# src/app/flashcards_tools.py

import json
import re
from pathlib import Path
from typing import Dict, List, Tuple, Optional

from deep_translator import GoogleTranslator

from .config import get_user_dir


def _get_decks_dir(username: str) -> Path:
    """
    Returns the directory where all of a user's decks are stored.
    """
    user_dir = get_user_dir(username)
    decks_dir = user_dir / "decks"
    decks_dir.mkdir(parents=True, exist_ok=True)
    return decks_dir


def list_user_decks(username: str) -> Dict[str, Path]:
    """
    Returns a mapping of deck name -> deck json path.
    Deck name is taken from the deck's "name" field if present,
    otherwise the filename stem.
    """
    decks_dir = _get_decks_dir(username)
    deck_files = sorted(decks_dir.glob("*.json"))
    decks: Dict[str, Path] = {}

    for path in deck_files:
        try:
            data = json.loads(path.read_text(encoding="utf-8"))
            name = data.get("name") or path.stem
        except Exception:
            name = path.stem

        # ensure uniqueness by appending stem if needed
        if name in decks and decks[name] != path:
            name = f"{name} ({path.stem})"
        decks[name] = path

    return decks


def _ensure_card_stats(card: Dict) -> None:
    """
    Ensure that a card has simple spaced-repetition stats.
    """
    if "score" not in card:  # learning strength
        card["score"] = 0
    if "reviews" not in card:
        card["reviews"] = 0


def load_deck(path: Path) -> Dict:
    """
    Loads a deck from JSON, ensuring 'cards' exists and that
    each card has basic stats for spaced repetition.
    """
    try:
        data = json.loads(path.read_text(encoding="utf-8"))
    except Exception:
        data = {}
    if "cards" not in data or not isinstance(data["cards"], list):
        data["cards"] = []
    if "name" not in data:
        data["name"] = path.stem
    if "tags" not in data or not isinstance(data["tags"], list):
        data["tags"] = []

    for card in data["cards"]:
        _ensure_card_stats(card)

    return data


def save_deck(path: Path, deck: Dict) -> None:
    """
    Saves deck to JSON.
    """
    if "cards" not in deck:
        deck["cards"] = []
    if "name" not in deck:
        deck["name"] = path.stem
    if "tags" not in deck or not isinstance(deck["tags"], list):
        deck["tags"] = []

    # make sure stats are present
    for card in deck["cards"]:
        _ensure_card_stats(card)

    path.write_text(json.dumps(deck, indent=2, ensure_ascii=False), encoding="utf-8")


# ------------------------------------------------------------
# Shared tokenization
# ------------------------------------------------------------

def _extract_candidate_words(text: str) -> List[str]:
    """
    Simple tokenizer & filter for candidate vocab words.
    """
    tokens = re.findall(r"\b\w+\b", text, flags=re.UNICODE)
    out = []
    seen = set()
    for t in tokens:
        t_norm = t.strip()
        if len(t_norm) < 2:
            continue
        if any(ch.isdigit() for ch in t_norm):
            continue
        lower = t_norm.lower()
        if lower in seen:
            continue
        seen.add(lower)
        out.append(t_norm)
    return out


# ------------------------------------------------------------
# OCR → Flashcards
# ------------------------------------------------------------

def generate_flashcards_from_ocr_results(
    username: str,
    ocr_results: List[Dict],
    deck_name: str = "ocr",
    target_lang: str = "en",
    tags: Optional[List[str]] = None,
) -> Path:
    """
    Takes OCR results (as produced by ocr_tools.ocr_and_translate_batch)
    and constructs a simple vocab deck.

    ocr_results: list of dict with keys:
        - "text": original text
        - optionally other fields (ignored)
    """
    all_text = []
    for res in ocr_results:
        t = res.get("text") or res.get("raw_text") or ""
        if t:
            all_text.append(t)
    joined = "\n".join(all_text)

    words = _extract_candidate_words(joined)
    if not words:
        raise ValueError("No candidate words found in OCR results.")

    translator = GoogleTranslator(source="auto", target=target_lang)
    cards = []
    for w in words:
        try:
            trans = translator.translate(w)
        except Exception:
            continue
        if not trans:
            continue
        if trans.strip().lower() == w.strip().lower():
            continue
        card = {
            "front": w,
            "back": trans,
            "content_type": "ocr_vocab",
            "language": target_lang,
        }
        _ensure_card_stats(card)
        cards.append(card)

    if not cards:
        raise ValueError("No translatable words found to build cards.")

    decks_dir = _get_decks_dir(username)
    deck_path = decks_dir / f"{deck_name}.json"

    deck = {
        "name": deck_name,
        "cards": cards,
        "tags": tags or [],
    }
    save_deck(deck_path, deck)
    return deck_path


# ------------------------------------------------------------
# Conversation/Text → Flashcards
# ------------------------------------------------------------

def generate_flashcards_from_text(
    username: str,
    text: str,
    deck_name: str = "conversation",
    target_lang: str = "en",
    tags: Optional[List[str]] = None,
) -> Path:
    """
    Build a vocab deck from raw conversation text.
    """
    words = _extract_candidate_words(text)
    if not words:
        raise ValueError("No candidate words found in text.")

    translator = GoogleTranslator(source="auto", target=target_lang)
    cards = []
    for w in words:
        try:
            trans = translator.translate(w)
        except Exception:
            continue
        if not trans:
            continue
        if trans.strip().lower() == w.strip().lower():
            continue
        card = {
            "front": w,
            "back": trans,
            "content_type": "conversation_vocab",
            "language": target_lang,
        }
        _ensure_card_stats(card)
        cards.append(card)

    if not cards:
        raise ValueError("No translatable words found to build cards.")

    decks_dir = _get_decks_dir(username)
    deck_path = decks_dir / f"{deck_name}.json"

    deck = {
        "name": deck_name,
        "cards": cards,
        "tags": tags or ["conversation"],
    }
    save_deck(deck_path, deck)
    return deck_path