app.py

import os
import subprocess
from playwright.async_api import async_playwright, TimeoutError as PlaywrightTimeoutError
import asyncio
import logging
from urllib.parse import urlparse
import re
from pathlib import Path
from io import BytesIO
import random
import streamlit as st
from bs4 import BeautifulSoup
from PyPDF2 import PdfReader
import zipfile
import tempfile
import mimetypes
import requests

# -------------------- spaCy Model Setup --------------------
import spacy
import spacy.cli
from spacy.language import Language

# Register a dummy factory under the exact key that the transformer model expects.
@Language.factory("spacy-curated-transformers_RobertaTransformer_v1")
def dummy_roberta_transformer(nlp, name):
    # This dummy component simply passes the Doc through.
    def dummy(doc):
        return doc
    return dummy

# Try to load the transformer-based model.
@st.cache_resource
def load_nlp_model():
    try:
        nlp_model = spacy.load("en_core_web_trf")
    except OSError:
        st.write("Model en_core_web_trf not found. Downloading it now...")
        spacy.cli.download("en_core_web_trf")
        try:
            nlp_model = spacy.load("en_core_web_trf")
        except Exception as e:
            st.error(f"Error loading model after download: {e}")
            st.write("Falling back to en_core_web_sm...")
            spacy.cli.download("en_core_web_sm")
            nlp_model = spacy.load("en_core_web_sm")
    return nlp_model

nlp_model = load_nlp_model()

# Also load SentenceTransformer for semantic re-ranking.
from sentence_transformers import SentenceTransformer, util
@st.cache_resource
def load_semantic_model():
    return SentenceTransformer('all-MiniLM-L6-v2')

semantic_model = load_semantic_model()

# -------------------- Transformers Summarization Setup --------------------
from transformers import pipeline
@st.cache_resource
def load_summarizer():
    return pipeline("summarization")

summarizer = load_summarizer()

def summarize_pdf_url(pdf_url):
    """
    Downloads a PDF from the given URL, extracts text using PyPDF2,
    and returns a summary of (up to) the first 3000 characters.
    """
    try:
        with st.spinner("Downloading and processing PDF..."):
            response = requests.get(pdf_url, stream=True)
            temp_pdf = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf")
            with open(temp_pdf.name, "wb") as f:
                f.write(response.content)
            reader = PdfReader(temp_pdf.name)
            text = " ".join([page.extract_text() or "" for page in reader.pages])
            os.remove(temp_pdf.name)
            limited_text = text[:3000]  # Limit text for summarization
            summary = summarizer(limited_text, max_length=200, min_length=50, do_sample=False)
            return summary[0]["summary_text"]
    except Exception as e:
        return f"Error summarizing PDF: {e}"

# -------------------- Google API Setup --------------------
GOOGLE_OAUTH_CONFIG = {
    "web": {
        "client_id": "90798824947-u25obg1q844qeikjoh4jdmi579kn9p1c.apps.googleusercontent.com",
        "project_id": "huggingface-449214",
        "auth_uri": "https://accounts.google.com/o/oauth2/auth",
        "token_uri": "https://oauth2.googleapis.com/token",
        "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
        "client_secret": "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",
        "redirect_uris": ["https://euler314-craw-web.hf.space/"]
    }
}

import google_auth_oauthlib.flow
import googleapiclient.discovery
import google.auth.transport.requests

def get_google_auth_url():
    client_config = GOOGLE_OAUTH_CONFIG["web"]
    flow = google_auth_oauthlib.flow.Flow.from_client_config(
        {"web": client_config},
        scopes=["https://www.googleapis.com/auth/drive.file"]
    )
    flow.redirect_uri = client_config["redirect_uris"][0]
    authorization_url, _ = flow.authorization_url(
        access_type="offline",
        include_granted_scopes="true",
        prompt="consent"
    )
    return authorization_url

def exchange_code_for_credentials(auth_code):
    if not auth_code.strip():
        return None, "No code provided."
    try:
        client_config = GOOGLE_OAUTH_CONFIG["web"]
        flow = google_auth_oauthlib.flow.Flow.from_client_config(
            {"web": client_config},
            scopes=["https://www.googleapis.com/auth/drive.file"]
        )
        flow.redirect_uri = client_config["redirect_uris"][0]
        flow.fetch_token(code=auth_code.strip())
        creds = flow.credentials
        if not creds or not creds.valid:
            return None, "Could not validate credentials. Check code and try again."
        return creds, "Google Sign-In successful!"
    except Exception as e:
        return None, f"Error during token exchange: {e}"
# -------------------- Playwright Setup --------------------
def install_playwright_dependencies():
    os.environ['PLAYWRIGHT_BROWSERS_PATH'] = os.path.expanduser("~/.cache/ms-playwright")
    os.environ['LD_LIBRARY_PATH'] = '/usr/lib/playwright:/usr/lib/x86_64-linux-gnu'
    try:
        subprocess.run(['apt-get', 'update', '-y'], check=True)
        packages = [
            'libnss3', 'libnss3-tools', 'libnspr4', 'libatk1.0-0',
            'libatk-bridge2.0-0', 'libatspi2.0-0', 'libcups2', 'libxcomposite1',
            'libxdamage1', 'libdrm2', 'libgbm1', 'libpango-1.0-0'
        ]
        subprocess.run(['apt-get', 'install', '-y', '--no-install-recommends'] + packages, check=True)
        os.makedirs('/usr/lib/playwright', exist_ok=True)
        symlinks = {
            'libnss3.so': '/usr/lib/x86_64-linux-gnu/libnss3.so',
            'libnssutil3.so': '/usr/lib/x86_64-linux-gnu/libnssutil3.so',
            'libsmime3.so': '/usr/lib/x86_64-linux-gnu/libsmime3.so',
            'libnspr4.so': '/usr/lib/x86_64-linux-gnu/libnspr4.so',
            'libatk-1.0.so.0': '/usr/lib/x86_64-linux-gnu/libatk-1.0.so.0',
            'libatk-bridge-2.0.so.0': '/usr/lib/x86_64-linux-gnu/libatk-bridge-2.0.so.0',
            'libcups.so.2': '/usr/lib/x86_64-linux-gnu/libcups.so.2',
            'libatspi.so.0': '/usr/lib/x86_64-linux-gnu/libatspi.so.0',
            'libXcomposite.so.1': '/usr/lib/x86_64-linux-gnu/libXcomposite.so.1',
            'libXdamage.so.1': '/usr/lib/x86_64-linux-gnu/libXdamage.so.1'
        }
        for link_name, target in symlinks.items():
            link_path = os.path.join('/usr/lib/playwright', link_name)
            if not os.path.exists(link_path):
                os.symlink(target, link_path)
        subprocess.run(['python3', '-m', 'playwright', 'install', 'chromium'], check=True)
        browser_path = os.path.expanduser("~/.cache/ms-playwright")
        os.makedirs(browser_path, exist_ok=True)
        subprocess.run(['chmod', '-R', '755', browser_path], check=True)
    except subprocess.CalledProcessError as e:
        st.error(f"Error installing dependencies: {e}")
    except Exception as e:
        st.error(f"Error: {e}")

# Initialize Playwright dependencies
install_playwright_dependencies()

# -------------------- Logging Setup --------------------
logging.basicConfig(
    filename='advanced_download_log.txt',
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger()

# -------------------- Shared Utils --------------------
USER_AGENTS = [
    'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/115.0.0.0 Safari/537.36',
    'Mozilla/5.0 (Macintosh; Intel Mac OS X 12_6_3) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.4 Safari/605.1.15',
    'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/115.0.0.0 Safari/537.36',
    'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:115.0) Gecko/20100101 Firefox/115.0',
]

def get_random_user_agent():
    return random.choice(USER_AGENTS)

def sizeof_fmt(num, suffix='B'):
    for unit in ['', 'K', 'M', 'G', 'T', 'P', 'E', 'Z']:
        if abs(num) < 1024.0:
            return f"{num:3.1f}{unit}{suffix}"
        num /= 1024.0
    return f"{num:.1f}Y{suffix}"

# ---------- Human-like Interactions -------------
async def human_like_scroll(page):
    scroll_height = await page.evaluate('document.body.scrollHeight')
    viewport_height = await page.evaluate('window.innerHeight')
    current_scroll = 0
    while current_scroll < scroll_height:
        await page.evaluate(f'window.scrollTo(0, {current_scroll})')
        await asyncio.sleep(random.uniform(0.5, 1.5))
        current_scroll += viewport_height * random.uniform(0.5, 1.5)
        scroll_height = await page.evaluate('document.body.scrollHeight')

async def human_like_interactions(page):
    await page.mouse.move(random.randint(0, 1000), random.randint(0, 1000))
    await asyncio.sleep(random.uniform(0.5, 1.5))
    await page.mouse.click(random.randint(0, 1000), random.randint(0, 1000))
    await asyncio.sleep(random.uniform(0.5, 1.5))
    await page.evaluate("window.scrollBy(0, window.innerHeight / 2)")
    await asyncio.sleep(random.uniform(0.5, 1.5))

# ---------- NLP Helpers -------------
def nlp_preprocess(query: str) -> str:
    doc = nlp_model(query)
    tokens = [token.lemma_.lower() for token in doc if not token.is_stop and token.is_alpha]
    processed = " ".join(tokens)
    return processed if processed.strip() else query

def nlp_extract_entities(text: str):
    doc = nlp_model(text)
    return [(ent.text, ent.label_) for ent in doc.ents]

# ---------- AI-enhanced Query Preprocessing -------------
def ai_preprocess_query(query: str) -> str:
    return query
# ---------- Download Manager -------------
class DownloadManager:
    def __init__(self, use_proxy=False, proxy=None, query=None, num_results=5):
        self.use_proxy = use_proxy
        self.proxy = proxy
        self.query = query
        self.num_results = num_results
        self.playwright = None
        self.browser = None
        self.context = None
        self.page = None

    async def __aenter__(self):
        self.playwright = await async_playwright().start()
        opts = {"headless": True}
        if self.use_proxy and self.proxy:
            opts["proxy"] = {"server": self.proxy}
        self.browser = await self.playwright.chromium.launch(**opts)
        self.context = await self.browser.new_context(user_agent=get_random_user_agent())
        self.page = await self.context.new_page()
        await self.page.set_extra_http_headers({
            'Accept-Language': 'en-US,en;q=0.9',
            'Accept-Encoding': 'gzip, deflate, br',
            'Referer': 'https://www.bing.com/'
        })
        return self

    async def __aexit__(self, exc_type, exc_val, exc_tb):
        if self.browser:
            await self.browser.close()
        if self.playwright:
            await self.playwright.stop()

    async def get_file_size(self, url):
        try:
            response = await self.page.request.head(url)
            length = response.headers.get('Content-Length', None)
            if length:
                return sizeof_fmt(int(length))
            else:
                return "Unknown Size"
        except Exception:
            return "Unknown Size"

    async def get_pdf_metadata(self, url):
        try:
            resp = await self.page.request.get(url, timeout=15000)
            if resp.ok:
                content = await resp.body()
                pdf = BytesIO(content)
                reader = PdfReader(pdf)
                return {
                    'Title': reader.metadata.title if reader.metadata.title else 'N/A',
                    'Author': reader.metadata.author if reader.metadata.author else 'N/A',
                    'Pages': len(reader.pages),
                }
            else:
                return {}
        except Exception:
            return {}

    async def search_bing(self):
        if not self.query:
            return [], []
        query = self.query
        if "filetype:pdf" not in query.lower():
            query += " filetype:pdf"
        if "site:" not in query.lower():
            query += " site:edu OR site:arxiv.org OR site:openstax.org"
        query = ai_preprocess_query(query)
        query_processed = nlp_preprocess(query)
        logger.info(f"BING SEARCH NLP: Original='{query}' -> Processed='{query_processed}'")
        
        bing_url = f"https://www.bing.com/search?q={query_processed.replace(' ', '+')}&count={self.num_results}"
        try:
            await self.page.goto(bing_url, timeout=30000)
            await self.page.wait_for_selector('li.b_algo', timeout=30000)
            await human_like_scroll(self.page)
            html = await self.page.content()
            soup = BeautifulSoup(html, 'html.parser')
            raw_results = soup.find_all('li', class_='b_algo')
            url_list = []
            info_list = []
            snippets = []
            
            for r in raw_results:
                link_tag = r.find('a')
                snippet_tag = r.find('p')
                snippet_text = snippet_tag.get_text(strip=True) if snippet_tag else ""
                snippets.append(snippet_text)
                entities = nlp_extract_entities(snippet_text)
                
                if link_tag and 'href' in link_tag.attrs:
                    link_url = link_tag['href']
                    url_list.append(link_url)
                    info_list.append({
                        'url': link_url,
                        'snippet': snippet_text,
                        'entities': entities
                    })
                    if len(url_list) >= self.num_results:
                        break

            query_emb = semantic_model.encode(query, convert_to_tensor=True)
            snippet_embs = semantic_model.encode(snippets, convert_to_tensor=True)
            scores = util.cos_sim(query_emb, snippet_embs)[0]
            sorted_indices = scores.argsort(descending=True).cpu().numpy().tolist()
            sorted_url_list = [url_list[i] for i in sorted_indices]
            sorted_info_list = [info_list[i] for i in sorted_indices]
            
            return sorted_url_list, sorted_info_list
        except PlaywrightTimeoutError:
            logger.error("Bing search timed out.")
            return [], []
        except Exception as e:
            logger.error(f"Bing search error: {e}")
            return [], []

    async def extract_downloadable_files(self, url, custom_ext_list):
        found_files = []
        try:
            await self.page.goto(url, timeout=30000)
            await self.page.wait_for_load_state('networkidle', timeout=30000)
            await human_like_interactions(self.page)
            content = await self.page.content()
            soup = BeautifulSoup(content, 'html.parser')
            
            default_exts = [
                '.pdf', '.docx', '.zip', '.rar', '.exe', '.mp3',
                '.mp4', '.avi', '.mkv', '.png', '.jpg', '.jpeg', '.gif'
            ]
            all_exts = set(default_exts + [ext.strip().lower() for ext in custom_ext_list if ext.strip()])
            
            anchors = soup.find_all('a', href=True)
            for a in anchors:
                href = a['href'].strip()
                if any(href.lower().endswith(ext) for ext in all_exts):
                    if href.startswith('http'):
                        file_url = href
                    elif href.startswith('/'):
                        parsed = urlparse(url)
                        file_url = f"{parsed.scheme}://{parsed.netloc}{href}"
                    else:
                        continue
                        
                    size_str = await self.get_file_size(file_url)
                    meta = {}
                    if file_url.lower().endswith('.pdf'):
                        meta = await self.get_pdf_metadata(file_url)
                        
                    found_files.append({
                        'url': file_url,
                        'filename': os.path.basename(file_url.split('?')[0]),
                        'size': size_str,
                        'metadata': meta
                    })
                    
                elif ("drive.google.com" in href) or ("drive.com" in href):
                    file_id = None
                    for pattern in [
                        r'/file/d/([^/]+)/',
                        r'open\?id=([^&]+)',
                        r'id=([^&]+)'
                    ]:
                        match = re.search(pattern, href)
                        if match:
                            file_id = match.group(1)
                            break
                            
                    if file_id:
                        direct = f"https://drive.google.com/uc?export=download&id={file_id}"
                        filename = f"drive_file_{file_id}"
                        try:
                            resp = await self.page.request.head(direct, timeout=15000)
                            cd = resp.headers.get("Content-Disposition", "")
                            if cd:
                                mt = re.search(r'filename\*?="?([^";]+)', cd)
                                if mt:
                                    filename = mt.group(1).strip('"').strip()
                            else:
                                ctype = resp.headers.get("Content-Type", "")
                                ext_guess = mimetypes.guess_extension(ctype) or ""
                                filename = f"drive_file_{file_id}{ext_guess}"
                        except Exception:
                            pass
                            
                        size_str = await self.get_file_size(direct)
                        found_files.append({
                            'url': direct,
                            'filename': filename,
                            'size': size_str,
                            'metadata': {}
                        })
                        
            return found_files
        except PlaywrightTimeoutError:
            logger.error(f"Timeout extracting from {url}")
            return []
        except Exception as e:
            logger.error(f"Error extracting from {url}: {e}")
            return []
async def download_file(self, file_info, save_dir, referer):
        file_url = file_info['url']
        fname = file_info['filename']
        path = os.path.join(save_dir, fname)
        base, ext = os.path.splitext(fname)
        i = 1
        while os.path.exists(path):
            path = os.path.join(save_dir, f"{base}({i}){ext}")
            i += 1
            
        os.makedirs(save_dir, exist_ok=True)
        try:
            if file_url.lower().endswith(".pdf") and "drive.google.com" not in file_url.lower():
                response = requests.get(file_url, stream=True)
                with open(path, "wb") as f:
                    f.write(response.content)
                logger.info(f"Directly downloaded PDF: {path}")
                return path

            if "drive.google.com" in file_url.lower():
                import gdown
                try:
                    result = gdown.download(file_url, output=path, quiet=False, fuzzy=True)
                    if result is None:
                        logger.error(f"gdown failed to download: {file_url}")
                        return None
                    current_ext = os.path.splitext(path)[1].lower()
                    allowed_exts = {'.pdf', '.jpg', '.jpeg', '.png', '.docx', '.zip', '.rar', '.mp3', '.mp4', '.avi', '.mkv'}
                    if current_ext not in allowed_exts:
                        try:
                            r = requests.head(file_url, allow_redirects=True, timeout=15)
                            ctype = r.headers.get("Content-Type", "")
                            guessed_ext = mimetypes.guess_extension(ctype) or ".pdf"
                        except Exception as e:
                            logger.error(f"Error in HEAD request for extension: {e}")
                            guessed_ext = ".pdf"
                        new_path = os.path.splitext(path)[0] + guessed_ext
                        os.rename(path, new_path)
                        path = new_path
                    logger.info(f"Downloaded using gdown: {path}")
                    return path
                except Exception as e:
                    logger.error(f"Error downloading using gdown: {e}")
                    return None

            headers = {
                'Accept-Language': 'en-US,en;q=0.9',
                'Accept-Encoding': 'gzip, deflate, br',
                'Referer': referer
            }
            await human_like_interactions(self.page)
            resp = await self.page.request.get(file_url, headers=headers, timeout=30000)
            if resp.status == 403:
                logger.error(f"403 Forbidden: {file_url}")
                return None
            if not resp.ok:
                logger.error(f"Failed to download {file_url}: Status {resp.status}")
                return None
            data = await resp.body()
            with open(path, 'wb') as f:
                f.write(data)
            logger.info(f"Downloaded: {path}")
            return path
        except PlaywrightTimeoutError:
            logger.error(f"Timeout downloading {file_url}")
            return None
        except Exception as e:
            logger.error(f"Error downloading {file_url}: {e}")
            return None

    async def deep_search(self, url, custom_ext_list, sublink_limit=2000, max_concurrency=500):
        progress_text = st.empty()
        progress_bar = st.progress(0)
        
        progress_text.text("Analyzing main page...")
        all_files = []
        main_files = await self.extract_downloadable_files(url, custom_ext_list)
        all_files.extend(main_files)
        
        progress_text.text("Getting sublinks...")
        sublinks = await self.get_sublinks(url, sublink_limit)
        total_links = len(sublinks)
        
        progress_text.text(f"Processing {total_links} sublinks...")
        sem = asyncio.Semaphore(max_concurrency)
        
        async def analyze_one_sublink(link, idx):
            async with sem:
                progress_text.text(f"Processing link {idx}/{total_links}: {link}")
                progress_bar.progress(idx/total_links)
                return await self.extract_downloadable_files(link, custom_ext_list)
        
        tasks = [analyze_one_sublink(link, i) for i, link in enumerate(sublinks, 1)]
        sub_results = await asyncio.gather(*tasks)
        
        for sr in sub_results:
            all_files.extend(sr)
            
        unique_map = {f['url']: f for f in all_files}
        combined = list(unique_map.values())
        
        progress_text.text(f"Found {len(combined)} unique files.")
        progress_bar.progress(1.0)
        return combined

    async def get_sublinks(self, url, limit=20000):
        try:
            await self.page.goto(url, timeout=30000)
            content = await self.page.content()
            soup = BeautifulSoup(content, "html.parser")
            links = []
            for a in soup.find_all('a', href=True):
                href = a['href'].strip()
                if href.startswith('http'):
                    links.append(href)
                elif href.startswith('/'):
                    parsed = urlparse(url)
                    links.append(f"{parsed.scheme}://{parsed.netloc}{href}")
            return list(set(links))[:limit]
        except Exception as e:
            logger.error(f"Error getting sublinks: {e}")
            return []

def main():
    st.set_page_config(page_title="Advanced File Downloader", layout="wide")
    
    if 'session_state' not in st.session_state:
        st.session_state.session_state = {
            'discovered_files': [],
            'current_url': None,
            'download_manager': None,
            'google_creds': None
        }

    st.title("Advanced File Downloader")
    
    mode = st.sidebar.radio("Select Mode", ["Manual URL", "Bing Search", "PDF Summarizer"])
    
    with st.sidebar.expander("Advanced Options"):
        custom_extensions = st.text_input(
            "Custom File Extensions",
            placeholder=".csv, .txt, .epub"
        )
        max_concurrency = st.slider(
            "Max Concurrency",
            min_value=1,
            max_value=1000,
            value=200
        )
        use_proxy = st.checkbox("Use Proxy")
        proxy = st.text_input("Proxy URL", placeholder="http://proxy:port")

    # Google OAuth Section
    with st.expander("Google Drive Integration"):
        if st.button("Start Google Sign-In"):
            auth_url = get_google_auth_url()
            st.markdown(f"[Click here to authorize]({auth_url})")
        
        auth_code = st.text_input("Enter authorization code")
        if st.button("Complete Sign-In") and auth_code:
            creds, msg = exchange_code_for_credentials(auth_code)
            st.session_state.session_state['google_creds'] = creds
            st.write(msg)

    if mode == "Manual URL":
        manual_url_mode()
    elif mode == "Bing Search":
        bing_search_mode()
    else:
        pdf_summarizer_mode()

def manual_url_mode():
    st.header("Manual URL Mode")
    
    url = st.text_input("Enter URL", placeholder="https://example.com")
    
    if st.button("Deep Search"):
        if url:
            async def run_deep_search():
                async with DownloadManager(
                    use_proxy=st.session_state.get('use_proxy', False),
                    proxy=st.session_state.get('proxy', None)
                ) as dm:
                    files = await dm.deep_search(
                        url=url,
                        custom_ext_list=st.session_state.get('custom_extensions', '').split(','),
                        max_concurrency=st.session_state.get('max_concurrency', 200)
                    )
                    st.session_state.session_state['discovered_files'] = files
                    st.session_state.session_state['current_url'] = url
                    
                    if files:
                        st.write(f"Found {len(files)} files:")
                        for f in files:
                            st.write(f"- {f['filename']} ({f['size']})")
                    else:
                        st.warning("No files found.")
            
            asyncio.run(run_deep_search())

def bing_search_mode():
    st.header("Bing Search Mode")
    
    query = st.text_input("Enter search query")
    num_results = st.slider("Number of results", 1, 50, 5)
    
    if st.button("Search"):
        if query:
            async def run_search():
                async with DownloadManager(
                    use_proxy=st.session_state.get('use_proxy', False),
                    proxy=st.session_state.get('proxy', None),
                    query=query,
                    num_results=num_results
                ) as dm:
                    urls, info = await dm.search_bing()
                    if urls:
                        st.write("Search Results:")
                        for i, (url, info) in enumerate(zip(urls, info), 1):
                            st.write(f"{i}. {url}")
                            st.write(f"   Snippet: {info['snippet']}")
                    else:
                        st.warning("No results found.")
            
            asyncio.run(run_search())

def pdf_summarizer_mode():
    st.header("PDF Summarizer")
    
    pdf_url = st.text_input("Enter PDF URL")
    
    if st.button("Summarize"):
        if pdf_url:
            summary = summarize_pdf_url(pdf_url)
            st.write("Summary:")
            st.write(summary)

if __name__ == "__main__":
    main()