#!/usr/bin/env python3
"""
小红书线索进线×成单 lag 表分析
数据源：xiaoxi_xhs_lead_detail.csv（陈逸鸫提供）
匹配：手机号→bi_vala_app_account.tel，备UID→bi_vala_app_account.id
成单口径：bi_vala_order order_status IN (3,4), pay_success_date IS NOT NULL
"""
import csv
import sys
import os
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from phone_encrypt import encrypt_phone
import psycopg2
import psycopg2.extras
from collections import defaultdict
from datetime import datetime, date
import openpyxl
from openpyxl.styles import Font, Alignment, PatternFill, Border, Side

# ── Database connection ──
PG_CONFIG = {
    "host": "bj-postgres-16pob4sg.sql.tencentcdb.com",
    "port": 28591,
    "user": "ai_member",
    "password": "LdfjdjL83h3h3^$&**YGG*",
    "dbname": "vala_bi",
}

def get_conn():
    return psycopg2.connect(**PG_CONFIG)

# ── Step 1: Load CSV ──
print("=" * 60)
print("Step 1: Loading CSV...")
leads = []
with open("/root/.openclaw/workspace/tmp/xiaoxi_xhs_lead_detail.csv", "r", encoding="utf-8-sig") as f:
    reader = csv.DictReader(f)
    for row in reader:
        lead_month = row.get("进线月", "").strip()
        if lead_month in ("2026-03", "2026-04", "2026-05"):
            leads.append(row)

print(f"  3-5月进线线索总数: {len(leads)}")

# Extract unique phones and UIDs
phones = set()
uids = set()
for r in leads:
    phone = r.get("手机号", "").strip()
    uid = r.get("用户ID", "").strip()
    if phone and phone != "0":
        phones.add(phone)
    if uid and uid != "0" and uid.isdigit():
        uids.add(int(uid))

print(f"  唯一手机号: {len(phones)}, 唯一UID: {len(uids)}")

# ── Step 2: Match phone → account_id ──
print("\nStep 2: Matching phone → bi_vala_app_account...")
conn = get_conn()
cur = conn.cursor()

phone_to_account = {}
if phones:
    # XXTEA encrypt phones and match against tel_encrypt
    phone_encrypt_map = {}
    for p in phones:
        try:
            phone_encrypt_map[encrypt_phone(p)] = p
        except Exception as e:
            pass
    
    encrypted_list = list(phone_encrypt_map.keys())
    batch_size = 500
    for i in range(0, len(encrypted_list), batch_size):
        batch = encrypted_list[i:i+batch_size]
        placeholders = ",".join(["%s"] * len(batch))
        cur.execute(
            f"SELECT tel_encrypt, id FROM bi_vala_app_account WHERE tel_encrypt IN ({placeholders}) AND status = 1 AND deleted_at IS NULL",
            batch
        )
        for tel_enc, aid in cur.fetchall():
            original_phone = phone_encrypt_map.get(tel_enc)
            if original_phone:
                phone_to_account[original_phone] = aid

print(f"  手机号匹配到账号: {len(phone_to_account)}/{len(phones)}")

# ── Step 3: Match UID → account_id ──
print("\nStep 3: Matching UID → bi_vala_app_account...")
uid_to_account = {}
if uids:
    uid_list = list(uids)
    for i in range(0, len(uid_list), batch_size):
        batch = uid_list[i:i+batch_size]
        placeholders = ",".join(["%s"] * len(batch))
        cur.execute(
            f"SELECT id FROM bi_vala_app_account WHERE id IN ({placeholders}) AND status = 1 AND deleted_at IS NULL",
            batch
        )
        for (aid,) in cur.fetchall():
            uid_to_account[aid] = aid

print(f"  UID匹配到有效账号: {len(uid_to_account)}/{len(uids)}")

# ── Step 4: Build lead → account_id mapping ──
print("\nStep 4: Building lead→account mapping...")
lead_account_map = {}  # lead_row_index → account_id
matched_by_phone = 0
matched_by_uid = 0
unmatched = 0

for idx, r in enumerate(leads):
    phone = r.get("手机号", "").strip()
    uid = r.get("用户ID", "").strip()
    
    aid = None
    match_method = None
    
    # Try phone first
    if phone and phone in phone_to_account:
        aid = phone_to_account[phone]
        match_method = "phone"
        matched_by_phone += 1
    # Fallback to UID
    elif uid and uid.isdigit() and int(uid) in uid_to_account:
        aid = int(uid)
        match_method = "uid"
        matched_by_uid += 1
    else:
        unmatched += 1
    
    lead_account_map[idx] = {"account_id": aid, "match_method": match_method}

print(f"  手机号匹配: {matched_by_phone}, UID匹配: {matched_by_uid}, 未匹配: {unmatched}")
print(f"  总匹配率: {(matched_by_phone + matched_by_uid) / len(leads) * 100:.1f}%")

# ── Step 5: Get all orders for matched accounts ──
print("\nStep 5: Fetching orders for matched accounts...")
matched_aids = set(
    v["account_id"] for v in lead_account_map.values() if v["account_id"] is not None
)
print(f"  去重匹配账号数: {len(matched_aids)}")

# Get orders with order_status 3 or 4, pay_success_date not null
# Also join refund info
aid_list = list(matched_aids)
account_orders = defaultdict(list)  # account_id → list of order dicts

for i in range(0, len(aid_list), batch_size):
    batch = aid_list[i:i+batch_size]
    placeholders = ",".join(["%s"] * len(batch))
    cur.execute(f"""
        SELECT 
            o.account_id,
            o.id as order_id,
            o.pay_success_date,
            o.pay_amount_int,
            o.order_status,
            o.key_from,
            o.trade_no,
            o.out_trade_no,
            r.refund_amount_int,
            r.status as refund_status
        FROM bi_vala_order o
        LEFT JOIN bi_refund_order r ON (
            (o.trade_no = r.trade_no OR o.out_trade_no = r.out_trade_no)
            AND r.status = 3
        )
        WHERE o.account_id IN ({placeholders})
          AND o.pay_success_date IS NOT NULL
          AND o.order_status IN (3, 4)
        ORDER BY o.account_id, o.pay_success_date
    """, batch)
    
    for row in cur.fetchall():
        aid, oid, psd, amt, ost, kf, tn, otn, ref_amt, ref_st = row
        account_orders[aid].append({
            "order_id": oid,
            "pay_success_date": psd,
            "pay_amount_int": amt,
            "order_status": ost,
            "key_from": kf,
            "trade_no": tn,
            "out_trade_no": otn,
            "refund_amount_int": ref_amt or 0,
            "refund_status": ref_st,
        })

total_orders = sum(len(v) for v in account_orders.values())
accounts_with_orders = len(account_orders)
print(f"  有订单的账号数: {accounts_with_orders}, 总订单数: {total_orders}")

cur.close()
conn.close()

# ── Step 6: Build lag table ──
print("\nStep 6: Building lag table...")

# For each lead, determine lead_month and find orders
# lag = (order_month - lead_month) in months

def month_diff(d1, d2):
    """Months between two dates: d2 - d1"""
    return (d2.year - d1.year) * 12 + (d2.month - d1.month)

# Aggregate: lead_month × lag_month
# For each lead_month, count leads, and for each lag bucket count orders
lag_data = defaultdict(lambda: defaultdict(lambda: {
    "lead_count": 0,
    "order_count": 0,
    "order_accounts": set(),
    "gmv": 0,
    "refund": 0,
    "gsv": 0,
}))

# Also track per-lead details for debugging
lead_details = []

for idx, r in enumerate(leads):
    lead_month = r.get("进线月", "").strip()
    lead_date_str = r.get("进线日期", "").strip()
    
    info = lead_account_map[idx]
    aid = info["account_id"]
    
    # Count lead
    lag_data[lead_month]["total"]["lead_count"] += 1
    
    if aid is None:
        lead_details.append({
            "lead_month": lead_month,
            "lead_date": lead_date_str,
            "phone": r.get("手机号", "").strip(),
            "uid": r.get("用户ID", "").strip(),
            "account_id": None,
            "matched": False,
            "match_method": "",
            "has_order": False,
            "order_month": None,
            "lag": None,
            "gmv": 0,
            "refund": 0,
            "gsv": 0,
        })
        continue
    
    # Parse lead date
    try:
        lead_dt = datetime.strptime(lead_date_str, "%Y-%m-%d").date()
    except:
        lead_dt = datetime.strptime(lead_month + "-01", "%Y-%m-%d").date()
    
    orders = account_orders.get(aid, [])
    
    if not orders:
        lead_details.append({
            "lead_month": lead_month,
            "lead_date": lead_date_str,
            "phone": r.get("手机号", "").strip(),
            "uid": r.get("用户ID", "").strip(),
            "account_id": aid,
            "matched": True,
            "match_method": info.get("match_method", "") or "",
            "has_order": False,
            "order_month": None,
            "lag": None,
            "gmv": 0,
            "refund": 0,
            "gsv": 0,
        })
        continue
    
    # For each order, determine lag
    has_any_order = False
    for order in orders:
        psd = order["pay_success_date"]
        if isinstance(psd, datetime):
            order_dt = psd.date()
        else:
            order_dt = psd
        
        lag = month_diff(lead_dt, order_dt)
        
        # Only count orders at or after lead month (lag >= 0)
        if lag < 0:
            continue
        
        has_any_order = True
        order_month = order_dt.strftime("%Y-%m")
        lag_key = f"M{lag}"
        
        lag_data[lead_month][lag_key]["order_count"] += 1
        lag_data[lead_month][lag_key]["order_accounts"].add(aid)
        lag_data[lead_month][lag_key]["gmv"] += order["pay_amount_int"]
        lag_data[lead_month][lag_key]["refund"] += order["refund_amount_int"]
        lag_data[lead_month][lag_key]["gsv"] += (order["pay_amount_int"] - order["refund_amount_int"])
    
    # Record first order for lead detail
    first_order = min(
        [o for o in orders if month_diff(lead_dt, o["pay_success_date"] if isinstance(o["pay_success_date"], datetime) else o["pay_success_date"]) >= 0],
        key=lambda o: o["pay_success_date"],
        default=None
    ) if orders else None
    
    if first_order:
        psd = first_order["pay_success_date"]
        order_dt = psd.date() if isinstance(psd, datetime) else psd
        first_lag = month_diff(lead_dt, order_dt)
        lead_details.append({
            "lead_month": lead_month,
            "lead_date": lead_date_str,
            "phone": r.get("手机号", "").strip(),
            "uid": r.get("用户ID", "").strip(),
            "account_id": aid,
            "matched": True,
            "match_method": info.get("match_method", "") or "",
            "has_order": True,
            "order_month": order_dt.strftime("%Y-%m"),
            "lag": first_lag,
            "gmv": first_order["pay_amount_int"] / 100,
            "refund": first_order["refund_amount_int"] / 100,
            "gsv": (first_order["pay_amount_int"] - first_order["refund_amount_int"]) / 100,
        })
    else:
        lead_details.append({
            "lead_month": lead_month,
            "lead_date": lead_date_str,
            "phone": r.get("手机号", "").strip(),
            "uid": r.get("用户ID", "").strip(),
            "account_id": aid,
            "matched": True,
            "match_method": info.get("match_method", "") or "",
            "has_order": False,
            "order_month": None,
            "lag": None,
            "gmv": 0,
            "refund": 0,
            "gsv": 0,
        })

# ── Step 7: Print summary ──
print("\n" + "=" * 60)
print("Lag Table Summary")
print("=" * 60)

for lead_month in sorted(lag_data.keys()):
    print(f"\n── 进线月: {lead_month} ──")
    total_leads = lag_data[lead_month]["total"]["lead_count"]
    print(f"  留资总数: {total_leads}")
    
    all_lag_keys = sorted(
        [k for k in lag_data[lead_month].keys() if k.startswith("M")],
        key=lambda x: int(x[1:])
    )
    
    cum_orders = 0
    cum_gmv = 0
    cum_refund = 0
    
    for lag_key in all_lag_keys:
        d = lag_data[lead_month][lag_key]
        cum_orders += d["order_count"]
        cum_gmv += d["gmv"]
        cum_refund += d["refund"]
        rate = d["order_count"] / total_leads * 100 if total_leads > 0 else 0
        cum_rate = cum_orders / total_leads * 100 if total_leads > 0 else 0
        print(f"  {lag_key}: 成单{d['order_count']}单 "
              f"({len(d['order_accounts'])}人) "
              f"成单率{rate:.1f}% "
              f"累计{cum_orders}单({cum_rate:.1f}%) "
              f"GMV¥{d['gmv']/100:,.0f} "
              f"退¥{d['refund']/100:,.0f} "
              f"GSV¥{d['gsv']/100:,.0f}")

# ── Step 8: Export to Excel ──
print("\nStep 8: Exporting to Excel...")

wb = openpyxl.Workbook()

# ── Sheet 1: Lag Summary ──
ws1 = wb.active
ws1.title = "Lag汇总表"

header_font = Font(bold=True, size=11)
header_fill = PatternFill(start_color="4472C4", end_color="4472C4", fill_type="solid")
header_font_white = Font(bold=True, size=11, color="FFFFFF")
thin_border = Border(
    left=Side(style='thin'), right=Side(style='thin'),
    top=Side(style='thin'), bottom=Side(style='thin')
)

# Title
ws1.merge_cells("A1:J1")
ws1["A1"] = "小红书线索进线×成单 Lag 表（2026年3-5月）"
ws1["A1"].font = Font(bold=True, size=14)
ws1["A1"].alignment = Alignment(horizontal="center")

ws1.merge_cells("A2:J2")
ws1["A2"] = "数据源：xiaoxi_xhs_lead_detail.csv | 成单口径：bi_vala_order order_status IN (3,4) pay_success_date NOT NULL | 测试账号已剔除"
ws1["A2"].font = Font(size=9, color="666666")
ws1["A2"].alignment = Alignment(horizontal="center")

# Headers
headers = ["进线月", "留资数", "Lag", "成单数", "成单人数", "成单率%", "累计成单数", "累计成单率%", "GMV(元)", "退款(元)", "GSV(元)"]
for col, h in enumerate(headers, 1):
    cell = ws1.cell(row=4, column=col, value=h)
    cell.font = header_font_white
    cell.fill = header_fill
    cell.alignment = Alignment(horizontal="center")
    cell.border = thin_border

row = 5
for lead_month in sorted(lag_data.keys()):
    total_leads = lag_data[lead_month]["total"]["lead_count"]
    all_lag_keys = sorted(
        [k for k in lag_data[lead_month].keys() if k.startswith("M")],
        key=lambda x: int(x[1:])
    )
    
    cum_orders = 0
    
    # First row for this lead_month
    for li, lag_key in enumerate(all_lag_keys):
        d = lag_data[lead_month][lag_key]
        cum_orders += d["order_count"]
        rate = d["order_count"] / total_leads * 100 if total_leads > 0 else 0
        cum_rate = cum_orders / total_leads * 100 if total_leads > 0 else 0
        
        values = [
            lead_month if li == 0 else "",
            total_leads if li == 0 else "",
            lag_key,
            d["order_count"],
            len(d["order_accounts"]),
            round(rate, 1),
            cum_orders,
            round(cum_rate, 1),
            round(d["gmv"] / 100, 2),
            round(d["refund"] / 100, 2),
            round(d["gsv"] / 100, 2),
        ]
        for col, v in enumerate(values, 1):
            cell = ws1.cell(row=row, column=col, value=v)
            cell.border = thin_border
            cell.alignment = Alignment(horizontal="center")
        row += 1
    
    # Total row for this lead_month
    total_orders = cum_orders
    total_rate = total_orders / total_leads * 100 if total_leads > 0 else 0
    total_gmv = sum(lag_data[lead_month][k]["gmv"] for k in all_lag_keys)
    total_refund = sum(lag_data[lead_month][k]["refund"] for k in all_lag_keys)
    total_gsv = sum(lag_data[lead_month][k]["gsv"] for k in all_lag_keys)
    
    total_values = [
        f"{lead_month} 合计",
        total_leads,
        "全部",
        total_orders,
        len(set().union(*[lag_data[lead_month][k]["order_accounts"] for k in all_lag_keys])),
        round(total_rate, 1),
        "",
        "",
        round(total_gmv / 100, 2),
        round(total_refund / 100, 2),
        round(total_gsv / 100, 2),
    ]
    for col, v in enumerate(total_values, 1):
        cell = ws1.cell(row=row, column=col, value=v)
        cell.font = Font(bold=True)
        cell.fill = PatternFill(start_color="D9E2F3", end_color="D9E2F3", fill_type="solid")
        cell.border = thin_border
        cell.alignment = Alignment(horizontal="center")
    row += 1
    row += 1  # blank row

# Adjust column widths
for col in range(1, 12):
    ws1.column_dimensions[openpyxl.utils.get_column_letter(col)].width = 14

# ── Sheet 2: Lead Detail ──
ws2 = wb.create_sheet("线索明细")

detail_headers = ["进线月", "进线日期", "手机号", "用户ID", "匹配账号ID", "匹配方式", "是否成单", "首单月份", "Lag月数", "GMV(元)", "退款(元)", "GSV(元)"]
for col, h in enumerate(detail_headers, 1):
    cell = ws2.cell(row=1, column=col, value=h)
    cell.font = header_font_white
    cell.fill = header_fill
    cell.alignment = Alignment(horizontal="center")
    cell.border = thin_border

for i, ld in enumerate(lead_details, 2):
    match_method = ld.get("match_method", "") or ""
    
    values = [
        ld["lead_month"],
        ld["lead_date"],
        ld["phone"],
        ld["uid"],
        ld["account_id"] or "",
        match_method,
        "是" if ld["has_order"] else "否",
        ld["order_month"] or "",
        ld["lag"] if ld["lag"] is not None else "",
        ld["gmv"],
        ld["refund"],
        ld["gsv"],
    ]
    for col, v in enumerate(values, 1):
        cell = ws2.cell(row=i, column=col, value=v)
        cell.border = thin_border
        cell.alignment = Alignment(horizontal="center")

for col in range(1, 13):
    ws2.column_dimensions[openpyxl.utils.get_column_letter(col)].width = 14

# ── Sheet 3: Match Stats ──
ws3 = wb.create_sheet("匹配统计")

stats = [
    ["指标", "数值"],
    ["3-5月进线线索总数", len(leads)],
    ["手机号匹配成功", matched_by_phone],
    ["UID匹配成功", matched_by_uid],
    ["未匹配", unmatched],
    ["匹配率", f"{(matched_by_phone + matched_by_uid) / len(leads) * 100:.1f}%"],
    ["匹配到去重账号数", len(matched_aids)],
    ["有订单的账号数", accounts_with_orders],
    ["总订单数", total_orders],
]

for i, (k, v) in enumerate(stats, 1):
    ws3.cell(row=i, column=1, value=k).font = Font(bold=True)
    ws3.cell(row=i, column=2, value=v)
    ws3.cell(row=i, column=1).border = thin_border
    ws3.cell(row=i, column=2).border = thin_border

ws3.column_dimensions["A"].width = 25
ws3.column_dimensions["B"].width = 20

# Save
output_path = "/root/.openclaw/workspace/output/xhs_lead_lag_analysis_202603-202605.xlsx"
wb.save(output_path)
print(f"\n✅ 输出文件: {output_path}")
print("Done!")