goodreads-backup/backup.py

"""Backup list of books from Goodreads."""

import configparser
import os
from datetime import date

from playwright.sync_api import Page, Playwright, sync_playwright

import_url = "https://www.goodreads.com/review/import"
wait_mins = 10
refresh_backup = True

today = date.today().isoformat()  # current date in ISO format

config = configparser.ConfigParser()

config_file_path = os.path.expanduser(
    os.path.join(os.getenv("XDG_CONFIG_HOME", "~/.config"), "goodreads", "config")
)
assert os.path.exists(config_file_path)
config.read(os.path.expanduser(config_file_path))
script_dir = os.path.dirname(os.path.abspath(__file__))


def login(page: Page) -> None:
    """Login to Goodreads."""
    page.goto("https://www.goodreads.com/")
    page.get_by_role("link", name="Sign In").click()
    page.get_by_role("button", name="Sign in with email").click()
    page.get_by_label("Email").fill(config.get("login", "email"))
    page.get_by_label("Password").fill(config.get("login", "password"))
    page.get_by_label("Keep me signed in").click()
    page.get_by_label("Sign in").click()


def navigate_to_import_and_export(page: Page) -> None:
    """Navigate to import and export."""
    page.goto("https://www.goodreads.com/")
    page.get_by_role("link", name="My Books").click()
    page.get_by_role("link", name="Import and export").click()


def run_backup(page: Page) -> None:
    """Run backup."""
    page.goto(import_url)

    if refresh_backup:
        print("backup requested")
        page.get_by_role("button", name="Export Library").click()
        print(f"waiting for {wait_mins} minutes")

        page.wait_for_timeout(wait_mins * 60 * 1000)

        print("reloading page")
        page.reload()

    print("download export")
    export_link = page.get_by_role("link", name="Your export from")
    print(export_link.text_content())
    with page.expect_download() as download_info:
        page.get_by_role("link", name="Your export from").click()
    download = download_info.value
    backup_dir = config.get("backup", "dir")

    save_to = os.path.join(backup_dir, f"{today}_goodreads_library_export.csv")
    download.save_as(save_to)


def run(playwright: Playwright) -> None:
    """Download export."""
    browser = playwright.chromium.launch(headless=False)
    auth_json = os.path.join(script_dir, "auth.json")
    context = browser.new_context(storage_state=auth_json)
    page = context.new_page()

    # login(page)

    page.goto("https://www.goodreads.com/")

    run_backup(page)

    page.close()

    context.storage_state(path=auth_json)
    context.close()
    browser.close()


with sync_playwright() as playwright:
    run(playwright)