import psycopg2

conn = psycopg2.connect(
    host="192.168.0.61",
    port=55432,
    database="njts_acct",
    user="njts_app",
    password="njts_app2025"
)

cur = conn.cursor()

print("=" * 140)
print("【识别可能的6月→7月/8月复制数据】")
print("=" * 140)

# 第1步：找出6月所有有的仕訳（描述+金额的组合）
print("\n第1步：提取6月的所有仕訳（描述和金额作为KEY）")

cur.execute("""
    SELECT DISTINCT
        je.description,
        COALESCE(SUM(jl.debit), 0)::BIGINT as total_debit,
        COALESCE(SUM(jl.credit), 0)::BIGINT as total_credit
    FROM journal_entries je
    LEFT JOIN journal_lines jl ON je.journal_entry_id = jl.journal_entry_id
    WHERE je.is_deleted = false
      AND je.is_latest = true
      AND EXTRACT(YEAR FROM je.entry_date) = 2025
      AND EXTRACT(MONTH FROM je.entry_date) = 6
    GROUP BY je.journal_entry_id, je.description
    ORDER BY je.description
""")

june_patterns = cur.fetchall()
print(f"找到 {len(june_patterns)} 个6月的仕訳模式")

# 第2步：在7月和8月中查找相同的仕訳
print("\n第2步：在7月和8月中查找相同描述和金额的仕訳")

duplicates_for_deletion = []

for description, debit, credit in june_patterns:
    cur.execute("""
        SELECT 
            je.journal_entry_id,
            TO_CHAR(je.entry_date, 'YYYY-MM-DD') as entry_date,
            EXTRACT(MONTH FROM je.entry_date)::INT as month,
            COALESCE(SUM(jl.debit), 0)::BIGINT as calc_debit,
            COALESCE(SUM(jl.credit), 0)::BIGINT as calc_credit
        FROM journal_entries je
        LEFT JOIN journal_lines jl ON je.journal_entry_id = jl.journal_entry_id
        WHERE je.is_deleted = false
          AND je.is_latest = true
          AND EXTRACT(YEAR FROM je.entry_date) = 2025
          AND EXTRACT(MONTH FROM je.entry_date) IN (7, 8)
          AND je.description = %s
        GROUP BY je.journal_entry_id, je.entry_date
        HAVING COALESCE(SUM(jl.debit), 0) = %s
          AND COALESCE(SUM(jl.credit), 0) = %s
    """, (description, debit, credit))
    
    matches = cur.fetchall()
    if matches:
        for entry_id, entry_date, month, calc_debit, calc_credit in matches:
            duplicates_for_deletion.append({
                'entry_id': entry_id,
                'date': entry_date,
                'month': month,
                'description': description,
                'debit': debit,
                'credit': credit
            })

# 按月份排序
duplicates_for_deletion.sort(key=lambda x: (x['month'], x['entry_id']))

print(f"\n找到 {len(duplicates_for_deletion)} 个可能的复制记录（7月/8月中与6月相同的）")

if duplicates_for_deletion:
    print("\n【可能的复制记录列表】\n")
    
    current_month = None
    for dup in duplicates_for_deletion:
        if dup['month'] != current_month:
            print(f"\n--- {dup['month']}月 ---")
            current_month = dup['month']
        
        print(f"ID={dup['entry_id']:3d} | {dup['date']} | 借={dup['debit']:>12,}  貸={dup['credit']:>12,}")
        print(f"         └─ {dup['description'][:80]}")

print("\n" + "=" * 140)
print(f"【预期删除】{len(duplicates_for_deletion)} 条记录")
print("=" * 140)

# 计算删除前后的普通預金期末残高变化
cur.execute("""
    SELECT 
        COALESCE(SUM(COALESCE(jl.debit, 0) - COALESCE(jl.credit, 0)), 0) as balance
    FROM journal_lines jl
    INNER JOIN journal_entries je ON jl.journal_entry_id = je.journal_entry_id
    WHERE jl.account_id = 2  -- 普通預金
    AND je.is_deleted = false
    AND je.is_latest = true
""")

current_balance = cur.fetchone()[0]
print(f"\n删除前普通預金期末残高: {current_balance:>15,.0f}円")

# 计算这些重复记录对普通預金的影响
deletion_impact = 0
for dup in duplicates_for_deletion:
    cur.execute("""
        SELECT COALESCE(SUM(COALESCE(jl.debit, 0) - COALESCE(jl.credit, 0)), 0)
        FROM journal_lines jl
        WHERE jl.journal_entry_id = %s
        AND jl.account_id = 2  -- 普通預金
    """, (dup['entry_id'],))
    
    impact = cur.fetchone()[0]
    deletion_impact += impact

expected_balance_after = current_balance - deletion_impact

print(f"这些记录对普通預金的总影响: {deletion_impact:>15,.0f}円")
print(f"删除后预期普通預金期末残高: {expected_balance_after:>15,.0f}円")
print(f"\n期望目标值: 14,916,322円")
print(f"删除后vs目标值差距: {14_916_322 - expected_balance_after:>15,.0f}円")

cur.close()
conn.close()

# 保存删除列表到Python变量以供后续使用
print("\n" + "=" * 140)
print(f"识别完成。找到 {len(duplicates_for_deletion)} 条可能的复制记录")
print("=" * 140)

# 返回delete_list供后续使用
import json
print("\n【识别的删除对象IDs】")
delete_ids = [dup['entry_id'] for dup in duplicates_for_deletion]
print(delete_ids)