#!/usr/bin/env python3
"""
第三轮修复：内料继承主产品供应商 + CSV中有但供应商为空的补充
规则：
  1. 内料(-NL/【内料】): 去掉 -NL 后缀，找到主产品，继承主产品的供应商
  2. CSV中有记录但供应商为空的包材: 去掉后缀找主产品继承
  3. CSV中有记录且有供应商的：不再处理（v2已完成）
"""
import csv
import pymysql
from config import DB_CONFIG

CSV_FILE = "/home/gxd/.openclaw/workspace-dingtalk-assistant/media/inbound/采购订单-20260531-1780213945661.csv"

def strip_suffix(code):
    """去掉已知后缀返回base_code"""
    SUFFIXES = [
        '-FM/GZ', '-TMGJ-NL', '-ALJK-BT', '-ALJK-FM/GZ',
        '-TM-BT', '-TM-PZ', '-TM-CH', '-TM-BQ', '-TM-SSM',
        '-TM-HMT', '-TM-RG', '-TM-GZ', '-TM-FM/GZ',
        '-PDD-BT', '-PDD-PZ', '-PDD-CH', '-PDD-HMT',
        '-NL', '-BT', '-PZ', '-CH', '-BQ', '-SSM', '-HMT',
        '-RG', '-GZ', '-JD', '-PDD', '-ALJK', '-TM',
        '-WPH', '-DW',
    ]
    for s in SUFFIXES:
        if code.endswith(s):
            base = code[:-len(s)]
            if base:
                return base
    return code

def main():
    conn = pymysql.connect(**DB_CONFIG)
    cur = conn.cursor()

    # 1. 读取CSV——构建精确匹配索引
    csv_best = {}
    with open(CSV_FILE, encoding="utf-8-sig") as f:
        reader = csv.DictReader(f)
        for row in reader:
            code = row["物料编码"].strip()
            supplier = row["供应商"].strip()
            price_str = row["含税单价"].strip()
            try:
                price = float(price_str) if price_str else None
            except ValueError:
                price = None
            if code not in csv_best:
                csv_best[code] = (supplier, price)
            else:
                old_sup, old_price = csv_best[code]
                if supplier and not old_sup:
                    csv_best[code] = (supplier, old_price if old_price is not None else price)
                if old_price is None and price is not None:
                    csv_best[code] = (old_sup if old_sup else supplier, price)

    print(f"CSV唯一物料编码: {len(csv_best)}")

    # 2. 找出所有包材类且supplier_name_match为空的产品
    keywords = ['瓶子','泵头','彩盒','海绵头','瓶贴','标签','收缩膜','瓶体','瓶标','软管','盖子','膜袋','标贴','唇膏管','内料','海绵','瓶子+内塞','泵头','瓶标','海绵头']
    clauses = " OR ".join(["product_name LIKE '%%{}%%'".format(k) for k in keywords])

    cur.execute("""
        SELECT id, product_code, product_name, cost_price
        FROM products
        WHERE ({}) AND (supplier_name_match IS NULL OR supplier_name_match = '')
    """.format(clauses))
    need_fix = cur.fetchall()
    print(f"需要修复的包材(当前supplier为空): {len(need_fix)}")

    # 3. 加载所有产品的supplier_name_match（包括已有值的），用于继承
    cur.execute("SELECT product_code, supplier_name_match, cost_price FROM products")
    all_prod = {}
    for code, sup, price in cur.fetchall():
        all_prod[code.strip()] = {
            'supplier': sup or '',
            'cost_price': float(price) if price else 0
        }

    # 4. 逐个处理
    update_1 = []  # 内料继承主产品供应商
    update_2 = []  # CSV中供应商为空但可继承主产品的
    update_3 = []  # 通过找主产品获得供应商（其他后缀）

    for pid, product_code, product_name, old_price in need_fix:
        code = product_code.strip()
        is_nl = code.endswith('-NL') or '【内料】' in product_name
        new_supplier = None

        # 4a. 检查CSV是否有该编码且供应商不为空
        if code in csv_best:
            csv_sup, csv_price = csv_best[code]
            if csv_sup:
                # CSV有供应商——v2应该已经覆盖了，但如果还没覆盖到
                new_supplier = csv_sup
                new_price = csv_price if csv_price is not None else (float(old_price) if old_price else None)
                update_2.append((new_supplier, new_price, pid))
                continue

        # 4b. 去掉后缀找主产品
        base_code = strip_suffix(code)
        if base_code and base_code != code and base_code in all_prod:
            base = all_prod[base_code]
            if base['supplier']:
                new_supplier = base['supplier']

        # 4c. 对于内料，特别处理：多剥一层后缀
        if not new_supplier and is_nl:
            # 去-NL后缀后再去一次
            stripped = code
            if stripped.endswith('-NL'):
                stripped = stripped[:-3]
            # 再去其他后缀
            base_code2 = strip_suffix(stripped)
            if base_code2 and base_code2 in all_prod:
                base = all_prod[base_code2]
                if base['supplier']:
                    new_supplier = base['supplier']

        if new_supplier:
            new_price = float(old_price) if old_price else None
            if is_nl:
                update_1.append((new_supplier, new_price, pid))
            else:
                update_2.append((new_supplier, new_price, pid))

    print(f"\n内料继承主产品供应商: {len(update_1)}")
    print(f"其他后缀继承主产品供应商: {len(update_2)}")
    print(f"仍然无法匹配: {len(need_fix) - len(update_1) - len(update_2)}")

    # 5. 批量更新
    def run_updates(label, updates_data):
        if not updates_data:
            return
        sql = "UPDATE products SET supplier_name_match=%s, cost_price=%s WHERE id=%s"
        n = 0
        for i in range(0, len(updates_data), 500):
            batch = updates_data[i:i+500]
            cur.executemany(sql, batch)
            conn.commit()
            n += len(batch)
        print(f"  {label}: {n} 条已更新")

    run_updates("内料继承", update_1)
    run_updates("其他继承", update_2)

    # 6. 最终统计
    cur.execute("SELECT COUNT(*) FROM products WHERE ({}) AND (supplier_name_match IS NULL OR supplier_name_match = '')".format(clauses))
    still_empty = cur.fetchone()[0]
    print(f"\n修复后，包材仍然为空: {still_empty}")

    # 按后缀分类查看剩余为空的情况
    if still_empty > 0:
        cur.execute("""
            SELECT product_code, product_name FROM products
            WHERE ({}) AND (supplier_name_match IS NULL OR supplier_name_match = '')
        """.format(clauses))
        remain = cur.fetchall()
        nl_left = sum(1 for c, _ in remain if c.endswith('-NL'))
        print(f"  其中-NL内料: {nl_left}")
        print(f"  其他: {still_empty - nl_left}")

    cur.execute("SELECT COUNT(*) FROM products WHERE supplier_name_match != ''")
    total_sup = cur.fetchone()[0]
    print(f"\n全部产品有供应商: {total_sup}")

    conn.close()
    print("\n✅ 完成！")


if __name__ == "__main__":
    main()