app.py

from pptx import Presentation
import gradio as gr
from pdf2image import convert_from_path
import pdfplumber
from docx import Document
import subprocess
import os
from typing import Optional, List
import string
import random
import re
import requests
from bs4 import BeautifulSoup
import logging
import time
from urllib.parse import urlparse


class URLTextExtractor:
    """
    A comprehensive utility for extracting text content from web pages with advanced features.
    
    Features:
    - Rotating User-Agents to mimic different browsers
    - Robust error handling and retry mechanism
    - Section preservation for maintaining document structure
    - Configurable extraction options
    - Logging support
    
    Attributes:
        USER_AGENTS (list): A comprehensive list of user agent strings to rotate through.
        logger (logging.Logger): Logger for tracking extraction attempts and errors.
    
    Example:
        >>> extractor = URLTextExtractor()
        >>> text = extractor.extract_text_from_url('https://example.com')
        >>> print(text)
    """
    
    # Expanded list of user agents including mobile and less common browsers
    USER_AGENTS = [
        # Desktop Browsers
        'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36',
        'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.1 Safari/605.1.15',
        'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:95.0) Gecko/20100101 Firefox/95.0',
        
        # Mobile Browsers
        'Mozilla/5.0 (iPhone; CPU iPhone OS 14_6 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.1.1 Mobile/15E148 Safari/604.1',
        'Mozilla/5.0 (Linux; Android 10; SM-G970F) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.101 Mobile Safari/537.36',
    ]
    
    def __init__(self, logger=None):
        """
        Initialize the URLTextExtractor.
        
        Args:
            logger (logging.Logger, optional): Custom logger. 
                If not provided, creates a default logger.
        """
        self.logger = logger or self._create_default_logger()
    
    def _create_default_logger(self):
        """
        Create a default logger for tracking extraction process.
        
        Returns:
            logging.Logger: Configured logger instance
        """
        logger = logging.getLogger(__name__)
        logger.setLevel(logging.INFO)
        handler = logging.StreamHandler()
        formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
        handler.setFormatter(formatter)
        logger.addHandler(handler)
        return logger
    
    def _process_element_text(self, element):
        """
        Process text within an element, handling anchor tags specially.
        
        Args:
            element (bs4.element.Tag): BeautifulSoup element to process
        
        Returns:
            str: Processed text with proper spacing
        """
        # Replace anchor tags with spaced text
        for a_tag in element.find_all('a'):
            # Add spaces around the anchor text
            a_tag.replace_with(f' {a_tag.get_text(strip=True)} ')
        
        # Get text with separator
        return element.get_text(separator=' ', strip=True)
    
    def extract_text_from_url(self, url, max_retries=3, preserve_sections=True, 
                               min_section_length=30, allowed_tags=None):
        """
        Extract text content from a given URL with advanced configuration.
        
        Args:
            url (str): The URL of the webpage to extract text from.
            max_retries (int, optional): Maximum number of retry attempts. Defaults to 3.
            preserve_sections (bool, optional): Whether to preserve section separations. Defaults to True.
            min_section_length (int, optional): Minimum length of text sections to include. Defaults to 30.
            allowed_tags (list, optional): Specific HTML tags to extract text from. 
                If None, uses a default set of content-rich tags.
        
        Returns:
            str: Extracted text content from the webpage
        
        Raises:
            ValueError: If URL cannot be fetched after maximum retries
            requests.RequestException: For network-related errors
        
        Examples:
            >>> extractor = URLTextExtractor()
            >>> text = extractor.extract_text_from_url('https://example.com')
            >>> text = extractor.extract_text_from_url('https://example.com', preserve_sections=False)
        """
        # Default allowed tags if not specified
        if allowed_tags is None:
            allowed_tags = ['p', 'div', 'article', 'section', 'main', 
                            'h1', 'h2', 'h3', 'h4', 'h5', 'h6']
        
        # Validate URL
        try:
            parsed_url = urlparse(url)
            if not all([parsed_url.scheme, parsed_url.netloc]):
                # raise ValueError("Invalid URL format")
                return None
        except Exception as e:
            self.logger.error(f"URL parsing error: {e}")
            raise
        
        for attempt in range(max_retries):
            try:
                # Randomly select a user agent
                headers = {
                    'User-Agent': random.choice(self.USER_AGENTS),
                    'Accept-Language': 'en-US,en;q=0.9',
                    'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8'
                }
                
                # Send a GET request to the URL
                response = requests.get(
                    url, 
                    headers=headers, 
                    timeout=10,
                    allow_redirects=True
                )
                
                # Raise an exception for bad status codes
                response.raise_for_status()
                
                # Log successful fetch
                self.logger.info(f"Successfully fetched URL: {url}")
                
                # Parse the HTML content
                soup = BeautifulSoup(response.text, 'html.parser')
                
                # Remove unwanted elements
                for script in soup(["script", "style", "head", "header", "footer", "nav"]):
                    script.decompose()
                
                # Extract text with section preservation
                if preserve_sections:
                    # Extract text from specified tags
                    sections = []
                    for tag in allowed_tags:
                        for element in soup.find_all(tag):
                            # Process element text, handling anchor tags
                            section_text = self._process_element_text(element)
                            
                            # Only add sections meeting minimum length
                            if len(section_text) >= min_section_length:
                                sections.append(section_text)
                    
                    # Join sections with newline
                    text = '\n'.join(sections)
                else:
                    # If not preserving sections, use modified text extraction
                    text = ' '.join(self._process_element_text(element) 
                                    for tag in allowed_tags 
                                    for element in soup.find_all(tag))
                
                # Remove excessive whitespace and empty lines
                text = '\n'.join(line.strip() for line in text.split('\n') if line.strip())
                
                return text
            
            except (requests.RequestException, ValueError) as e:
                # Log error details
                self.logger.warning(f"Attempt {attempt + 1} failed: {e}")
                
                # If it's the last retry, raise the error
                if attempt == max_retries - 1:
                    self.logger.error(f"Failed to fetch URL after {max_retries} attempts")
                    raise ValueError(f"Error fetching URL after {max_retries} attempts: {e}")
                
                # Exponential backoff
                wait_time = 2 ** attempt
                self.logger.info(f"Waiting {wait_time} seconds before retry")
                time.sleep(wait_time)
        
        # Fallback (though this should never be reached due to the raise in the loop)
        return None

def extract_text_from_pptx(file_path):
    prs = Presentation(file_path)
    text_content = []

    for slide in prs.slides:
        slide_text = []
        for shape in slide.shapes:
            if hasattr(shape, "text"):
                slide_text.append(shape.text)
        text_content.append("\n".join(slide_text))

    return "\n\n".join(text_content)


def extract_text_from_ppt(file_path):
    try:
        # Convert PPT to PPTX using unoconv
        pptx_file_path = os.path.splitext(file_path)[0] + ".pptx"
        subprocess.run(["unoconv", "-f", "pptx", file_path], check=True)

        # Extract text from PPTX
        presentation = Presentation(pptx_file_path)
        text_content = []

        for slide in presentation.slides:
            slide_text = []
            for shape in slide.shapes:
                if hasattr(shape, "text"):
                    slide_text.append(shape.text)
            text_content.append("\n".join(slide_text))

        # Remove the converted PPTX file
        os.remove(pptx_file_path)

        return "\n\n".join(text_content)
    except Exception as e:
        print(f"Error extracting text from PPT file: {e}")
        return "Error extracting text from PPT file"


def extract_text_from_ppt_or_pptx(file_path):
    if file_path.endswith(".pptx"):
        return extract_text_from_pptx(file_path)
    elif file_path.endswith(".ppt"):
        return extract_text_from_ppt(file_path)
    else:
        return "Unsupported file type. Please provide a .ppt or .pptx file."


def convert_pdf_to_image(file):
    images = convert_from_path(file)
    return images


def extract_text_from_pdf(file):
    text = ""
    with pdfplumber.open(file) as pdf:
        for page in pdf.pages:
            text += page.extract_text() + "\n"
    return text


def extract_text_from_docx(file):
    text = ""
    doc = Document(file.name)
    for paragraph in doc.paragraphs:
        text += paragraph.text + "\n"
    return text


def convert_doc_to_text(doc_path):
    try:
        subprocess.run(
            ["unoconv", "--format", "txt", doc_path],
            capture_output=True,
            text=True,
            check=True,
        )
        txt_file_path = doc_path.replace(".doc", ".txt")
        with open(txt_file_path, "r") as f:
            text = f.read()
        text = text.lstrip("\ufeff")
        os.remove(txt_file_path)
        return text
    except subprocess.CalledProcessError as e:
        print(f"Error converting {doc_path} to text: {e}")
        return ""


def extract_text_from_doc_or_docx(file):
    if file.name.endswith(".docx"):
        return extract_text_from_docx(file)
    elif file.name.endswith(".doc"):
        return convert_doc_to_text(file.name)
    else:
        return "Unsupported file type. Please upload a .doc or .docx file."


# function that generates a random string
def generate_random_string(length=23):
    characters = string.ascii_letters + string.digits  # Includes letters and digits
    random_string = "".join(random.choice(characters) for _ in range(length))
    return random_string


# function that adds the necessary json fields
def handle_json_output(json_list: list):
    n = len(json_list)
    for i in range(n):
        # not last element
        random_string1 = generate_random_string()
        random_string2 = generate_random_string()
        element = json_list[i]
        front = element["frontText"]
        back = element["backText"]
        element["frontHTML"] = (
            f'<div id="element-richtextarea-{random_string1}" style="position:absolute;left:100px;top:50px;width:800px;height:300px;text-align:center;display:flex;align-items:center;font-size:40px;">'
            f"<p>{front}</p></div>"
        )
        element["backHTML"] = (
            f'<div id="element-richtextarea-{random_string2}" style="position:absolute;left:100px;top:50px;width:800px;height:300px;text-align:center;display:flex;align-items:center;font-size:40px;">'
            f"<p>{back}</p></div>"
        )
        element["termType"] = "basic"
        cloze_matches = re.findall(r"_{2,}", front)
        # match only the first one, if there is multiple don't do anything
        if (cloze_matches != []) & (len(cloze_matches) <= 2):
            # It's a cloze type card
            element["termType"] = "cloze"

            # inject the back in a span format into the front
            def replace_cloze(match):
                return f'</p><p><span class="closure">{back}</span></p><p>'

            front_html = re.sub(r"_{2,}", replace_cloze, front)
            element["frontHTML"] = (
                f'<div id="element-richtextarea-{random_string1}" style="position:absolute;left:100px;top:50px;width:800px;height:300px;text-align:center;display:flex;align-items:center;font-size:40px;">'
                f"<p>{front_html}</p></div>"
            )

            def replace_underscores(match):
                return f" {back} "

            element["frontText"] = re.sub(r"_{2,}", replace_underscores, front)
            element["backText"] = ""

            element["backHTML"] = (
                f'<div id="element-richtextarea-{random_string2}" style="position:absolute;left:100px;top:50px;width:800px;height:300px;text-align:center;display:flex;align-items:center;font-size:40px;">'
                f"<p><br></p></div>"
            )

    return json_list


def sanitize_list_of_lists(text: str) -> Optional[List[List]]:
    left = text.find("[")
    right = text.rfind("]")
    text = text[left : right + 1]
    try:
        # Safely evaluate the string to a Python object
        list_of_lists = eval(text)
        if isinstance(list_of_lists, list):  # Ensure it's a list
            out = []
            try:
                # parse list of lists
                for front, back in list_of_lists:
                    out.append({"frontText": front, "backText": back})
                return handle_json_output(out)
            # errors
            except Exception as e:
                print(e)
                # return anything that was already parsed
                if out != []:
                    return handle_json_output(out)
                # original schedma is not respected
                else:
                    return None
        else:
            print("The evaluated object is not a list.")
            return None
    except Exception as e:
        print(f"Error parsing the list of lists: {e}")
        return None

extractor = URLTextExtractor()
def parse_url(url):
    return extractor.extract_text_from_url(url)

pdf_to_img = gr.Interface(
    convert_pdf_to_image, gr.File(), gr.Gallery(), api_name="pdf_to_img"
)
pdf_to_text = gr.Interface(
    extract_text_from_pdf,
    gr.File(),
    gr.Textbox(placeholder="Extracted text will appear here"),
    api_name="pdf_to_text",
)

doc_or_docx_to_text = gr.Interface(
    extract_text_from_doc_or_docx,
    gr.File(),
    gr.Textbox(placeholder="Extracted text from DOC or DOCX will appear here"),
    api_name="doc_or_docx_to_text",
)

pptx_or_ppt_to_text = gr.Interface(
    extract_text_from_ppt_or_pptx,
    gr.File(),
    gr.Textbox(placeholder="Extracted text from PPTX will appear here"),
    api_name="pptx_or_ppt_to_text",
)

str_to_json = gr.Interface(
    sanitize_list_of_lists,
    gr.Text(),
    gr.JSON(),
    api_name="str_to_json",
    examples=[
        """[
  ["What year was the Carthaginian Empire founded?", "Around 814 BCE"],
  ["Where was the center of the Carthaginian Empire located?", "Carthage, near present-day Tunis, Tunisia"],
  ["Which powerful ancient republic did Carthage have conflicts with?", "The Roman Republic"],
  ["Fill in the blank: Hannibal famously crossed the ________ with war elephants.", "Alps"],
  ["What were the series of conflicts between Carthage and Rome called?", "The Punic Wars"],
  ["Multiple Choice: What was a significant military advantage of Carthage? A) Strong infantry, B) Powerful navy, C) Fortified cities", "B) Powerful navy"],
  ["In what year was Carthage captured and destroyed by Rome?", "146 BCE"],
  ["What did Carthage excel in that allowed it to amass wealth?", "Maritime trade"]
]"""
    ],
)

url_parser = gr.Interface(
    parse_url,
    inputs=["text"],
    outputs=["text"],
    api_name="url_to_text",
)
demo = gr.TabbedInterface(
    [pdf_to_img, pdf_to_text, doc_or_docx_to_text, pptx_or_ppt_to_text, url_parser, str_to_json],
    [
        "PDF to Image",
        "Extract PDF Text",
        "Extract DOC/DOCX Text",
        "Extract PPTX/PPT Text",
        "Extract text from URL",
        "Extract Json",
    ],
)

demo.launch(server_name="0.0.0.0.", server_port=7860, debug=True)