Spaces:
Running
Running
File size: 6,465 Bytes
d81c653 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 |
import pandas as pd
import numpy as np
def load_data(ticker):
try:
# Load data from CSV
data = pd.read_csv(f'tickers/{ticker}.csv', index_col="Date", parse_dates=['Date'])
data.sort_index(inplace=True) # Ensure data is sorted by date
return data
except FileNotFoundError:
print(f"Data for {ticker} not found.")
return None
def sma(data, period):
return data['Close'].rolling(window=period).mean()
def ema(data, period):
return data['Close'].ewm(span=period, adjust=False).mean()
def score_downward_trend(data, window=4):
"""
Score the downward trend based on price action and volume.
"""
if len(data) < window:
return 0 # Not enough data
score = 0
for j in range(1, window):
if data['High'].iloc[j] < data['High'].iloc[j-1]:
score += 1 # Increment score for each lower high
if data['Close'].iloc[j] < data['Close'].iloc[j-1]:
score += 1 # Increment score for each lower close
if data['Volume'].iloc[j] > data['Volume'].iloc[j-1]:
score += 0.5 # Increment score for increasing volume during downtrend
return score
def score_candle(candle, data, i):
"""
Score the candle based on its pattern, volume, and position relative to moving averages.
"""
open_price, close_price, low_price, high_price = candle['Open'], candle['Close'], candle['Low'], candle['High']
prev_candle = data.iloc[i-1]
score = 0
body = abs(close_price - open_price)
bottom_wick_length = min(open_price, close_price) - low_price
top_wick_length = high_price - max(open_price, close_price)
# Add 16 points if candle is green and there's a significant gap
if close_price > open_price and low_price > prev_candle['High']:
score += 16
# Rest of the existing scoring logic
if bottom_wick_length > 2 * body:
score += 10
if abs(open_price - close_price) < (0.1 * (high_price - low_price)):
score += 15
if bottom_wick_length > top_wick_length:
score += 8
# Volume analysis
if candle['Volume'] > data['Volume'].rolling(window=20).mean().iloc[i]:
score += 5
# Moving average analysis
ema_20 = ema(data.iloc[:i+1], 20).iloc[-1]
sma_50 = sma(data.iloc[:i+1], 50).iloc[-1]
sma_200 = sma(data.iloc[:i+1], 200).iloc[-1]
if close_price > ema_20 and open_price < ema_20:
score += 5
if close_price > sma_50:
score += 3
if close_price > sma_200:
score += 2
# Momentum indicator
rsi = calculate_rsi(data.iloc[:i+1], period=14).iloc[-1]
if rsi < 30:
score += 5
penalty = 0
conditions_met = 0
if candle['High'] > prev_candle['High']:
conditions_met += 1
if candle['Low'] > prev_candle['High']:
conditions_met += 1
if candle['Close'] > max(prev_candle['Close'], prev_candle['Open']):
conditions_met += 1
if candle['Open'] > max(prev_candle['Open'], prev_candle['Close']):
conditions_met += 1
current_avg = (candle['Open'] + candle['Close'] + candle['High'] + candle['Low']) / 4
prev_avg = (prev_candle['Open'] + prev_candle['Close'] + prev_candle['High'] + prev_candle['Low']) / 4
if current_avg > prev_avg:
conditions_met += 1
if conditions_met == 3:
penalty = -10
elif conditions_met == 4:
penalty = -12
elif conditions_met == 5:
penalty = -17
score += penalty
return score
def calculate_rsi(data, period=14):
delta = data['Close'].diff()
gain = (delta.where(delta > 0, 0)).rolling(window=period).mean()
loss = (-delta.where(delta < 0, 0)).rolling(window=period).mean()
rs = gain / loss
return 100 - (100 / (1 + rs))
def calculate_risk_reward(data, entry_index, stop_loss_percent=0.02, target_percent=0.06):
entry_price = data['Close'].iloc[entry_index]
stop_loss = entry_price * (1 - stop_loss_percent)
target = entry_price * (1 + target_percent)
risk = entry_price - stop_loss
reward = target - entry_price
return reward / risk
def find_reversal_patterns(data, window=4, candle_score_threshold=20, trend_score_threshold=5, risk_reward_threshold=2):
patterns = []
for i in range(window, len(data)):
trend_score = score_downward_trend(data.iloc[i-window:i], window=window)
if trend_score >= trend_score_threshold:
candle_score = score_candle(data.iloc[i], data, i)
if candle_score >= candle_score_threshold:
risk_reward = calculate_risk_reward(data, i)
if risk_reward >= risk_reward_threshold:
format_date = data.index[i].strftime('%Y-%m-%d')
patterns.append((format_date, trend_score, candle_score, risk_reward))
return patterns
def back_reversal_finder(data, window=4, candle_score_threshold=20, trend_score_threshold=4.5, risk_reward_threshold=1.5):
patterns = []
for i in range(window, len(data)):
trend_score = score_downward_trend(data.iloc[i-window:i], window=window)
if trend_score >= trend_score_threshold:
candle_score = score_candle(data.iloc[i], data, i)
if candle_score >= candle_score_threshold:
risk_reward = calculate_risk_reward(data, i)
if risk_reward >= risk_reward_threshold:
format_date = data.index[i].strftime('%Y-%m-%d')
patterns.append(format_date)
return patterns
def check_for_reversal_patterns(ticker, window=4, candle_score_threshold=20, trend_score_threshold=5, risk_reward_threshold=2):
data = load_data(ticker)
if data is None:
return
patterns = find_reversal_patterns(data, window=window, candle_score_threshold=candle_score_threshold,
trend_score_threshold=trend_score_threshold, risk_reward_threshold=risk_reward_threshold)
if patterns:
print(f"{ticker}: Potential reversal patterns found:")
for date, trend_score, candle_score, risk_reward in patterns:
print(f"Date: {date}, Trend Score: {trend_score:.2f}, Candle Score: {candle_score:.2f}, Risk-Reward: {risk_reward:.2f}")
else:
print(f"{ticker}: No clear reversal patterns detected.")
def main():
ticker = input("Enter Ticker: ").upper()
check_for_reversal_patterns(ticker)
if __name__ == '__main__':
main() |