import requests
import sqlite3
import os
import PyPDF2
from io import BytesIO
from concurrent.futures import ThreadPoolExecutor, as_completed

# Function to download a single PDF
def download_pdf(i):
    url = f'https://eksempel.org/{i}.pdf'
    
    # Check if the PDF has already been downloaded
    c.execute("SELECT * FROM pdfs WHERE id=?", (i,))
    data = c.fetchone()
    if data is not None and data[1] == 1:
        print(f"Skipping {url}, already downloaded.")
        return

    # Download the PDF
    response = requests.get(url, stream=True)
    if response.status_code == 404:
        print(f"{url} not found.")
        return

    # Check if the PDF is not empty
    file = BytesIO(response.content)
    try:
        reader = PyPDF2.PdfFileReader(file)
        if reader.getNumPages() == 0:
            print(f"Skipping {url}, empty file.")
            return
    except PyPDF2.utils.PdfReadError:
        print(f"Skipping {url}, not a valid PDF.")
        return

    # Save the PDF
    with open(f'{i}.pdf', 'wb') as f:
        f.write(response.content)

    # Record the PDF as downloaded
    c.execute("INSERT OR IGNORE INTO pdfs VALUES (?, ?, ?)", (i, url, 1))
    conn.commit()

# Connect to SQLite3 database, create if not exists
conn = sqlite3.connect('pdfs.db')
c = conn.cursor()

# Create table if not exists
c.execute('''
    CREATE TABLE IF NOT EXISTS pdfs
    (id INTEGER PRIMARY KEY, url TEXT, downloaded INTEGER)
''')

# Find the maximum id in the database, and start from there.
# If the database is empty, start from 1.
c.execute("SELECT MAX(id) FROM pdfs")
result = c.fetchone()
start = 1 if result[0] is None else result[0] + 1

# Use ThreadPoolExecutor to download multiple PDFs at the same time
with ThreadPoolExecutor(max_workers=100) as executor:
    futures = {executor.submit(download_pdf, i) for i in range(start, start+1000)}
    for future in as_completed(futures):
        future.result()  # to raise exceptions if any

# Close the database connection
conn.close()