feat: Add Bookys scraper alongside EBoek with startup mode chooser

Adds a second scraper for bookys-ebooks.com that harvests direct 1fichier
download URLs into a .txt file. The existing EBoek scraper core is unchanged;
only the GUI gained a Mode menu entry to switch between the two.

Bookys specifics:
- Visible (never headless) browser, since the site is behind Cloudflare.
  The challenge is auto-detected: the scraper polls for real listing items
  and continues on its own once it clears, no button needed.
- Persistent Chrome profile in ~/.eboek_scraper/bookys_chrome_profile so the
  Cloudflare clearance cookie survives between runs. Removed the hardcoded
  user-agent override, which claimed Chrome/120 against a real Chrome/150 and
  invalidated that cookie on every launch.
- Pop-up ads are avoided by never clicking host links: hrefs are read and
  navigated to directly, with window.open neutered via CDP as a backstop.
- Host links are read with textContent rather than Selenium's .text, which
  returns empty for these elements because they sit in a collapsed container.
- Only 1fichier hosts are followed; other hosts on a page are ignored.

Build configs list the new modules as hidden imports in all four places.
gui_main imports the Bookys window lazily, so PyInstaller's static analysis
would otherwise miss it and the .exe would crash on switching modes.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NVJc2XuTvqBHSuqotetseS

scripts/build_exe.bat

@@ -43,6 +43,10 @@ python -m pyinstaller --onefile --windowed --name "EBoek_Scraper" ^
     --hidden-import "gui.login_dialog" ^
     --hidden-import "gui.progress_dialog" ^
     --hidden-import "utils.validators" ^
+    --hidden-import "core.bookys_scraper" ^
+    --hidden-import "core.bookys_thread" ^
+    --hidden-import "gui.bookys_window" ^
+    --hidden-import "gui.startup_dialog" ^
     --exclude-module "tkinter" ^
     --exclude-module "matplotlib" ^
     ..\gui_main.py