Create FibonacciLevels.py

indicators/FibonacciLevels.py

@@ -0,0 +1,116 @@
+import pandas as pd
+import numpy as np
+
+def FibonacciLevels(dataframe, fibo_range=1, x=0, fibo_level=61.8, custom_fibo_level=None, start_date=None):
+    """
+    Calculate Fibonacci levels for price ranges based on input parameters.
+
+    Parameters:
+    dataframe (DataFrame): Input DataFrame with columns ["high", "low", "open", "close", "volume"].
+    fibo_range (int): Fibo range mode [1-10] (default is 1).
+        - 1  : High-Low previous day
+        - 2  : High-low previous week
+        - 3  : High-low previous month
+        - 4  : High-Low of last X days
+        - 5  : Open-Close previous day
+        - 6  : Open-Close previous week
+        - 7  : Open-Close previous month
+        - 8  : Open-Close of last X days
+        - 9  : Highest-Lowest for last X bars back
+        - 10 : Open-Close for last X bars back    
+    x (int): Custom days/bars count (default is 0).
+    fibo_level (float): Default Fibonacci level (default is 61.8).
+    custom_fibo_level (float): Custom Fibonacci level (default is None).
+    start_date (datetime): Start point for calculations (default is None).
+
+    Returns:
+    DataFrame: A new DataFrame with the calculated Fibonacci levels added as a "fibo_level" column.
+    """
+    df = dataframe.copy()
+
+    # Initialize variables
+    tf_end_time = 0
+    bars_used = -1
+    prev_tf_open = 0
+    prev_tf_high = 0
+    prev_tf_low = 0
+    prev_tf_close = 0
+    fibo_level = fibo_level
+    fibo_range_used = fibo_range
+    start_date_used = False
+
+    # Create an empty array to store Fibonacci levels
+    buffer = np.empty(len(df))
+    buffer[:] = np.nan
+
+    for i in range(len(df)):
+        if _is_new_tf_start(df.index[i], start_date_used, start_date, fibo_range_used, tf_end_time, bars_used, x):
+            upper_value, lower_value = _calculate_price_range(df, i, fibo_range_used, prev_tf_open, prev_tf_close, prev_tf_high, prev_tf_low)
+
+            if not np.isnan(upper_value) and not np.isnan(lower_value):
+                percent_step = (upper_value - lower_value) / 100
+                delta = fibo_level * percent_step
+                bullish = prev_tf_close > prev_tf_open
+                fibo_level = (upper_value - delta) if bullish else (lower_value + delta)
+
+            prev_tf_open = df["open"].iloc[i]
+            prev_tf_high = df["high"].iloc[i]
+            prev_tf_low = df["low"].iloc[i]
+            prev_tf_close = df["close"].iloc[i]
+            bars_used = 0 if i == 0 else 1
+
+        else:
+            if i != 0:
+                prev_tf_high = max(prev_tf_high, df["high"].iloc[i])
+                prev_tf_low = min(prev_tf_low, df["low"].iloc[i])
+                prev_tf_close = df["close"].iloc[i]
+                bars_used += 1
+
+        buffer[i] = fibo_level
+
+    return buffer
+
+def _is_new_tf_start(time, start_date_used, start_date, fibo_range_used, tf_end_time, bars_used, x):
+    if start_date is not None and not start_date_used and start_date <= time:
+        _set_end_time(time, fibo_range_used, tf_end_time)
+        return True
+
+    if fibo_range_used in [9, 10]:
+        if bars_used == -1 or bars_used == x:
+            return True
+        else:
+            return False
+
+    if fibo_range_used in [1, 2, 3, 4, 5, 6, 7, 8]:
+        if tf_end_time == 0 or tf_end_time <= time:
+            _set_end_time(time, fibo_range_used, tf_end_time)
+            return True
+        else:
+            return False
+
+    return False
+
+def _set_end_time(time, fibo_range_used, tf_end_time):
+    cur_day_start = pd.Timestamp(time).replace(hour=0, minute=0, second=0, microsecond=0)
+
+    if fibo_range_used in [1, 5]:
+        tf_end_time = cur_day_start + pd.Timedelta(days=1)
+    elif fibo_range_used in [2, 6]:
+        cur_day_start = cur_day_start - pd.Timedelta(days=cur_day_start.dayofweek)
+        tf_end_time = cur_day_start + pd.Timedelta(weeks=1)
+    elif fibo_range_used in [3, 7]:
+        cur_day_start = cur_day_start - pd.Timedelta(days=cur_day_start.day - 1)
+        tf_end_time = cur_day_start + pd.Timedelta(days=cur_day_start.days_in_month)
+
+def _calculate_price_range(df, i, fibo_range_used, prev_tf_open, prev_tf_close, prev_tf_high, prev_tf_low):
+    if fibo_range_used in [1, 2, 3, 4, 9]:
+        upper_value = prev_tf_high
+        lower_value = prev_tf_low
+    elif fibo_range_used in [5, 6, 7, 8, 10]:
+        upper_value = max(prev_tf_open, prev_tf_close)
+        lower_value = min(prev_tf_open, prev_tf_close)
+    else:
+        upper_value = np.nan
+        lower_value = np.nan
+
+    return upper_value, lower_value