DOWNLOAD FULL - Metoda 2.py

DOWNLOAD FULL - Metoda 2.py

# -*- coding: utf-8 -*-
"""
DOWNLOAD FULL - Metoda 1
========================
Bazat pe structura din "Claude-FINAL 15 ... Firefox.py", dar SINGURA diferenta
este METODA de preluare a datelor: NU se mai descarca PDF (are limita). In schimb,
pentru fiecare pagina se preia IMAGINEA SURSA din browser (METODA 1 = fetch la
blob-ul <img class="page-canvas">), apoi toate imaginile unui document se pun
intr-un PDF.

Stocare (ca scriptul mare - fara spatiu pe D:):
  - imaginile (staging/backup):  g:\\Temporare\\<Colectie>\\<Document>\\pageNNNN.jpg
  - PDF final per document:      G:\\<Colectie>\\<Document>.pdf
  - state de resume:             d:\\TEST\\arcanum_capture\\state.json

Resume: la repornire se sar colectiile/documentele deja terminate (din state.json)
si, in plus, paginile deja salvate pe disc -> un document intrerupt se reia de unde
a ramas.

Login: copiaza profilul Firefox activ in temp (cookie-uri => deja logat);
Firefox-ul tau normal ramane deschis si neatins.
"""

import os
import re
import sys
import json
import time
import glob
import base64
import shutil
import tempfile
from datetime import datetime, date, time as dtime

# consola Windows e cp1252 -> titlurile cu ș/ţ/etc ar crapa la print
try:
    sys.stdout.reconfigure(encoding="utf-8", errors="replace")
    sys.stderr.reconfigure(encoding="utf-8", errors="replace")
except Exception:
    pass

from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.firefox.options import Options as FirefoxOptions
from selenium.webdriver.firefox.service import Service as FirefoxService
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.common.exceptions import WebDriverException
from PIL import Image

# ======================= CONFIG =======================
ADDITIONAL_COLLECTIONS = [
    # "https://adt.arcanum.com/ro/collection/FilmeNoi/",
    # "https://adt.arcanum.com/ro/collection/ITTrends/",
    "https://adt.arcanum.com/ro/collection/SzatmariMuzeumKiadvanyai_Evkonyv_ADT/",
    "https://adt.arcanum.com/ro/collection/Afirmarea/",
    "https://adt.arcanum.com/ro/collection/AdevarulJurnalAradean/",
    "https://adt.arcanum.com/ro/collection/AdevarulJurnalAradean/?decade=1990#collection-contents",
    "https://adt.arcanum.com/ro/collection/AdevarulJurnalAradean/?decade=2000#collection-contents",
    "https://adt.arcanum.com/ro/collection/AdevarulJurnalAradean/?decade=2010#collection-contents",
    "https://adt.arcanum.com/ro/collection/AdevarulJurnalAradean/?decade=2020#collection-contents",
    "https://adt.arcanum.com/ro/collection/Timpul/",
    "https://adt.arcanum.com/ro/collection/Timpul/?decade=2000#collection-contents",
    "https://adt.arcanum.com/ro/collection/Carpatii/",
    "https://adt.arcanum.com/ro/collection/CurierulRecreatiilorIntelectuale/",
    "https://adt.arcanum.com/ro/collection/RadioRomania/",
    "https://adt.arcanum.com/ro/collection/RevistaVanatorilor/",
    "https://adt.arcanum.com/ro/collection/CurierulFinanciar/",
    "https://adt.arcanum.com/ro/collection/EWeekRomania/",
    "https://adt.arcanum.com/ro/collection/EvenimentulSibian/",
    "https://adt.arcanum.com/ro/collection/EvenimentulSibian/?decade=2000#collection-contents",
    "https://adt.arcanum.com/ro/collection/NazuintaZilah/",
    "https://adt.arcanum.com/ro/collection/NazuintaZilah/?decade=1970#collection-contents",
    "https://adt.arcanum.com/ro/collection/NazuintaZilah/?decade=1980#collection-contents",
    "https://adt.arcanum.com/ro/collection/RevistaIstoricaRomana/",

]

G_ROOT     = "G:\\"                                       # PDF-urile finale: G:\<Colectie>\<Document>.pdf
TEMP_ROOT  = r"g:\Temporare"                              # imaginile (staging): g:\Temporare\<Colectie>\<Document>\
STATE_PATH = r"d:\TEST\arcanum_capture\state.json"        # resume

PAGE_WAIT = 4      # secunde de asteptare intre pagini (cerinta)
PDF_WAIT  = 120    # 2 minute pauza dupa PDF-ul fiecarui document

# inchidere automata in fereastra 03:40 - 04:00 (la 04:00 porneste celalalt script)
SHUTDOWN_START = dtime(3, 40)
SHUTDOWN_END   = dtime(4, 0)

# --- mod test (dezactivat: rulam complet pe toate colectiile) ---
TEST_MODE      = False  # False = rulare completa pe toate documentele/paginile
TEST_MAX_DOCS  = 1
TEST_MAX_PAGES = 3
if TEST_MODE:
    PDF_WAIT = 5
# ======================================================

SKIP_DIRS = {
    "cache2", "startupCache", "shader-cache", "OfflineCache", "thumbnails",
    "crashes", "datareporting", "saved-telemetry-pings", "minidumps",
    "security_state", "settings", "gmp", "gmp-gmpopenh264", "gmp-widevinecdm",
}
IMG_EXT = ("jpg", "png", "webp")


class ScheduledStop(Exception):
    """Oprire programata (fereastra 03:40-04:00)."""


def in_shutdown_window():
    now = datetime.now().time()
    return SHUTDOWN_START <= now < SHUTDOWN_END


def check_schedule():
    if in_shutdown_window():
        raise ScheduledStop()


# ----------------------- state / resume -----------------------
# Format (ca scriptul mare):
# {
#   "date": "YYYY-MM-DD",
#   "count": <suma paginilor descarcate>,
#   "downloaded_issues": [
#       {"url","title","pages","completed_at","last_successful_segment_end","total_pages"}
#   ]
# }
def load_state():
    if os.path.exists(STATE_PATH):
        try:
            with open(STATE_PATH, "r", encoding="utf-8") as fh:
                s = json.load(fh)
            if isinstance(s, dict):
                s.setdefault("downloaded_issues", [])
                return s
        except Exception:
            pass
    return {"date": date.today().isoformat(), "count": 0, "downloaded_issues": []}


def save_state(state):
    state["date"] = date.today().isoformat()
    state["count"] = sum(int(it.get("pages", 0)) for it in state["downloaded_issues"])
    os.makedirs(os.path.dirname(STATE_PATH), exist_ok=True)
    tmp = STATE_PATH + ".tmp"
    with open(tmp, "w", encoding="utf-8") as fh:
        json.dump(state, fh, ensure_ascii=False, indent=2)
    os.replace(tmp, STATE_PATH)


def issue_url_norm(view_url):
    return view_url.rstrip("/") + "/"


def get_issue(state, view_url):
    u = issue_url_norm(view_url)
    for it in state["downloaded_issues"]:
        if it.get("url") == u:
            return it
    return None


def issue_is_complete(entry):
    # complet = PDF-ul a fost facut cu succes (completed_at setat)
    return entry is not None and bool(entry.get("completed_at"))


def upsert_issue(state, view_url, title, pages, total_pages, last_idx, completed=False, pdf=None):
    u = issue_url_norm(view_url)
    entry = get_issue(state, view_url)
    if entry is None:
        entry = {}
        state["downloaded_issues"].append(entry)
    # ordine ca in exemplu + campul "pdf" (calea PDF-ului facut)
    entry.clear()
    entry["url"] = u
    entry["title"] = title
    entry["pages"] = pages
    entry["completed_at"] = datetime.now().isoformat(timespec="seconds") if completed else None
    entry["last_successful_segment_end"] = last_idx
    entry["total_pages"] = total_pages
    entry["pdf"] = pdf
    save_state(state)
    return entry


# ----------------------- login / profil -----------------------
def find_active_profile():
    base = os.path.join(os.environ["APPDATA"], r"Mozilla\Firefox\Profiles")
    cands = glob.glob(os.path.join(base, "*.default-release")) \
        or glob.glob(os.path.join(base, "*.default")) \
        or [p for p in glob.glob(os.path.join(base, "*")) if os.path.isdir(p)]
    if not cands:
        raise RuntimeError("Nu am gasit niciun profil Firefox.")
    cands.sort(key=lambda p: os.path.getmtime(os.path.join(p, "cookies.sqlite"))
               if os.path.exists(os.path.join(p, "cookies.sqlite")) else 0, reverse=True)
    return cands[0]


# doar fisierele necesare pentru sesiunea logata (profilul complet poate avea sute de MB!)
ESSENTIAL_FILES = [
    "cookies.sqlite", "cookies.sqlite-wal", "cookies.sqlite-shm",
    "key4.db", "logins.json", "cert9.db", "prefs.js", "permissions.sqlite",
    "webappsstore.sqlite", "webappsstore.sqlite-wal", "webappsstore.sqlite-shm",
    "handlers.json", "containers.json",
]


def copy_profile(src):
    dst = tempfile.mkdtemp(prefix="ff_dl1_")
    for name in ESSENTIAL_FILES:
        s = os.path.join(src, name)
        if not os.path.exists(s):
            continue
        d = os.path.join(dst, name)
        try:
            shutil.copy2(s, d)
        except Exception as e:
            print(f"     ({name}: copy2 a esuat - {e}; incerc citire bruta)")
            try:
                with open(s, "rb") as fh:
                    data = fh.read()
                with open(d, "wb") as fh:
                    fh.write(data)
            except Exception as e2:
                print(f"     ({name}: sarit - {e2})")
    return dst


def _gecko_service():
    # fixam calea geckodriver ca Selenium sa NU mai porneasca Selenium Manager (care poate atinge reteaua si bloca)
    path = shutil.which("geckodriver") or r"C:\Windows\geckodriver.exe"
    try:
        if os.path.exists(path):
            print(f"   geckodriver: {path}")
            return FirefoxService(executable_path=path)
    except Exception as e:
        print(f"   (nu pot fixa geckodriver explicit: {e})")
    return FirefoxService()


def _firefox_binary():
    cands = [
        r"C:\Program Files\Mozilla Firefox\firefox.exe",
        r"C:\Program Files (x86)\Mozilla Firefox\firefox.exe",
        os.path.join(os.environ.get("LOCALAPPDATA", ""), r"Mozilla Firefox\firefox.exe"),
    ]
    for c in cands:
        if c and os.path.exists(c):
            return c
    try:
        import winreg
        with winreg.OpenKey(winreg.HKEY_LOCAL_MACHINE,
                            r"SOFTWARE\Microsoft\Windows\CurrentVersion\App Paths\firefox.exe") as k:
            val, _ = winreg.QueryValueEx(k, None)
            if val and os.path.exists(val):
                return val
    except Exception:
        pass
    return None


def cleanup_stale_automation():
    """Curata procesele de automatizare ramase din rulari anterioare (crapate/oprite),
    ca sa nu mai fie nevoie de RESTART la PC. NU atinge Firefox-ul normal al userului:
    omoara doar firefox.exe lansat cu profilul nostru temporar 'ff_dl1_' + toate geckodriver."""
    print("Curat geckodriver + Firefox de automatizare ramase (NU si Firefox-ul tau normal)...")
    try:
        import subprocess
        flags = getattr(subprocess, "CREATE_NO_WINDOW", 0)
        # /T omoara si procesele-copil (Firefox-ul de automatizare pornit de geckodriver);
        # Firefox-ul tau normal NU e copil de geckodriver, deci ramane neatins.
        subprocess.run(["taskkill", "/F", "/T", "/IM", "geckodriver.exe"],
                       timeout=15, capture_output=True, creationflags=flags)
    except Exception as e:
        print(f"   (taskkill geckodriver: {e})")
    # sterge profilele temporare vechi (nu mai sunt folosite)
    removed = 0
    for d in glob.glob(os.path.join(tempfile.gettempdir(), "ff_dl1_*")):
        try:
            shutil.rmtree(d, ignore_errors=True)
            removed += 1
        except Exception:
            pass
    print(f"   curatat. (profile temporare vechi sterse: {removed})")


def start_firefox(profile_dir):
    opts = FirefoxOptions()
    opts.add_argument("--no-remote")
    opts.add_argument("-profile")
    opts.add_argument(profile_dir)
    opts.set_preference("pdfjs.disabled", False)
    opts.set_preference("browser.tabs.remote.autostart", False)
    opts.set_preference("general.useragent.override",
        "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:115.0) Gecko/20100101 Firefox/115.0")
    binpath = _firefox_binary()
    if binpath:
        opts.binary_location = binpath
        print(f"   Firefox binar: {binpath}")
    else:
        print("   (nu am gasit firefox.exe explicit; las geckodriver sa caute)")
    print("   lansez Firefox prin geckodriver (poate dura 5-15s)...")
    drv = webdriver.Firefox(options=opts, service=_gecko_service())
    drv.set_window_size(1500, 1200)
    drv.set_script_timeout(60)
    print("   Firefox a pornit cu succes.")
    return drv


class Browser:
    """Manager Firefox cu auto-recuperare daca fereastra se inchide din greseala."""
    def __init__(self):
        self.drv = None
        self.tmp = None

    def start(self):
        t = time.time()
        print("   caut profilul Firefox activ...")
        src = find_active_profile()
        print(f"   profil sursa: {src}")
        print("   copiez fisierele de login (cateva MB)...")
        self.tmp = copy_profile(src)
        print(f"   profil copiat in {time.time() - t:.1f}s -> {self.tmp}")
        self.drv = start_firefox(self.tmp)

    def quit(self):
        try:
            if self.drv:
                self.drv.quit()
        except Exception:
            pass
        self.drv = None
        if self.tmp:
            shutil.rmtree(self.tmp, ignore_errors=True)
            self.tmp = None

    def alive(self):
        try:
            _ = self.drv.current_url
            return True
        except Exception:
            return False

    def restart(self):
        print("   >>> repornesc Firefox (recuperare)...")
        self.quit()
        time.sleep(3)
        self.start()
        print("   >>> Firefox repornit, continui de unde am ramas.")


def retry_browser(br, fn, what, retries=6):
    """Ruleaza fn(); daca browserul a fost inchis/pierdut, repornește Firefox si reincearca."""
    for attempt in range(1, retries + 1):
        try:
            return fn()
        except WebDriverException as e:
            msg = (str(e) or type(e).__name__).splitlines()[0]
            print(f"   !! eroare browser la {what} (incercare {attempt}/{retries}): {msg[:120]}")
            if attempt >= retries:
                raise
            if not br.alive():
                print("   ... fereastra inchisa/pierduta -> recuperare ...")
                for tryno in range(3):
                    try:
                        br.restart()
                        break
                    except Exception as e2:
                        print(f"   restart esuat ({e2}); reincerc in 5s...")
                        time.sleep(5)
            else:
                time.sleep(2)
    return None


# ----------------------- JS -----------------------
JS_PAGECOUNT = "return document.querySelectorAll('ul.thumbs li.thumb-item').length;"

JS_BIGIMG = r"""
var imgs = Array.from(document.querySelectorAll('img.page-canvas, img[src^="blob:"]'))
             .filter(function(i){ return i.naturalWidth > 0; });
if(!imgs.length){ return null; }
imgs.sort(function(a,b){ return b.naturalWidth*b.naturalHeight - a.naturalWidth*a.naturalHeight; });
var i = imgs[0];
return {nw:i.naturalWidth, nh:i.naturalHeight};
"""

JS_GRAB_BLOB = r"""
var cb = arguments[arguments.length-1];
var imgs = Array.from(document.querySelectorAll('img.page-canvas, img[src^="blob:"]'))
             .filter(function(i){ return i.naturalWidth > 0; });
if(!imgs.length){ cb({ok:false, err:'no img'}); return; }
imgs.sort(function(a,b){ return b.naturalWidth*b.naturalHeight - a.naturalWidth*a.naturalHeight; });
var img = imgs[0];
fetch(img.src).then(function(r){return r.blob();}).then(function(b){
  var fr = new FileReader();
  fr.onload = function(){ cb({ok:true, ct:b.type, nw:img.naturalWidth, nh:img.naturalHeight, data:fr.result}); };
  fr.onerror = function(){ cb({ok:false, err:'reader'}); };
  fr.readAsDataURL(b);
}).catch(function(e){ cb({ok:false, err:String(e)}); });
"""


def save_dataurl(data_url, path):
    # scriere atomica: .part -> rename, ca o oprire brusca sa nu lase imagine corupta
    b64 = data_url.split(",", 1)[1]
    tmp = path + ".part"
    with open(tmp, "wb") as fh:
        fh.write(base64.b64decode(b64))
    os.replace(tmp, path)


# ----------------------- logica colectie -----------------------
def collection_name(coll_url):
    return coll_url.rstrip("/").split("/collection/")[-1].split("/")[0]


def doc_name(view_url):
    return view_url.split("/view/")[-1].strip("/").split("/")[0]


def extract_document_urls(drv):
    """Ca in scriptul mare: linkuri unice de /view/ din colectie (documentele/anii)."""
    try:
        WebDriverWait(drv, 30).until(
            EC.presence_of_element_located((By.CSS_SELECTOR, 'a[href*="/view/"]')))
    except Exception:
        pass
    time.sleep(2)
    anchors = drv.find_elements(By.CSS_SELECTOR, 'li.list-group-item a[href*="/view/"]')
    if not anchors:
        anchors = drv.find_elements(By.CSS_SELECTOR, 'a[href*="/view/"]')
    seen, unique = set(), []
    for a in anchors:
        href = a.get_attribute("href")
        if href and "/view/" in href:
            norm = href.split("?")[0].rstrip("/")
            if norm not in seen:
                seen.add(norm)
                unique.append(norm)
    return unique


def existing_page_file(stage_dir, pg):
    for ext in IMG_EXT:
        p = os.path.join(stage_dir, f"page{pg:04d}.{ext}")
        if os.path.exists(p) and os.path.getsize(p) > 1024:
            return p
    return None


def collect_page_files(stage_dir):
    files = []
    for ext in IMG_EXT:
        files += glob.glob(os.path.join(stage_dir, f"page*.{ext}"))
    files.sort(key=lambda p: os.path.basename(p))
    return files


def page_index(path):
    m = re.search(r"page(\d+)\.", os.path.basename(path))
    return int(m.group(1)) if m else -1


def get_issue_title(drv):
    """Titlul din breadcrumb activ (ex: 'Filme Noi, 1971 (nr. 1-10)1971 / nr. 1')."""
    for _ in range(20):
        try:
            t = drv.find_element(By.CSS_SELECTOR, "li.breadcrumb-item.active").text.strip()
            if t:
                return t
        except Exception:
            pass
        time.sleep(0.5)
    return ""


def wait_for_page_image(drv, timeout=30):
    end = time.time() + timeout
    while time.time() < end:
        info = drv.execute_script(JS_BIGIMG)
        if info and info["nw"] > 600:
            return (info["nw"], info["nh"])
        time.sleep(0.5)
    return None


def capture_document(br, view_url, stage_dir, state):
    name = doc_name(view_url)
    print(f"\n=== DOCUMENT: {name} ===")

    def _open():
        drv = br.drv
        drv.get(view_url + "/?pg=0&layout=s")
        WebDriverWait(drv, 40).until(EC.presence_of_element_located((By.TAG_NAME, "body")))
        title = get_issue_title(drv)
        total = 0
        for _ in range(40):
            total = drv.execute_script(JS_PAGECOUNT) or 0
            if total > 0:
                break
            time.sleep(1)
        return title, total

    title, total = retry_browser(br, _open, f"deschidere {name}")
    if not total:
        print("   !! Nu am putut afla numarul de pagini, sar peste document.")
        return None
    print(f"   titlu: {title}")
    print(f"   pagini in document (total_pages): {total}")
    total_pages = total

    if TEST_MODE:
        total = min(total, TEST_MAX_PAGES)
        print(f"   [TEST] capturez doar primele {total} pagini")

    os.makedirs(stage_dir, exist_ok=True)
    pages_done = 0
    last_idx = -1
    for pg in range(total):
        check_schedule()   # oprire automata 03:40-04:00
        # RESUME: pagina deja salvata -> sarim (dar o numaram)
        ex = existing_page_file(stage_dir, pg)
        if ex:
            print(f"   pg {pg:04d}: deja exista ({os.path.basename(ex)}), sar")
            pages_done += 1
            last_idx = pg
            upsert_issue(state, view_url, title, pages_done, total_pages, last_idx)
            continue

        def _capture():
            drv = br.drv
            drv.get(f"{view_url}/?pg={pg}&layout=s")
            time.sleep(PAGE_WAIT)
            wait_for_page_image(drv, timeout=30)
            return drv.execute_async_script(JS_GRAB_BLOB)

        res = None
        for attempt in range(3):
            res = retry_browser(br, _capture, f"pagina {pg} din {name}")
            if res and res.get("ok"):
                break
            print(f"   pg {pg:04d}: fetch nereusit (incercare {attempt + 1}/3), reincerc...")
            time.sleep(2)
        if not res or not res.get("ok"):
            print(f"   pg {pg:04d}: ESEC final ({res.get('err') if res else 'None'}) - "
                  f"se reia la urmatoarea rulare")
            continue
        ct = res.get("ct", "")
        ext = {"image/jpeg": "jpg", "image/png": "png", "image/webp": "webp"}.get(ct, "jpg")
        img_path = os.path.join(stage_dir, f"page{pg:04d}.{ext}")
        save_dataurl(res["data"], img_path)
        pages_done += 1
        last_idx = pg
        print(f"   pg {pg:04d}: OK  {res.get('nw')}x{res.get('nh')}  -> {os.path.basename(img_path)}")
        # SALVAM PROGRESUL DUPA FIECARE PAGINA (resume exact de unde s-a oprit)
        upsert_issue(state, view_url, title, pages_done, total_pages, last_idx)

    complete = pages_done >= total_pages
    return {"title": title, "total_pages": total_pages,
            "pages_done": pages_done, "complete": complete}


def open_image_robust(p, retries=4):
    """Deschide o imagine cu reincercari (trece peste blip-uri de I/O pe G:)."""
    for attempt in range(retries):
        try:
            im = Image.open(p)
            im.load()                 # forteaza citirea completa (prinde I/O tranzitoriu)
            return im.convert("RGB")
        except Exception as e:
            if attempt == retries - 1:
                print(f"   !! pagina ilizibila dupa {retries} incercari: {os.path.basename(p)} ({e})")
            time.sleep(1.0)
    return None


def build_pdf(image_paths, pdf_path, total_pages):
    # nu facem PDF daca lipsesc pagini
    if total_pages and len(image_paths) < total_pages:
        print(f"   (PDF amanat: doar {len(image_paths)}/{total_pages} pagini pe disc)")
        return False
    if not image_paths:
        print("   (fara imagini, nu fac PDF)")
        return False
    # verificare rapida: toate prezente si non-goale (fara a le decoda)
    for p in image_paths:
        if not os.path.exists(p) or os.path.getsize(p) < 1024:
            print(f"   !! pagina lipsa/mica: {os.path.basename(p)} - NU fac PDF acum")
            return False

    os.makedirs(os.path.dirname(pdf_path), exist_ok=True)
    tmp = pdf_path + ".part"

    # 1) img2pdf = inglobeaza JPEG-urile direct, FARA decodare -> memorie/CPU minime
    try:
        import img2pdf
        try:
            layout = img2pdf.get_fixed_dpi_layout_fun((200, 200))  # pagini la 200 DPI
        except Exception:
            layout = None
        for attempt in range(3):
            try:
                with open(tmp, "wb") as f:
                    if layout:
                        f.write(img2pdf.convert(image_paths, layout_fun=layout))
                    else:
                        f.write(img2pdf.convert(image_paths))
                os.replace(tmp, pdf_path)
                print(f"   PDF salvat (img2pdf): {pdf_path}  ({len(image_paths)} pagini)")
                return True
            except Exception as e:
                print(f"   img2pdf incercare {attempt + 1}/3 esuata: {str(e)[:120]}")
                time.sleep(2)
    except Exception as e:
        print(f"   (img2pdf indisponibil: {e}) - folosesc PIL")

    # 2) fallback PIL (decodeaza in memorie - mai greu)
    imgs = []
    for p in image_paths:
        im = open_image_robust(p)
        if im is None:
            print("   !! NU fac PDF (o pagina e ilizibila) - se reincearca la urmatoarea rulare")
            return False
        imgs.append(im)
    imgs[0].save(tmp, "PDF", resolution=200.0, save_all=True, append_images=imgs[1:])
    os.replace(tmp, pdf_path)
    print(f"   PDF salvat (PIL): {pdf_path}  ({len(imgs)} pagini)")
    return True


def finalize_pending_pdfs(state):
    """La pornire: pentru orice document cu imaginile complete dar fara PDF, face PDF-ul.
    Nu are nevoie de browser - lucreaza doar de pe disc + total_pages din state.json.
    Asa nu se mai pierde niciun PDF chiar daca oprești des aplicatia."""
    print("Verific PDF-uri restante (imagini complete dar fara PDF)...")
    totals, entries_by_name = {}, {}
    for e in state["downloaded_issues"]:
        nm = e.get("url", "").rstrip("/").split("/")[-1]
        totals[nm] = e.get("total_pages", 0)
        entries_by_name[nm] = e
    if not os.path.isdir(TEMP_ROOT):
        return
    facute = 0
    for cname in sorted(os.listdir(TEMP_ROOT)):
        cdir = os.path.join(TEMP_ROOT, cname)
        if not os.path.isdir(cdir):
            continue
        for name in sorted(os.listdir(cdir)):
            stage = os.path.join(cdir, name)
            if not os.path.isdir(stage):
                continue
            e = entries_by_name.get(name)
            if e is not None and e.get("completed_at"):
                continue  # deja finalizat in json -> nu reface (poate userul a sters PDF-ul intentionat)
            tot = totals.get(name, 0)
            files = collect_page_files(stage)
            if not tot or len(files) < tot:
                continue  # incomplet -> il termina bucla de download
            pdf_path = os.path.join(G_ROOT, cname, name + ".pdf")
            if os.path.exists(pdf_path):
                continue  # deja are PDF
            check_schedule()
            print(f"  [finalize] {cname}/{name}: {len(files)}/{tot} imagini, PDF lipsa -> il fac acum")
            if build_pdf(files, pdf_path, tot):
                e = entries_by_name.get(name)
                if e is not None:
                    e["completed_at"] = datetime.now().isoformat(timespec="seconds")
                    e["pdf"] = pdf_path
                    save_state(state)
                facute += 1
    print(f"Finalize: {facute} PDF-uri restante create." if facute else "Finalize: niciun PDF restant.")


def main():
    state = load_state()

    # 1) intai facem PDF-urile restante (imagini complete dar fara PDF), fara browser
    try:
        finalize_pending_pdfs(state)
    except ScheduledStop:
        print("\n[oprire programata 03:40-04:00] inchid aplicatia.")
        return

    cleanup_stale_automation()   # ca sa nu mai fie nevoie de restart la PC dupa multe rulari

    print("Login: copiez profilul Firefox activ (Firefox-ul tau ramane deschis)...")
    br = Browser()
    try:
        br.start()

        for coll_url in ADDITIONAL_COLLECTIONS:
            check_schedule()
            cname = collection_name(coll_url)
            print(f"\n########## COLECTIE: {cname}  ({coll_url}) ##########")

            def _load_collection():
                drv = br.drv
                drv.get(coll_url)
                WebDriverWait(drv, 40).until(EC.presence_of_element_located((By.TAG_NAME, "body")))
                time.sleep(2)
                return extract_document_urls(drv)

            docs = retry_browser(br, _load_collection, f"enumerare {cname}") or []
            print(f"  documente (sub-colectii) gasite: {len(docs)}")
            for d in docs[:8]:
                print("    -", d)
            if not docs:
                print("  !! niciun document gasit, trec la urmatoarea colectie.")
                continue

            if TEST_MODE:
                docs = docs[:TEST_MAX_DOCS]
                print(f"  [TEST] procesez doar primele {len(docs)} document(e)")

            for view_url in docs:
                check_schedule()
                name = doc_name(view_url)
                pdf_path = os.path.join(G_ROOT, cname, name + ".pdf")
                entry = get_issue(state, view_url)

                # COMPLET = finalizat in json (completed_at e setat DOAR dupa un PDF facut cu succes).
                # Daca e in json ca finalizat -> SKIP, chiar daca userul a sters PDF-ul (l-a urcat pe archive.org).
                if issue_is_complete(entry):
                    print(f"\n=== DOCUMENT {name}: deja finalizat in json, sar ===")
                    continue

                stage_dir = os.path.join(TEMP_ROOT, cname, name)
                info = capture_document(br, view_url, stage_dir, state)
                if info is None:
                    continue

                if info["complete"]:
                    # toate paginile sunt pe disc -> incercam PDF-ul
                    files = collect_page_files(stage_dir)
                    if build_pdf(files, pdf_path, info["total_pages"]):
                        # marcam COMPLET doar daca PDF-ul s-a facut cu succes
                        upsert_issue(state, view_url, info["title"], info["pages_done"],
                                     info["total_pages"], info["total_pages"] - 1,
                                     completed=True, pdf=pdf_path)
                        print(f"   [state] COMPLET {info['pages_done']}/{info['total_pages']}  PDF OK")
                        print(f"   ... pauza {PDF_WAIT}s (PDF) ...")
                        time.sleep(PDF_WAIT)
                    else:
                        print("   !! PDF nereusit acum - documentul ramane neterminat "
                              "(se reia la urmatoarea rulare)")
                else:
                    print(f"   document INCOMPLET ({info['pages_done']}/{info['total_pages']}) "
                          f"- PDF-ul se va face cand documentul e gata")

        print("\nGATA.")
    except ScheduledStop:
        print("\n[oprire programata 03:40-04:00] inchid aplicatia (state.json salvat). "
              "La 04:00 porneste celalalt script.")
    except KeyboardInterrupt:
        print("\n[oprit manual] progresul e salvat in state.json - reia de aici la repornire.")
    except WebDriverException as e:
        print(f"\n[oprit] browserul nu s-a putut recupera: {str(e).splitlines()[0][:120]}")
        print("Progresul e salvat in state.json - reporneste scriptul ca sa reia de aici.")
    finally:
        br.quit()


if __name__ == "__main__":
    main()