#!/usr/bin/env python3
"""双激活用户 L1 vs L2 学习进度交叉分布 —— 按L1区间看L2分布"""
import psycopg2

conn = psycopg2.connect(
    host="bj-postgres-16pob4sg.sql.tencentcdb.com",
    port=28591,
    user="ai_member",
    password="LdfjdjL83h3h3^$&**YGG*",
    dbname="vala_bi"
)

cur = conn.cursor()

sql = """
WITH orders AS (
    SELECT o.out_trade_no, o.account_id, o.pay_success_date
    FROM bi_vala_order o
    JOIN bi_vala_app_account a ON o.account_id = a.id
    WHERE o.pay_success_date >= '2026-03-01'
      AND o.order_status != 4
      AND o.pay_success_date IS NOT NULL
      AND a.status = 1
),
l2_accounts AS (
    SELECT DISTINCT o.account_id
    FROM orders o
    JOIN bi_vala_seasonal_ticket t ON o.out_trade_no = t.out_trade_no
    WHERE t.status = 1 AND t.deleted_at IS NULL AND t.season_package_level = 'A2'
),
dual_accounts AS (
    SELECT account_id FROM l2_accounts
    INTERSECT
    SELECT DISTINCT o2.account_id FROM orders o2
    JOIN bi_vala_seasonal_ticket t2 ON o2.out_trade_no = t2.out_trade_no
    WHERE t2.status = 1 AND t2.deleted_at IS NULL AND t2.season_package_level = 'A1'
),
dual_tickets AS (
    SELECT o.out_trade_no, o.account_id, o.pay_success_date,
           t.character_id, t.season_package_level
    FROM orders o
    JOIN bi_vala_seasonal_ticket t ON o.out_trade_no = t.out_trade_no
    WHERE t.status = 1 AND t.deleted_at IS NULL
      AND t.season_package_level IN ('A1', 'A2')
      AND o.account_id IN (SELECT account_id FROM dual_accounts)
),
all_chapter_records AS (
    SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_0 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_1 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_2 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_3 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_4 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_5 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_6 WHERE play_status = 1
    UNION ALL SELECT user_id, chapter_id, created_at FROM bi_user_chapter_play_record_7 WHERE play_status = 1
),
level_records AS (
    SELECT pr.user_id, pr.chapter_id, pr.created_at, l.course_level
    FROM all_chapter_records pr
    JOIN bi_level_unit_lesson l ON pr.chapter_id = l.id
    WHERE l.course_level IN ('L1', 'L2')
),
ticket_progress AS (
    SELECT 
        t.account_id,
        t.character_id,
        t.season_package_level,
        COUNT(DISTINCT lr.chapter_id) AS completed_lessons
    FROM dual_tickets t
    LEFT JOIN level_records lr 
        ON t.character_id = lr.user_id
        AND lr.course_level = CASE 
            WHEN t.season_package_level = 'A1' THEN 'L1' 
            WHEN t.season_package_level = 'A2' THEN 'L2' 
        END
        AND lr.created_at >= t.pay_success_date
        AND lr.created_at <= t.pay_success_date + INTERVAL '30 days'
    GROUP BY t.account_id, t.character_id, t.season_package_level
),
account_level_progress AS (
    SELECT 
        account_id,
        CASE WHEN season_package_level = 'A1' THEN 'L1' ELSE 'L2' END AS course_level,
        MAX(completed_lessons) AS max_completed
    FROM ticket_progress
    GROUP BY account_id, season_package_level
)
SELECT 
    COALESCE(l1.max_completed, 0) AS l1_lessons,
    COALESCE(l2.max_completed, 0) AS l2_lessons,
    COUNT(*) AS account_count
FROM (SELECT DISTINCT account_id FROM dual_accounts) da
LEFT JOIN account_level_progress l1 ON da.account_id = l1.account_id AND l1.course_level = 'L1'
LEFT JOIN account_level_progress l2 ON da.account_id = l2.account_id AND l2.course_level = 'L2'
GROUP BY COALESCE(l1.max_completed, 0), COALESCE(l2.max_completed, 0)
ORDER BY l1_lessons, l2_lessons;
"""

print("正在查询...")
cur.execute(sql)
rows = cur.fetchall()

# 定义 L1 分桶
def bucket_l1(v):
    if v == 0: return (0, "0节")
    elif v <= 5: return (1, "1-5节")
    elif v <= 10: return (2, "6-10节")
    elif v <= 15: return (3, "11-15节")
    elif v <= 20: return (4, "16-20节")
    elif v <= 25: return (5, "21-25节")
    elif v <= 30: return (6, "26-30节")
    else: return (7, "31+节")

# 定义 L2 分桶（更细一些因为L2集中在低段）
def bucket_l2(v):
    if v == 0: return (0, "0节")
    elif v <= 3: return (1, "1-3节")
    elif v <= 5: return (2, "4-5节")
    elif v <= 10: return (3, "6-10节")
    elif v <= 15: return (4, "11-15节")
    elif v <= 20: return (5, "16-20节")
    else: return (6, "21+节")

# 构建交叉表
l1_buckets = [(0,"0节"),(1,"1-5节"),(2,"6-10节"),(3,"11-15节"),(4,"16-20节"),(5,"21-25节"),(6,"26-30节"),(7,"31+节")]
l2_buckets = [(0,"0节"),(1,"1-3节"),(2,"4-5节"),(3,"6-10节"),(4,"11-15节"),(5,"16-20节"),(6,"21+节")]

# 聚合到桶
from collections import defaultdict
cross = defaultdict(lambda: defaultdict(int))
l1_row_total = defaultdict(int)
l2_col_total = defaultdict(int)

for l1_val, l2_val, cnt in rows:
    l1b = bucket_l1(l1_val)[0]
    l2b = bucket_l2(l2_val)[0]
    cross[l1b][l2b] += cnt
    l1_row_total[l1b] += cnt
    l2_col_total[l2b] += cnt

total = sum(l1_row_total.values())

# 打印
print(f"\n{'='*80}")
print(f"双激活用户 L1 学习进度 × L2 学习进度 交叉分布（{total} 人）")
print(f"{'='*80}\n")

# 表头
header = f"{'L1 ↓ / L2 →':>14}"
for _, l2l in l2_buckets:
    header += f" | {l2l:>8}"
header += f" | {'L1合计':>8}"
print(header)
print("-" * len(header))

for l1k, l1l in l1_buckets:
    line = f"{l1l:>14}"
    row_total = l1_row_total[l1k]
    for l2k, _ in l2_buckets:
        val = cross[l1k][l2k]
        if row_total > 0:
            pct = val / row_total * 100
            line += f" | {val:>4}({pct:>4.0f}%)"
        else:
            line += f" | {val:>4}(  0%)"
    line += f" | {row_total:>8}"
    print(line)

# 列合计行
print("-" * len(header))
line = f"{'L2合计':>14}"
for l2k, _ in l2_buckets:
    line += f" | {l2_col_total[l2k]:>8}"
line += f" | {total:>8}"
print(line)

# 对每个 L1 区间补充：L2 的均值和中位数
print(f"\n{'='*80}")
print("各 L1 区间下的 L2 学习进度摘要")
print(f"{'='*80}")
print(f"{'L1区间':>12} | {'L2均值':>8} | {'L2中位':>8} | {'L2=0节%':>8} | {'L2≤3节%':>9}")
print("-" * 55)

import statistics
for l1k, l1l in l1_buckets:
    vals = []
    for l1_val, l2_val, cnt in rows:
        if bucket_l1(l1_val)[0] == l1k:
            vals.extend([l2_val] * cnt)
    if vals:
        avg = sum(vals) / len(vals)
        med = statistics.median(vals)
        zero = sum(1 for v in vals if v == 0) / len(vals) * 100
        le3 = sum(1 for v in vals if v <= 3) / len(vals) * 100
        print(f"{l1l:>12} | {avg:>7.1f} | {med:>7.0f} | {zero:>7.1f}% | {le3:>8.1f}%")

cur.close()
conn.close()
print("\n查询完成。")