubeLoader.from_youtube_url(str(url), add_video_info=False)
    docs = loader.load()
    print(docs)
    return "\n\n".join(map(repr, docs))


@tool
def multiply(x: int, y: int) -> int:
    """Multiply two numbers."""
    return x * y


@tool
def add(x: int, y: int) -> int:
    """Add two numbers."""
    return x + y


@tool
def subtract(x: int, y: int) -> int:
    """Subtract two numbers."""
    return x - y


@tool
def divide(x: int, y: int) -> float:
    """Divide two numbers."""
    return x / y


@tool
def modulus(x: int, y: int) -> int:
    """Get the modulus of two numbers."""
    return x % y


@tool
def wiki_search(query: str) -> str:
    """Search Wikipedia for a query. Returns the first result 3.
    Args:
        query (str): The search query."""
    loader = WikipediaLoader(query=query, load_max_docs=3)
    docs = loader.load()

    if docs:
        # For each document Extract the content and metadata from the documents
        results = []
        for doc in docs:
            content = doc.page_content
            metadata = doc.metadata
            # Handle missing metadata fields gracefully
            result = {
                "source": metadata.get('source', 'Unknown source'),
                "page": metadata.get('page', 'Unknown page'),
                "content": content
            }
            results.append(result)

    return results.__str__() if results else "No results found."


@tool
def arxiv_search(query: str) -> str:
    """Search Arxiv for a query. Returns the first result 3.
    Args:
        query (str): The search query."""
    loader = ArxivLoader(query=query, load_max_docs=3)
    docs = loader.load()

    if docs:
        # For each document Extract the content and metadata from the documents
        results = []
        for doc in docs:
            content = doc.page_content
            metadata = doc.metadata
            # Format the result as json object
            result = {
                "source": metadata['source'],
                "page": metadata['page'],
                "content": content[:1000]  # Limit content to 1000 characters
            }
            results.append(result)

    return results.__str__() if results else "No results found."


@tool
def tavily_search(query: str) -> str:
    """Search Tavily for a query. Returns the first result 3.
    Args:
        query (str): The search query."""
    docs = TavilySearchResults(max_results=2).invoke(query)

    if docs:
        # For each document Extract the content and metadata from the documents
        results = []
        for doc in docs:
            content = doc['content']
            titel = doc['title']
            url = doc['url']
            # Format the result as json object
            result = {
                "source": url,
                "titel": titel,
                "content": content  # Limit content to 1000 characters
            }
            results.append(result)

    return results.__str__() if results else "No results found."


@tool
def wiki_loader(query: str, lang: str='en', load_max_docs: int=3):
    """
    Fetches content from Wikipedia based on a given query.

    Parameters:
    - query (str): The search query for Wikipedia.
    - lang (str): The language of the Wikipedia to search in. Default is 'en'.
    - load_max_docs (int): The maximum number of documents to load. Default is 2.

    Returns:
    - list: A list of documents containing the fetched Wikipedia content.
    """
    try:
        # Initialize the WikipediaLoader with the given query, language, and max documents
        loader = WikipediaLoader(query=query, lang=lang, load_max_docs=load_max_docs)
        # Load the documents
        documents = loader.load()
        result = []
        for doc in documents:
            doc_dict = {
                "source": doc.metadata.get('source', 'Unknown source'),
                "page": doc.metadata.get('page', 'Unknown page'),
                "content": doc.page_content
            }
            result.append(doc_dict)
            print(result)
        return result
    except Exception as e:
        print(f"An error occurred: {e}")
        return []