#!/usr/bin/env python3
"""
KOCR v4 — 金蝶 K3 凭证 OCR 填表工具
小唯 A06 出品 | 自动调参版

用法:
  python3 kocr.py                    # 批量处理所有新照片
  python3 kocr.py --new              # 只处理未识别过的照片
  python3 kocr.py --status           # 查看处理进度
  python3 kocr.py IMG_001.jpg        # 单张处理
  python3 kocr.py --merge            # 合并输出排序好的汇总 Excel
  python3 kocr.py --config my.yaml   # 指定配置文件

特性:
  - 自动图像分析 + 自动调参（按照片亮度/阴影/模糊度选参数）
  - 大小写金额交叉验证（大写转数字比对）
  - PPStructure TableMaster 表格识别（失败回退 DB+SVTR）
  - 交错匹配（dr1, cr1, dr2, cr2...）
  - 多页凭证自动合并
  - 外置配置文件（部署时只需调整 config.yaml）
"""

import os
import sys
import argparse
import yaml
import cv2

# 项目根目录
PROJECT_DIR = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, PROJECT_DIR)

from core.recognizer import run_ocr
from core.parser import K3VoucherParser
from core.postprocessor import validate_voucher, validate_merged_vouchers, generate_review_file
from core.excel_writer import fill_template, merge_output, merge_voucher_data


# ======== 路径常量 ========
DEFAULT_CONFIG = os.path.join(PROJECT_DIR, 'config.yaml')
BATCH_INPUT_DIR = os.path.join(PROJECT_DIR, 'input')
BATCH_INPUT_DONE = os.path.join(PROJECT_DIR, 'input/done')
BATCH_OUTPUT_DIR = os.path.join(PROJECT_DIR, 'output')
INDEX_FILE = os.path.join(PROJECT_DIR, '.kocr_index.json')


# ======== 配置加载 ========
def load_config(config_path=None):
    """加载 YAML 配置，支持命令行覆盖"""
    path = config_path or DEFAULT_CONFIG
    if not os.path.exists(path):
        print(f"⚠️  配置文件不存在: {path}，使用默认配置")
        return _default_config()
    
    with open(path, 'r', encoding='utf-8') as f:
        config = yaml.safe_load(f)
    return config


def _default_config():
    """默认配置（当 config.yaml 不存在时使用）"""
    return {
        'preprocessing': {
            'max_side': 1600,
            'denoise': True,
            'deskew': True,
            'remove_shadow': True,
            'sharpen': True,
        },
        'ocr': {
            'use_gpu': False,
            'lang': 'ch',
            'show_log': False,
            'use_angle_cls': True,
            'det_db_thresh': 0.3,
            'det_db_box_thresh': 0.5,
            'det_db_unclip_ratio': 2.2,
            'max_side_len': 1600,
            'det_db_score_mode': 'slow',
            'min_size': 8,
            'rec_drop_score': 0.5,
        },
        'postprocessing': {
            'amount_balance_tolerance': 0.01,
        },
        'paths': {
            'input_dir': 'input',
            'output_dir': 'output',
            'done_dir': 'input/done',
            'template': '~/.hermes/cache/documents/doc_704eac18442d_凭证模板.xls',
        },
        'photo_analysis': {
            'brightness_thresholds': {'flash': 200, 'dark': 60},
            'blur_threshold': 100,
            'shadow_threshold': 0.3,
        },
    }


# ======== 索引管理 ========
def load_index():
    if os.path.exists(INDEX_FILE):
        try:
            with open(INDEX_FILE, 'r', encoding='utf-8') as f:
                return yaml.safe_load(f) or {}
        except:
            pass
    return {}


def save_index(idx):
    os.makedirs(os.path.dirname(INDEX_FILE), exist_ok=True)
    with open(INDEX_FILE, 'w', encoding='utf-8') as f:
        yaml.dump(idx, f, allow_unicode=True, default_flow_style=False)


def mark_processed(filename, voucher_info, output_path):
    idx = load_index()
    idx[filename] = {
        'voucher_word': voucher_info.get('voucher_word', '记'),
        'voucher_no': voucher_info.get('voucher_no'),
        'date': voucher_info.get('date').strftime('%Y-%m-%d') if voucher_info.get('date') else '',
        'entries': len(voucher_info.get('entries', [])),
        'total_debit': round(sum(e.get('debit', 0) or 0 for e in voucher_info.get('entries', [])), 2),
        'total_credit': round(sum(e.get('credit', 0) or 0 for e in voucher_info.get('entries', [])), 2),
        'output': os.path.basename(output_path) if output_path else '',
        'processed_at': __import__('datetime').datetime.now().strftime('%Y-%m-%d %H:%M:%S'),
    }
    save_index(idx)


def is_processed(filename):
    idx = load_index()
    return filename in idx


# ======== 单张凭证处理 ========
def process_single(image_path, config, output_path=None, verbose=False, mark_done=True):
    """处理单张凭证照片
    
    流程: 原图 → PaddleOCR → 解析 → 填Excel → 验证
    
    Returns:
        (voucher_data, quality)
    """
    filename = os.path.basename(image_path)
    
    if verbose:
        print(f"\n📷 {filename}")
    
    # 1. 读取图像
    img = cv2.imread(image_path)
    if img is None:
        raise ValueError(f"无法读取图片: {image_path}")
    
    # 2. 原图直出 OCR（无预处理）
    ocr_config = config.get('ocr', {})
    
    if verbose:
        print(f"   📋 OCR 识别中...")
    lines = run_ocr(img, ocr_config)
    
    if verbose:
        print(f"   OCR: {len(lines)} 行文字")
    
    # 3. 解析
    parser = K3VoucherParser(lines)
    voucher = parser.parse()
    
    if verbose:
        print(f"   公司: {voucher['company']}")
        print(f"   日期: {voucher['date'].strftime('%Y-%m-%d') if voucher['date'] else '未识别'}")
        print(f"   凭证: {voucher['voucher_word']}-{voucher['voucher_no']}")
        print(f"   分录: {len(voucher['entries'])} 行")
        for e in voucher['entries']:
            dr = f"借 {e['debit']:.2f}" if e['debit'] else ""
            cr = f"贷 {e['credit']:.2f}" if e['credit'] else ""
            print(f"     {e['account_code']} {e['account_name']:20s} {dr:>12s} {cr:>12s}")
    
    # 4. 大小写金额交叉验证
    if voucher['total_str'] and voucher['total_amount'] > 0:
        from core.postprocessor import amount_check
        total_dr = sum(e.get('debit', 0) or 0 for e in voucher['entries'])
        total_cr = sum(e.get('credit', 0) or 0 for e in voucher['entries'])
        total_max = max(total_dr, total_cr)
        if total_max > 0:
            cn_part = __import__('re').sub(r'[\\d,.\\s￥¥]', '', voucher['total_str'])
            if cn_part:
                ok = amount_check(total_max, cn_part)
                if verbose:
                    status = '✅' if ok else '❌'
                    print(f"   大小写校验: {status} 小写={total_max:.2f} 大写={cn_part}")
    
    # 5. 填 Excel
    if output_path is None:
        basename = os.path.splitext(filename)[0]
        output_path = os.path.join(BATCH_OUTPUT_DIR, f'{basename}_K3.xls')
    
    paths_config = config.get('paths', {})
    template = os.path.expanduser(paths_config.get('template', '~/.hermes/cache/documents/doc_704eac18442d_凭证模板.xls'))
    
    fill_template(voucher, template, output_path, verbose=verbose)
    
    # 6. 标记已处理
    if mark_done:
        mark_processed(filename, voucher, output_path)
    
    return voucher, None


# ======== 状态报告 ========
def show_status(config):
    paths_config = config.get('paths', {})
    input_dir = os.path.expanduser(paths_config.get('input_dir', 'input'))
    done_dir = os.path.expanduser(paths_config.get('done_dir', 'input/done'))
    
    idx = load_index()
    
    exts = ('.jpg', '.jpeg', '.png', '.bmp', '.tiff', '.tif', '.webp')
    pending_files = sorted([f for f in os.listdir(input_dir) if f.lower().endswith(exts)]) if os.path.exists(input_dir) else []
    done_files = sorted([f for f in os.listdir(done_dir) if f.lower().endswith(exts)]) if os.path.exists(done_dir) else []
    
    processed = set(idx.keys())
    
    print("📊 KOCR v4 处理状态")
    print(f"   {'='*40}")
    print(f"   索引记录: {len(processed)} 张已处理")
    print(f"   input/ 待处理: {len(pending_files)} 张")
    print(f"   done/ 已归档: {len(done_files)} 张")
    print()
    
    if idx:
        print("   最近处理的:")
        sorted_items = sorted(
            idx.items(),
            key=lambda x: (x[1].get('date', ''), x[1].get('voucher_no', 0)),
            reverse=True
        )[:10]
        for fname, info in sorted_items:
            v = f"{info.get('voucher_word', '记')}-{info.get('voucher_no', '?')}"
            d = info.get('date', '')
            e = info.get('entries', 0)
            t = info.get('processed_at', '')
            print(f"     {fname:30s} → {v:>8s}  {d}  {e}笔分录  {t}")
    
    if pending_files:
        new_pending = [f for f in pending_files if f not in processed]
        if new_pending:
            print(f"\n   📋 未处理的新照片 ({len(new_pending)} 张):")
            for f in new_pending[:10]:
                print(f"     {f}")
            if len(new_pending) > 10:
                print(f"     ... 还有 {len(new_pending)-10} 张")
    
    print()


# ======== 主入口 ========
def main():
    parser = argparse.ArgumentParser(description='KOCR v4 — 金蝶 K3 凭证 OCR 填表工具')
    parser.add_argument('input', nargs='?', default=None,
                        help='凭证照片路径（留空则批量处理 input 目录）')
    parser.add_argument('--output', '-o', default=None, help='输出 Excel 路径')
    parser.add_argument('--config', '-c', default=DEFAULT_CONFIG, help='配置文件路径')
    parser.add_argument('--verbose', '-v', action='store_true', help='显示详细结果')
    parser.add_argument('--status', '-s', action='store_true', help='显示处理进度')
    parser.add_argument('--new', '-n', action='store_true', help='只处理未识别过的新照片')
    parser.add_argument('--archive', '-a', action='store_true', help='归档已处理的照片到 done/')
    parser.add_argument('--review', '-r', action='store_true', help='生成数据校验报告')
    parser.add_argument('--merge', '-m', action='store_true', help='合并输出汇总 Excel')
    parser.add_argument('--force', '-f', action='store_true', help='强制重新处理所有图片')
    args = parser.parse_args()
    
    # 加载配置
    config = load_config(args.config)
    
    # 路径
    paths_config = config.get('paths', {})
    input_dir = os.path.expanduser(paths_config.get('input_dir', BATCH_INPUT_DIR))
    output_dir = os.path.expanduser(paths_config.get('output_dir', BATCH_OUTPUT_DIR))
    done_dir = os.path.expanduser(paths_config.get('done_dir', BATCH_INPUT_DONE))
    template_path = os.path.expanduser(paths_config.get('template', '~/.hermes/cache/documents/doc_704eac18442d_凭证模板.xls'))
    
    # 确保目录存在
    for d in [input_dir, done_dir, output_dir]:
        os.makedirs(d, exist_ok=True)
    
    if not os.path.exists(template_path):
        print(f"❌ 找不到模板: {template_path}")
        sys.exit(1)
    
    # ---- 状态查看 ----
    if args.status:
        show_status(config)
        return
    
    # ---- 归档 ----
    if args.archive:
        idx = load_index()
        if not os.path.exists(input_dir):
            print("📭 input 目录不存在")
            return
        moved = 0
        for fname in os.listdir(input_dir):
            if fname in idx:
                src = os.path.join(input_dir, fname)
                dst = os.path.join(done_dir, fname)
                os.rename(src, dst)
                moved += 1
        print(f"📦 已归档 {moved} 张照片到 input/done/")
        return
    
    # ---- 确定文件列表 ----
    input_path = args.input or input_dir
    
    if os.path.isdir(input_path):
        # 批量模式
        exts = ('.jpg', '.jpeg', '.png', '.bmp', '.tiff', '.tif', '.webp')
        files = sorted([f for f in os.listdir(input_path) if f.lower().endswith(exts)])
        
        if not files:
            print(f"📭 {input_path} 中没有图片文件")
            sys.exit(0)
        
        # --new 模式过滤
        idx = load_index()
        if args.new or not args.force:
            new_files = [f for f in files if f not in idx]
            skipped = len(files) - len(new_files)
            files = new_files
            if skipped > 0:
                print(f"⏭️  跳过 {skipped} 张已处理过的照片")
            if not files:
                print("✅ 全部照片都已处理过了，无需重复识别")
                return
        
        print(f"📂 KOCR v4 — 批量处理: {len(files)} 张凭证")
        print(f"   输入: {input_path}")
        print(f"   输出: {output_dir}")
        print(f"   {'='*40}")
        
        results = []
        for i, fname in enumerate(files):
            fpath = os.path.join(input_path, fname)
            basename = os.path.splitext(fname)[0]
            out_path = os.path.join(output_dir, f'{basename}_K3.xls')
            
            print(f"  [{i+1}/{len(files)}] {fname}")
            try:
                voucher, quality = process_single(fpath, config, out_path, verbose=args.verbose)
                vno = voucher.get('voucher_no', '?')
                vdate = voucher['date'].strftime('%Y-%m-%d') if voucher.get('date') else '?'
                results.append((fname, voucher, out_path, quality))
                print(f"  ✅ ['raw'] {vdate} 记-{vno} → {os.path.basename(out_path)}")
            except Exception as e:
                print(f"  ❌ 失败: {e}")
            print()
        
        # ---- 合并凭证数据（多页合并）+ 统一校验 ----
        merged_vouchers = merge_voucher_data(results)
        
        # ---- 科目名称归一化：同一代码取最高频名称 ---- #
        from collections import Counter
        code_name_counter = Counter()
        code_name_votes = {}
        for v in merged_vouchers:
            for e in v['entries']:
                c = e['account_code']
                n = e['account_name']
                if n and len(n) >= 2:
                    code_name_counter[(c, n)] += 1
        # 每个代码选最高频名称
        code_best_name = {}
        for (c, n), cnt in code_name_counter.items():
            if c not in code_best_name or cnt > code_best_name[c][1]:
                code_best_name[c] = (n, cnt)
        # 输出差异
        for v in merged_vouchers:
            for e in v['entries']:
                c = e['account_code']
                if c in code_best_name:
                    best_n, _ = code_best_name[c]
                    if e['account_name'] != best_n:
                        e['account_name'] = best_n
        
        # 写回流到 results（merge_output 用 results 写文件）
        for fname, voucher, out_path, quality in results:
            if voucher.get('entries'):
                for e in voucher['entries']:
                    c = e['account_code']
                    if c in code_best_name:
                        e['account_name'] = code_best_name[c][0]
        
        validation_results = validate_merged_vouchers(
            merged_vouchers,
            config.get('postprocessing', {}).get('amount_balance_tolerance', 0.01)
        )

        clean_count = sum(1 for _, is_clean, _, _ in validation_results if is_clean)
        multi_count = sum(1 for v in merged_vouchers if v.get('is_multi_page'))

        print(f"   {'='*40}")
        print(f"   📋 合并后校验结果（{len(merged_vouchers)} 笔凭证，含 {multi_count} 笔多页）:")
        for voucher, is_clean, issues, summary in validation_results:
            date_str = voucher['date'].strftime('%m-%d') if voucher.get('date') else '??'
            vno = voucher.get('voucher_no', '?')
            multi = '📑' if voucher.get('is_multi_page') else '  '
            entries = len(voucher['entries'])
            status = '✅' if is_clean else '⚠️'
            print(f"   {status} {multi} {date_str} 记-{vno} ({entries}条分录)")
            if not is_clean:
                for iss in issues[:3]:
                    print(f"      {iss}")
        print(f"   {'='*40}")
        print(f"   ✅ 校验通过: {clean_count} 笔")
        print(f"   ⚠️  需人工复核: {len(validation_results) - clean_count} 笔")

        # ---- 合并输出 ----
        if args.merge and len(results) > 1:
            merge_path = merge_output(results, template_path, output_dir)
            print(f"\n   📊 已生成: {merge_path}")

        # ---- 校验报告（基于合并后的数据） ----
        if args.review and results:
            review_results = []
            for voucher, is_clean, issues, summary in validation_results:
                fname = ' + '.join(voucher.get('pages', ['unknown']))
                review_results.append((fname, voucher, '', None))
            review_path = generate_review_file(review_results, output_dir)
            print(f"   📋 校验报告: {os.path.basename(review_path)}")
        
        # ---- 自动归档 ----
        archived = 0
        for fname, voucher, out_path, quality in results:
            src = os.path.join(input_path, fname)
            dst = os.path.join(done_dir, fname)
            if os.path.exists(src):
                os.rename(src, dst)
                archived += 1
        if archived > 0:
            print(f"   📦 已归档 {archived} 张照片到 done/")
        
        print(f"🎉 完成! {len(results)}/{len(files)} 成功")
    
    else:
        # 单文件模式
        if not os.path.exists(input_path):
            print(f"❌ 找不到图片: {input_path}")
            sys.exit(1)
        
        if args.output:
            out_path = args.output
        else:
            basename = os.path.splitext(os.path.basename(input_path))[0]
            out_path = os.path.join(output_dir, f'{basename}_K3.xls')
        
        voucher, quality = process_single(input_path, config, out_path, verbose=args.verbose)
        results = [(os.path.basename(input_path), voucher, out_path, quality)]
        
        if args.review:
            review_path = generate_review_file(results, output_dir)
            print(f"   📋 校验报告: {os.path.basename(review_path)}")


if __name__ == '__main__':
    main()