import re
from pathlib import Path
from collections import defaultdict

# ================= 配置区 =================
# 请将下方路径替换为你自己电脑上的实际路径
# 注意：Windows 系统路径前面的 r 不要删掉，它用来防止路径中的反斜杠被转义

# 1. 作品元数据文件路径 (即包含作品信息的 catalog.ndjson)
CATALOG_FILE = r"C:\Your\Path\Here\catalog.ndjson"

# 2. 用户行为数据文件路径 (即包含你的打分、短评、状态的 journal.ndjson)
JOURNAL_FILE = r"C:\Your\Path\Here\journal.ndjson"

# 3. 输出大文件夹路径 (脚本运行后，会在这个路径下自动生成 Book/Movie/Show 三个文件夹)
OUTPUT_DIR = r"D:\Your\Obsidian\Vault\Media"
# ==========================================

# 通用流派（标签）翻译字典
# NeoDB 导出的流派通常是英文，这里做一个简单的映射，方便在 Obsidian 里用中文检索
# 如果导出的标签不在这个字典里，脚本会自动保留原有的英文
GENRE_MAP = {
    "action": "动作", "adventure": "冒险", "animation": "动画", "anime": "动画",
    "biography": "传记", "comedy": "喜剧", "crime": "犯罪", "documentary": "纪录片",
    "drama": "剧情", "family": "家庭", "fantasy": "奇幻", "history": "历史",
    "horror": "恐怖", "music": "音乐", "musical": "歌舞", "mystery": "悬疑",
    "romance": "爱情", "sci-fi": "科幻", "science fiction": "科幻", "short": "短片",
    "sport": "体育", "thriller": "惊悚", "war": "战争", "western": "西部",
    "donghua": "国创"
}

# ================= 辅助函数区 =================

def sanitize_filename(name):
    """
    清洗文件名：剔除 Windows/Mac 系统不允许出现在文件名中的特殊字符（\ / * ? : " < > |）
    防止因为作品名字里带有冒号等符号导致文件创建失败。
    """
    name = str(name)
    return re.sub(r'[\\/*?:"<>|]', "", name).strip()

def escape_yaml(text):
    """
    YAML 防报错处理：如果文本中包含双引号或换行符，会破坏 Obsidian 的 Frontmatter 结构。
    这里将双引号转义，并把回车换行替换为空格。
    """
    if not text: return ""
    return str(text).replace('"', '\\"').replace('\n', ' ').replace('\r', '')

def translate_genres(genre_list):
    """
    翻译流派：遍历作品的标签列表，去上面的 GENRE_MAP 字典里找对应的中文。
    """
    if not isinstance(genre_list, list): return genre_list
    return [GENRE_MAP.get(str(g).lower(), g) for g in genre_list]

def format_list(item_list):
    """
    列表格式化：将 Python 的列表安全地转换为 YAML 识别的数组格式，例如 ["标签1", "标签2"]。
    """
    if not item_list: return "[]"
    if isinstance(item_list, list):
        items = [f'"{(escape_yaml(i))}"' for i in item_list if i]
        return "[" + ", ".join(items) + "]"
    return f'["{escape_yaml(item_list)}"]'

# ================= 核心执行逻辑 =================

def process_neodb():
    cat_path = Path(CATALOG_FILE)
    jou_path = Path(JOURNAL_FILE)
    out_path = Path(OUTPUT_DIR)

    # 1. 前置检查：确保配置文件路径正确
    if not cat_path.exists() or not jou_path.exists():
        print("❌ 找不到 JSON 文件，请检查路径。")
        return

    # 2. 初始化分类文件夹：自动在目标路径下生成 Book, Movie, Show
    # 注：因本人不在 NeoDB 记录游戏，故此处不包含 Game 分类
    dirs = {'book': out_path / "Book", 'movie': out_path / "Movie", 'tv': out_path / "Show"}
    for d in dirs.values():
        d.mkdir(parents=True, exist_ok=True)

    # 3. 收集用户行为数据 (Journal)
    # 使用 defaultdict，如果遇到还没记录过的作品，自动初始化一组空数据
    user_records = defaultdict(lambda: {'status': '', 'rating': '', 'comment': '', 'date_planned': '9999-12-31', 'date_completed': ''})

    # 读取 journal.ndjson，逐行解析
    with open(jou_path, 'r', encoding='utf-8') as f:
        for line in f:
            if not line.strip(): continue
            row = json.loads(line)
            t = row.get('type')
            
            # 解析：标记状态 (看过/想看等)
            if t == 'ShelfMember':
                content = row.get('content', {})
                uid = content.get('withRegardTo') or row.get('withRegardTo')
                status = content.get('status') or row.get('status')
                if uid: user_records[uid]['status'] = status
            
            # 解析：用户打分
            elif t == 'Rating':
                content = row.get('content', {})
                uid = content.get('withRegardTo') or row.get('withRegardTo')
                if uid: user_records[uid]['rating'] = content.get('value', '')
            
            # 解析：用户短评
            elif t == 'Comment':
                content = row.get('content', {})
                uid = content.get('withRegardTo') or row.get('withRegardTo')
                if uid: user_records[uid]['comment'] = content.get('content', '')
            
            # 解析：时间线 (想看日期、看完日期)
            elif t == 'ShelfLog' or ('item' in row and 'timestamp' in row):
                content = row.get('content', {})
                uid = row.get('item') or row.get('withRegardTo') or content.get('withRegardTo')
                log_status = row.get('status') or content.get('status')
                ts = row.get('timestamp', '')[:10] if row.get('timestamp') else '' # 截取 YYYY-MM-DD
                
                if not uid or not ts: continue
                # 如果是“想看”，记录最早的时间
                if log_status == 'wishlist' and ts < user_records[uid]['date_planned']:
                    user_records[uid]['date_planned'] = ts
                # 如果是“看过/读过”，记录最晚的时间
                elif log_status == 'complete' and ts > user_records[uid]['date_completed']:
                    user_records[uid]['date_completed'] = ts

    # 4. 过滤规则：只提取标记为“看过/完成 (complete)”的作品 ID
    completed_ids = {uid for uid, data in user_records.items() if data['status'] == 'complete'}

    # 5. 组合数据并生成 Markdown 文件
    count = 0
    with open(cat_path, 'r', encoding='utf-8') as f:
        for line in f:
            if not line.strip(): continue
            item = json.loads(line)
            uid = item.get('id')
            
            # 如果这件作品不在“已完成”列表里，直接跳过
            if uid not in completed_ids: continue
            
            category = item.get('category')
            # 如果属于不处理的分类（如 podcast, game 等），直接跳过
            if category not in dirs: continue
            
            title = item.get('display_title') or item.get('title')
            if not title: continue

            # 取出之前解析好的个人打分、评价、日期等数据
            my_data = user_records[uid]
            planned = "" if my_data['date_planned'] == '9999-12-31' else my_data['date_planned']
            completed = my_data['date_completed']
            comment = my_data['comment']
            rating = my_data['rating']
            
            # 处理文件名和路径
            safe_title = sanitize_filename(title)
            file_path = dirs[category] / f"{safe_title}.md"
            
            # 处理简介格式
            brief = item.get('description') or item.get('brief') or ""
            brief = brief.replace('\r\n', '\n')
            
            neodb_url = uid
            aliases = format_list(item.get('orig_title', []))
            
            raw_genres = item.get('genre') or item.get('tags') or []
            translated_genres = translate_genres(raw_genres)
            genres = format_list(translated_genres)

            # --- 开始拼接 Markdown 文件内容 ---
            md = "---\n"
            md += f'title: "{escape_yaml(title)}"\n'
            md += f'aliases: {aliases}\n'
            md += f'category: "{category}"\n'
            md += f'genres: {genres}\n'
            md += f'neodb_url: "{neodb_url}"\n'
            md += f'cover: "{item.get("cover_image_url", "")}"\n'
            
            # 针对书籍的专属属性
            if category == 'book':
                md += f'type: "book"\n' # 刻意避让 Hugo 的 type 字段，使用了独立分类属性
                md += f'author: {format_list(item.get("author"))}\n'
                md += f'translator: {format_list(item.get("translator"))}\n'
                
                # 尝试通过正则从作者字段提取国籍（如 [日]太宰治 提取出“日”）
                authors = item.get('author', [])
                country = ""
                if authors:
                    match = re.search(r'\[(.*?)\]', str(authors[0]))
                    if match: country = match.group(1)
                md += f'country: "{country}"\n'
                
                publishers = item.get('publisher', [])
                pub = publishers[0] if isinstance(publishers, list) and publishers else publishers
                md += f'publisher: "{escape_yaml(pub)}"\n'
                md += f'publish_year: {item.get("pub_year", "")}\n'
                md += f'pages: {item.get("pages", "")}\n'
                md += f'isbn: "{item.get("isbn", "")}"\n'
                md += f'format: []\n'
                md += f'status: "读过"\n'
                md += f'date_planned: {planned}\n'
                md += f'date_read: {completed}\n'
                
            # 针对影视的专属属性
            elif category in ['movie', 'tv']:
                md += f'type: "{category}"\n'
                md += f'director: {format_list(item.get("director"))}\n'
                md += f'year: {item.get("year", "")}\n'
                
                areas = item.get('area', [])
                area = areas[0] if isinstance(areas, list) and areas else areas
                md += f'country: "{escape_yaml(area)}"\n'
                
                if category == 'movie':
                    md += f'duration: {item.get("duration", "")}\n'
                else:
                    md += f'show_type: "{item.get("type", "")}"\n'
                    # 处理电视剧的季数
                    if item.get("type") == "TVSeason":
                        md += f'season: {item.get("season_number", "")}\n'
                    else:
                        md += f'season: ""\n'
                    md += f'episodes: {item.get("episode_count", "")}\n'
                    
                md += f'status: "看过"\n'
                md += f'date_planned: {planned}\n'
                md += f'date_watched: {completed}\n'

            # 通用收尾属性
            md += f'awards: []\n'
            md += f'rating: {rating}\n'
            md += "---\n\n"
            
            # --- 正文部分 ---
            # 使用 Obsidian 的 Callout 语法将长简介默认折叠，保持页面清爽
            md += "> [!quote]- 简介\n"
            for p in brief.split('\n'):
                if p.strip(): md += f"> {p}\n"
                
            md += "\n## 短评\n"
            md += f"{comment}\n"
            
            # 将拼接好的内容写入物理文件
            with open(file_path, 'w', encoding='utf-8') as out_f:
                out_f.write(md)
            count += 1

    print(f"🎉 导出完成！共成功处理并生成 {count} 个文件。")

if __name__ == "__main__":
    process_neodb()