# -*- coding: utf-8 -*-
"""
从 inkreach PostgreSQL 同步指定国家货盘商品数据到 SQLite (spu_sku.db)

支持多国累积同步（保留已存在数据）：
  - sync(plate_ids, country, clear_first)
      clear_first=True  -> 先 DELETE 整库再插入（重置该次目标）
      clear_first=False -> 仅追加，跳过已存在的 SPU code 与 SKU 组合

国家货盘 plate_id 映射：
  - 美国  : (2, 3, 4, 5)  country="US"
  - 墨西哥: (6,)          country="MX"
  - 日本  : (7,)          country="JP"
  - 韩国  : (8,)          country="KR"
  - 沙特  : (9,)          country="SA"
  - 巴西  : (10,)         country="BR"
  - 英国  : (11,)         country="GB"
  - 加拿大: (12,)         country="CA"
  - 波兰  : (13,)         country="PL"
  - 西班牙: (14,)         country="ES"
  - 德国  : (15,)         country="DE"
  - 澳洲  : (16,)         country="AU"
  - 意大利: (17,)         country="IT"

SPU = categories（code 去重）
SKU = 颜色 × 尺码 粒度：
  - code     = colors.code 颜色映射编码（脏数据兜底为 款号-颜色的英文名，英文名取自 color_detail_images.color_name_web）
  - color    = colors.name 颜色名
  - size     = product_extra.product_size JSON 的尺码列（缺 JSON 回退 size_chart）
  - 肩宽/胸围/衣长/袖长 = product_size JSON 对应列（动态识别表头）
  - 最长边/次长边/最短边 = packaging_spec JSON "包装尺寸（cm）" 列 长*宽*高 拆分排序
  - package_weight = packaging_spec JSON "含包装重量（g）" 列（单位: 克）
  - price    = 该品类 prices 首条价格（颜色尺码粒度无独立价格）
  - img_url_2~5 = 该品类首颜色 web_sku 的图（seq 2~5）
"""
import sqlite3
import psycopg2
import re
import json

PG = dict(host="localhost", port=5432, user="postgres", password="inkreach", dbname="inkreach")
SQLITE = r"C:\Users\Admin\Desktop\test模版\design_agent\pod_trend_agent\db\spu_sku.db"

# 尺码体系按国家区分：亚洲市场(JP/KR)用亚洲尺码，其余默认 欧美尺码/欧美尺码常规
# （库内约定：除 JP/KR 外所有国家 size_group 一律"欧美尺码"，与下方默认一致）
SIZE_SYSTEM = {
    "JP": ("亚洲尺码", "亚洲尺码亚洲常规"),
    "KR": ("亚洲尺码", "亚洲尺码亚洲常规"),
}


def clean(v):
    if v is None:
        return None
    s = str(v).strip()
    return s if s else None


def format_pct(v):
    """成分比例转百分比形式：'100'/'100.0' -> '100%'，空值保持 None"""
    s = clean(v)
    if s is None:
        return None
    try:
        num = float(s)
        if num == int(num):
            return f"{int(num)}%"
        return f"{num}%"
    except ValueError:
        return s


# 中文颜色名 -> 英文（color_name_web 缺失时的兜底）
ZH_TO_EN = {
    "黑": "Black", "黑色": "Black",
    "白": "White", "白色": "White",
    "灰": "Gray", "灰色": "Gray",
    "杏": "Apricot", "杏色": "Apricot",
    "咖色": "Coffee", "咖": "Coffee",
    "藏青": "Navy", "藏青色": "Navy",
    "蓝": "Blue", "蓝色": "Blue", "海蓝": "Blue", "海蓝色": "Ocean Blue",
    "翠绿": "Green", "翠绿色": "Green", "绿": "Green", "绿色": "Green",
    "紫": "Purple", "紫色": "Purple",
    "玫红": "Rose", "玫红色": "Rose", "粉": "Pink", "粉色": "Pink", "粉红": "Pink",
    "红": "Red", "红色": "Red",
    "黄": "Yellow", "黄色": "Yellow",
    "金": "Gold", "金色": "Gold",
    "银": "Silver", "银色": "Silver",
    "棕": "Brown", "棕色": "Brown",
    "米": "Beige", "米色": "Beige", "卡其": "Khaki", "卡其色": "Khaki",
    "橙": "Orange", "橙色": "Orange",
}


def clean_color_name(name):
    """清洗颜色名：
      'S-3XL黑色' -> '黑色'（去开头尺码前缀）
      '灰色（S~XXL）' -> '灰色'（去结尾尺码区间括号）
      '（M~3XL）' -> None（纯尺码无颜色）
    保留中文颜色核心部分。"""
    s = clean(name)
    if not s:
        return None
    # 开头若全是非中文（尺码前缀如 S-3XL / S（4-5）y），取其后中文部分
    m = re.match(r"^[^一-鿿]+([一-鿿].*)$", s)
    if m:
        s = m.group(1).strip()
    # 去掉结尾的尺码区间括号，如 （S~XXL）/（M~5XL）/（XS~3XL）
    s = re.sub(r"[（(][^一-鿿]*[）)]$", "", s).strip()
    # 去掉残留的纯尺码/空白字符
    s = s.strip(" ()（）~　 ")
    return s if s else None


def is_clean_code(ccode, catcode):
    """判断颜色 code 是否规范：必须以 品类code- 开头，
    且后缀为纯字母数字（如 BL01 / ESPRESSO），不含中文、~ 或括号。
    纯尺码区间（如 JPHM009-（S~XXXL））与多色拼接脏数据会被判为非规范。"""
    if not ccode or not catcode:
        return False
    prefix = catcode + "-"
    if not ccode.startswith(prefix):
        return False
    suffix = ccode[len(prefix):]
    if not suffix or re.search(r"[一-鿿~（(）)]", suffix):
        return False
    return True


def first_color_en(name):
    """从可能含多色的脏名称中取首个可识别颜色的英文名，否则原样返回。"""
    if not name:
        return None
    for zh, en in ZH_TO_EN.items():
        if zh in name:
            return en
    return name


def extract_weight(name):
    if not name:
        return None, None
    m = re.search(r"(\d+(?:\.\d+)?)\s*(?:G|g|克)", name)
    if m:
        return m.group(1), "G"
    return None, None


def parse_size_json(raw):
    """product_size JSON -> [{size, shoulder, bust, length, sleeve}...]"""
    if not raw:
        return []
    try:
        data = json.loads(raw)
    except Exception:
        return []
    if not isinstance(data, list) or len(data) < 2:
        return []
    header = [c.get("content", "") for c in data[0]]
    size_i = next((i for i, h in enumerate(header) if "尺码" in h), None)
    sh_i = next((i for i, h in enumerate(header) if "肩宽" in h), None)
    bu_i = next((i for i, h in enumerate(header) if "胸围" in h), None)
    le_i = next((i for i, h in enumerate(header) if "衣长" in h), None)
    sl_i = next((i for i, h in enumerate(header) if "袖长" in h), None)
    if size_i is None:
        return []
    rows = []
    for r in data[1:]:
        if not isinstance(r, list):
            continue
        def cell(i):
            if i is None or i >= len(r):
                return None
            return clean(r[i].get("content", "") if isinstance(r[i], dict) else r[i])
        size = cell(size_i)
        if not size or size == "尺码":
            continue
        rows.append(dict(size=size, shoulder=cell(sh_i), bust=cell(bu_i),
                         length=cell(le_i), sleeve=cell(sl_i)))
    return rows


def parse_pkg_json(raw):
    """packaging_spec JSON -> {size: (longest, second, shortest)} 按 cm 列拆分排序"""
    if not raw:
        return {}
    try:
        data = json.loads(raw)
    except Exception:
        return {}
    if not isinstance(data, list) or len(data) < 2:
        return {}
    header = data[0]
    size_i = next((i for i, c in enumerate(header) if "尺码" in c.get("content", "")), None)
    cm_i = next((i for i, c in enumerate(header)
                 if "包装尺寸" in c.get("content", "") and "cm" in c.get("content", "")), None)
    if size_i is None or cm_i is None:
        return {}
    out = {}
    for r in data[1:]:
        if not isinstance(r, list) or cm_i >= len(r):
            continue
        size = clean(r[size_i].get("content", "")) if isinstance(r[size_i], dict) else clean(r[size_i])
        val = clean(r[cm_i].get("content", "")) if isinstance(r[cm_i], dict) else clean(r[cm_i])
        if not size or not val:
            continue
        parts = re.findall(r"\d+(?:\.\d+)?", val)
        if len(parts) >= 3:
            nums = sorted((float(p) for p in parts[:3]), reverse=True)
            out[size] = (nums[0], nums[1], nums[2])
    return out


def parse_weight_json(raw):
    """packaging_spec JSON -> {size: 含包装重量(g)}，取"含包装重量（g）"列"""
    if not raw:
        return {}
    try:
        data = json.loads(raw)
    except Exception:
        return {}
    if not isinstance(data, list) or len(data) < 2:
        return {}
    header = data[0]
    size_i = next((i for i, c in enumerate(header) if "尺码" in c.get("content", "")), None)
    w_i = next((i for i, c in enumerate(header)
                if "重量" in c.get("content", "") and "（g）" in c.get("content", "")), None)
    if size_i is None or w_i is None:
        return {}
    out = {}
    for r in data[1:]:
        if not isinstance(r, list) or w_i >= len(r):
            continue
        size = clean(r[size_i].get("content", "")) if isinstance(r[size_i], dict) else clean(r[size_i])
        val = clean(r[w_i].get("content", "")) if isinstance(r[w_i], dict) else clean(r[w_i])
        if not size or not val:
            continue
        out[size] = val
    return out


def sync(plate_ids, country, clear_first=True):
    print(f"\n==== 同步 {country} (plate_ids={plate_ids}, clear_first={clear_first}) ====")
    pg = psycopg2.connect(**PG)
    pg.autocommit = True
    pc = pg.cursor()

    # ---------- SPU 数据 ----------
    pc.execute("""
        SELECT c.id, c.plate_id, c.code, c.name, c.sub_category, c.fabric, c.composition, p.display_name
        FROM categories c
        LEFT JOIN plates p ON c.plate_id = p.id
        WHERE c.plate_id IN %s AND c.code IS NOT NULL AND c.code <> ''
        ORDER BY c.plate_id, c.id
    """, (plate_ids,))
    cat_rows = pc.fetchall()

    spu_map = {}
    for cid, plate_id, code, name, sub, fabric, comp, display in cat_rows:
        if code in spu_map:
            prev = spu_map[code]
            prev_score = (0 if prev["sub"] == "组合款" else 1, -prev["plate_id"], -prev["cid"])
            new_score = (0 if sub == "组合款" else 1, -plate_id, -cid)
            if new_score > prev_score:
                spu_map[code] = dict(cid=cid, plate_id=plate_id, name=name, sub=sub,
                                     fabric=fabric, comp_raw=comp, display=display)
        else:
            spu_map[code] = dict(cid=cid, plate_id=plate_id, name=name, sub=sub,
                                 fabric=fabric, comp_raw=comp, display=display)

    pc.execute("""
        SELECT category_id, fabric, comp1, comp1_pct, comp2, comp2_pct, comp3, comp3_pct
        FROM composition
    """)
    comp_by_cat = {}
    for cid, fabric, c1, p1, c2, p2, c3, p3 in pc.fetchall():
        comp_by_cat[cid] = dict(c1=c1, p1=p1, c2=c2, p2=p2, c3=c3, p3=p3)

    pc.execute("""
        SELECT code, english_name, washing_instructions, design_explanation, texture,
               product_size, packaging_spec
        FROM product_extra
    """)
    extra_by_code = {}
    for code, en, wash, design, texture, psize, pkg in pc.fetchall():
        extra_by_code[code] = dict(en=en, wash=wash, design=design, texture=texture,
                                   psize=psize, pkg=pkg)

    # ---------- SKU 数据 ----------
    # 颜色
    pc.execute("""
        SELECT id, category_id, name, code, size_range
        FROM colors
        WHERE category_id IN (SELECT id FROM categories WHERE plate_id IN %s)
        ORDER BY category_id, seq
    """, (plate_ids,))
    colors_by_cat = {}
    for col_id, cid, name, code, sr in pc.fetchall():
        colors_by_cat.setdefault(cid, []).append(dict(id=col_id, name=name, code=code, sr=sr))

    # 颜色英文名（color_detail_images.color_name_web，按 color_id）
    pc.execute("""
        SELECT DISTINCT color_id, color_name_web
        FROM color_detail_images
        WHERE color_id IN (
            SELECT id FROM colors
            WHERE category_id IN (SELECT id FROM categories WHERE plate_id IN %s)
        )
    """, (plate_ids,))
    color_en_by_id = {}
    for col_id, en in pc.fetchall():
        en = clean(en)
        if en:
            color_en_by_id[col_id] = en

    # size_chart 回退
    pc.execute(f"""
        SELECT category_id, size, shoulder, bust, length, sleeve
        FROM size_chart
        WHERE category_id IN (SELECT id FROM categories WHERE plate_id IN %s)
        ORDER BY category_id, seq
    """, (plate_ids,))
    sizechart_by_cat = {}
    for cid, size, sh, bu, le, sl in pc.fetchall():
        sizechart_by_cat.setdefault(cid, []).append(dict(size=size, shoulder=sh, bust=bu,
                                                         length=le, sleeve=sl))

    # prices（首条价格）
    pc.execute("""
        SELECT code, price FROM prices
        WHERE code IN (SELECT DISTINCT code FROM categories WHERE plate_id IN %s)
        ORDER BY code, web_product_id
    """, (plate_ids,))
    price_by_code = {}
    for code, price in pc.fetchall():
        if code not in price_by_code:
            price_by_code[code] = price

    # 图片（按 code，取首个 web_sku 的图）
    pc.execute("""
        SELECT cd.code, cd.web_sku, cd.seq, cd.image_url
        FROM color_detail_images cd
        WHERE cd.code IN (SELECT DISTINCT code FROM categories WHERE plate_id IN %s)
        ORDER BY cd.code, cd.web_sku, cd.seq
    """, (plate_ids,))
    imgs_by_code = {}
    for code, wsku, seq, url in pc.fetchall():
        imgs_by_code.setdefault(code, []).append((wsku, seq, url))

    pg.close()

    # ---------- 写入 SQLite ----------
    db = sqlite3.connect(SQLITE)
    cur = db.cursor()

    if clear_first:
        cur.execute("DELETE FROM SKU")
        cur.execute("DELETE FROM SPU")
        db.commit()
        print("[clear] 已清空 SPU/SKU 旧数据")
    else:
        print("[append] 保留现有数据，仅追加新国家")

    # append 模式：收集已存在的 SPU code 与 SKU 组合，避免重复插入
    existing_spu_codes = set()
    existing_sku_keys = set()
    if not clear_first:
        existing_spu_codes = {r[0] for r in cur.execute("SELECT code FROM SPU")}
        existing_sku_keys = {(r[0], r[1], r[2])
                             for r in cur.execute("SELECT spu_id, code, size FROM SKU")}
        print(f"[append] 已有 SPU {len(existing_spu_codes)} 条, SKU {len(existing_sku_keys)} 条")

    # SPU
    spu_id_by_code = {}
    n_spu = 0
    n_spu_skip = 0
    for code, s in sorted(spu_map.items()):
        # append 模式：若该 SPU code 已存在则跳过（保留首次写入的数据）
        if code in existing_spu_codes:
            spu_id = cur.execute("SELECT id FROM SPU WHERE code=?", (code,)).fetchone()[0]
            spu_id_by_code[code] = spu_id
            n_spu_skip += 1
            continue
        comp = comp_by_cat.get(s["cid"], {}) or {}
        weight, wunit = extract_weight(s["name"])
        # 业务模板固定值（按用户要求，不随 PG 数据变化）
        fixed = dict(
            pattern="印花", details="无", collar_style="圆领", style="休闲",
            care_Instructions="数码印花类可机洗且不可干洗", fabric="微弹",
            target_audience="成人", season="四季", is_transparent="否", layout="常规",
            weaving_method="针织(含钩织、毛织面料)", printing_type="定位印花",
            fabric_texture_1="光面", fabric_weight_unit_1="g/㎡",
            lining_texture="无里料/无内衬", mark="1",
        )
        cur.execute("""
            INSERT INTO SPU (
                code, material, component_1, component_2, component_3,
                component_proportion_1, component_proportion_2, component_proportion_3,
                pattern, details, collar_style, style, care_Instructions, fabric,
                target_audience, season, is_transparent, layout, weaving_method,
                printing_type, fabric_texture_1, fabric_weight_1, fabric_weight_unit_1,
                lining_texture, country, mark
            ) VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)
        """, (
            code, clean(s["fabric"]),
            clean(comp.get("c1")), clean(comp.get("c2")), clean(comp.get("c3")),
            format_pct(comp.get("p1")), format_pct(comp.get("p2")), format_pct(comp.get("p3")),
            fixed["pattern"], fixed["details"], fixed["collar_style"], fixed["style"],
            fixed["care_Instructions"], fixed["fabric"], fixed["target_audience"],
            fixed["season"], fixed["is_transparent"], fixed["layout"], fixed["weaving_method"],
            fixed["printing_type"], fixed["fabric_texture_1"], weight, fixed["fabric_weight_unit_1"],
            fixed["lining_texture"],
            country, fixed["mark"],
        ))
        spu_id_by_code[code] = cur.lastrowid
        n_spu += 1

    # SKU（颜色 × 尺码）
    n_sku = 0
    n_sku_skip = 0
    detail = dict(no_color=0, no_size=0, no_pkg=0, no_img=0, color_sku=0)
    for code, s in sorted(spu_map.items()):
        spu_id = spu_id_by_code[code]
        cat_id = s["cid"]
        extra = extra_by_code.get(code, {})

        # 颜色列表
        color_list = colors_by_cat.get(cat_id, [])
        if not color_list:
            color_list = [dict(name=None, code=code, sr=None)]  # 兜底
            detail["no_color"] += 1
        # 品类尺码范围（取首个非空）
        sr = next((c["sr"] for c in color_list if c["sr"]), None)
        if sr is None:
            sr = None

        # 尺码表：product_size JSON 优先，回退 size_chart
        size_rows = parse_size_json(extra.get("psize"))
        if not size_rows:
            size_rows = sizechart_by_cat.get(cat_id, [])
        if not size_rows:
            detail["no_size"] += 1
        pkg_map = parse_pkg_json(extra.get("pkg"))
        weight_map = parse_weight_json(extra.get("pkg"))
        # 归一化兜底：尺码列可能带英寸后缀（如 '30*40/76.2*101.6' vs '30*40'）
        pkg_map_norm = {k.split("/")[0]: v for k, v in pkg_map.items()}
        weight_map_norm = {k.split("/")[0]: v for k, v in weight_map.items()}

        def lookup_pkg(size):
            v = pkg_map.get(size) or pkg_map_norm.get(size.split("/")[0])
            if v is None and ("one" in size.lower() or "均码" in size):
                v = pkg_map.get("均码") or pkg_map.get("Onesize") or pkg_map.get("OneSize")
            return v

        def lookup_weight(size):
            v = weight_map.get(size) or weight_map_norm.get(size.split("/")[0])
            if v is None and ("one" in size.lower() or "均码" in size):
                v = weight_map.get("均码") or weight_map.get("Onesize") or weight_map.get("OneSize")
            return v

        # 图片
        imgs = imgs_by_code.get(code, [])
        if imgs:
            first_sku = imgs[0][0]
            urls = [u for ws, _, u in imgs if ws == first_sku]
        else:
            urls = []
        if not urls:
            detail["no_img"] += 1
        img2 = urls[1] if len(urls) > 1 else None
        img3 = urls[2] if len(urls) > 2 else None
        img4 = urls[3] if len(urls) > 3 else None
        img5 = urls[4] if len(urls) > 4 else None

        price = price_by_code.get(code)
        # 尺码体系按国家区分：亚洲市场(JP/KR)=亚洲尺码，其余默认 欧美尺码/欧美尺码常规
        sgroup, stype = SIZE_SYSTEM.get(country, ("欧美尺码", "欧美尺码常规"))

        for col in color_list:
            raw_code = col["code"]
            raw_name = col["name"]
            col_id = col["id"]
            cname = clean_color_name(raw_name)          # 清洗颜色名（去尺码前缀/区间）
            # 颜色英文名：优先 PG color_name_web，否则用中文名查兜底字典/取首个颜色
            en = None
            if cname:
                en = color_en_by_id.get(col_id) or first_color_en(cname)
            # 命名规则：
            #   规范 code（品类code-英文/缩写，无中文，如 JPHM009-BL01 / JPTM007-ESPRESSO）
            #     -> 直接使用
            #   缺失/脏 code（含中文或尺码区间）-> 款号(品类code)-颜色的英文名
            if not is_clean_code(raw_code, code):
                ccode = f"{code}-{en}" if en else code
            else:
                ccode = raw_code
            for srow in size_rows:
                size = clean(srow["size"])
                if not size:
                    continue
                # append 模式：跳过已存在的 SKU 组合
                if (spu_id, ccode, size) in existing_sku_keys:
                    n_sku_skip += 1
                    continue
                pkg = lookup_pkg(size)
                pkg_weight = lookup_weight(size)
                cur.execute("""
                    INSERT INTO SKU (
                        spu_id, code, price, color, size, size_group, size_type,
                        shoulder_width, bust, clothing_length, sleeve_length,
                        longest_side, secondary_long_side, shortest_side,
                        package_weight, img_url_2, img_url_3, img_url_4, img_url_5
                    ) VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)
                """, (
                    spu_id, ccode, price, cname, size, sgroup, stype,
                    clean(srow.get("shoulder")), clean(srow.get("bust")),
                    clean(srow.get("length")), clean(srow.get("sleeve")),
                    str(pkg[0]) if pkg else None,
                    str(pkg[1]) if pkg else None,
                    str(pkg[2]) if pkg else None,
                    pkg_weight,
                    img2, img3, img4, img5,
                ))
                n_sku += 1

    db.commit()

    # ---------- 验证 ----------
    print(f"  本次新增 SPU: {n_spu} (跳过已存在 {n_spu_skip})")
    print(f"  本次新增 SKU: {n_sku} (跳过已存在 {n_sku_skip})  (颜色×尺码展开)")
    print(f"  无颜色品类: {detail['no_color']}, 无尺码品类: {detail['no_size']}, 无图品类: {detail['no_img']}")
    print(f"  库内 SPU 总数: {cur.execute('SELECT COUNT(*) FROM SPU').fetchone()[0]}")
    print(f"  库内 SKU 总数: {cur.execute('SELECT COUNT(*) FROM SKU').fetchone()[0]}")
    print(f"  country 分布: {cur.execute('SELECT country, COUNT(*) FROM SPU GROUP BY country').fetchall()}")
    db.close()


if __name__ == "__main__":
    # 追加模式（clear_first=False）：保留现有 US/MX/JP/GB 及手动修改，仅补齐其余 9 国
    # 任何国家若已存在会被跳过（SPU code / SKU 组合去重），不会覆盖手动编辑
    sync((2, 3, 4, 5), "US", clear_first=False)
    sync((6,), "MX", clear_first=False)
    sync((7,), "JP", clear_first=False)
    sync((11,), "GB", clear_first=False)
    # 新增国家
    sync((8,), "KR", clear_first=False)
    sync((9,), "SA", clear_first=False)
    sync((10,), "BR", clear_first=False)
    sync((12,), "CA", clear_first=False)
    sync((13,), "PL", clear_first=False)
    sync((14,), "ES", clear_first=False)
    sync((15,), "DE", clear_first=False)
    sync((16,), "AU", clear_first=False)
    sync((17,), "IT", clear_first=False)
    print("\n✅ 同步完成：13 国数据已包含（US/MX/JP/GB 保留，KR/SA/BR/CA/PL/ES/DE/AU/IT 已追加）")