bollinger_count_debug.py

"""
실리콘투 볼린저밴드 카운팅 디버깅 v5

확정 규칙:
1. 시초가 > 240일선  ← 신규 추가
2. 종가 > BB 상단 × 1.03
3. 60일선 아래 다녀온 후: 시초가 > 이전 피크 종가 (추가 조건)
4. 240일선 아래 → 완전 리셋
5. 60일선 아래 → 카운트 유지
"""

import pandas as pd
import numpy as np
import FinanceDataReader as fdr
import warnings
from datetime import datetime, timezone, timedelta
warnings.filterwarnings('ignore')

_KST = timezone(timedelta(hours=9))

TICKER       = '420770'
DEBUG_FROM   = '2025-05-02'
DEBUG_TO     = datetime.now(_KST).strftime('%Y-%m-%d')   # 오늘(KST) — 매 실행시 최신(장중이면 현재가)
BB_WINDOW    = 20
BB_STD       = 2
MA_240       = 240
MA_60        = 60
BB_THRESHOLD = 1.030

print(f"({TICKER}) 데이터 로드 중...")
df = fdr.DataReader(TICKER, '2022-01-01', DEBUG_TO)
print(f"데이터: {len(df)}일치\n")

df['ma20']     = df['Close'].rolling(BB_WINDOW).mean()
df['std20']    = df['Close'].rolling(BB_WINDOW).std()
df['bb_upper'] = df['ma20'] + BB_STD * df['std20']
df['bb_lower'] = df['ma20'] - BB_STD * df['std20']
df['ma60']     = df['Close'].rolling(MA_60).mean()
df['ma240']    = df['Close'].rolling(MA_240).mean()

count        = 0
peak_close   = 0
was_below_60 = False
counts       = []
peak_list    = []
flags        = []

for i in range(len(df)):
    c    = df['Close'].iloc[i]
    o    = df['Open'].iloc[i]
    u    = df['bb_upper'].iloc[i]
    m60  = df['ma60'].iloc[i]
    m240 = df['ma240'].iloc[i]

    if pd.isna(m240) or pd.isna(u) or pd.isna(m60):
        counts.append(count)
        peak_list.append(peak_close)
        flags.append('데이터부족')
        continue

    if c < m240:
        count        = 0
        peak_close   = 0
        was_below_60 = False
        flag = '240선이탈→리셋'

    elif c < m60:
        was_below_60 = True
        flag = f'60선이탈→유지(count={count})'

    else:
        # 핵심 조건들
        open_above_240 = (o > m240)          # ← 신규: 시초가 > 240일선
        clearly_above  = (c > u * BB_THRESHOLD)  # 종가 > BB상단 × 1.020

        if not open_above_240:
            flag = f'시초가({o:,.0f})<240선({m240:,.0f})→무효'
        elif clearly_above:
            if was_below_60:
                if o > peak_close:
                    count       += 1
                    peak_close   = c
                    was_below_60 = False
                    flag = f'+{count} (60선복귀,시초가>{peak_close:,.0f})'
                else:
                    flag = f'BB돌파but시초가({o:,.0f})≤피크({peak_close:,.0f})→스킵'
            else:
                count      += 1
                peak_close  = max(peak_close, c)
                flag = f'+{count}'
        else:
            pct = (c / u - 1) * 100
            flag = f'종가미달(BB대비{pct:+.1f}%)'

    counts.append(count)
    peak_list.append(peak_close)
    flags.append(flag)

df['bb_count']   = counts
df['peak_close'] = peak_list
df['flag']       = flags

mask     = (df.index >= DEBUG_FROM) & (df.index <= DEBUG_TO)
debug_df = df[mask].copy()

print("=" * 115)
print(f"날짜별 카운팅 ({DEBUG_FROM} ~ {DEBUG_TO})")
print("=" * 115)
print(f"{'날짜':<12} {'종가':>7} {'시가':>7} {'BB상단':>8} {'BB×1.03':>9} "
      f"{'60MA':>7} {'240MA':>7} {'피크':>8} {'카운트':>6}  판정")
print("-" * 115)

prev_count = 0
for date, row in debug_df.iterrows():
    cnt  = int(row['bb_count'])
    flag = row['flag']
    is_notable = (cnt != prev_count or
                  '리셋' in flag or '이탈' in flag or
                  '복귀' in flag or '스킵' in flag or
                  '무효' in flag)
    if is_notable:
        mark = ' ◀' if cnt > prev_count else ''
        print(f"{str(date)[:10]:<12} "
              f"{row['Close']:>7,.0f} "
              f"{row['Open']:>7,.0f} "
              f"{row['bb_upper']:>8,.0f} "
              f"{row['bb_upper']*BB_THRESHOLD:>9,.0f} "
              f"{row['ma60']:>7,.0f} "
              f"{row['ma240']:>7,.0f} "
              f"{row['peak_close']:>8,.0f} "
              f"{cnt:>6}  {flag}{mark}")
    prev_count = cnt

print("\n" + "=" * 60)
print("카운트 증가한 날들 요약")
print("=" * 60)
prev = 0
for date, row in debug_df.iterrows():
    cnt = int(row['bb_count'])
    if cnt > prev:
        print(f"  {str(date)[:10]}: {row['flag']}"
              f"  (종가:{row['Close']:,.0f}, 시가:{row['Open']:,.0f}, "
              f"BB×1.03:{row['bb_upper']*BB_THRESHOLD:,.0f}, "
              f"240MA:{row['ma240']:,.0f})")
    prev = cnt

print(f"\n최대 카운트: {debug_df['bb_count'].max()}")