stockstats

Indicators are accessed using specific string patterns in the DataFrame. You can control the column and the window size using these patterns:

1. Full Control: <column>_<window>_<indicator>

   • high_5_sma: 5-period simple moving average of the high price.
   • close_10_ema: 10-period exponential moving average of the close price.
   • high_-1_d: 1-period delta of the high price (negative window looks backward).

2. Indicator with varying window: <indicator>_<window>

   • rsi_6: 6-period RSI.
   • cci_10: 10-period CCI.
   • atr_13: 13-period ATR.

Install the stockstats package using pip to enable inline stock indicator calculations on pandas DataFrames.

pip install stockstats

The StockDataFrame class acts as a wrapper for a pandas.DataFrame. You can initialize it using the wrap function or StockDataFrame.retype.

To ensure compatibility, your DataFrame should include the following columns (case-insensitive):

• close: the close price of the period
• high: the highest price of the interval
• low: the lowest price of the interval
• volume: the volume of stocks traded during the interval
• date: timestamp of the record (optional, used as index by default)

You can use unwrap to convert the StockDataFrame back into a standard pandas.DataFrame.

import pandas as pd
from stockstats import wrap

# from CSV
df = wrap(pd.read_csv('stock.csv'))

# from yfinance (disable multi-level index for compatibility)
import yfinance as yf
df = wrap(yf.download('AAPL', multi_level_index=False))

Some indicators automatically generate multiple related columns. To ensure these columns are calculated and added to the DataFrame, use the .get() method on the StockDataFrame before attempting to access the individual columns.

# MACD generates three columns at once
df.get('macd')
print(df[['macd', 'macds', 'macdh']].tail())

# Bollinger Bands
df.get('boll')
print(df[['boll', 'boll_ub', 'boll_lb']].tail())

You can detect signal crossovers (e.g., a fast moving average crossing a slow moving average) using specific naming conventions for column access. The syntax for a cross-up (golden cross) is {column1}_{window1}_{indicator}_{direction}_{column2}_{window2}_{indicator}.

Example: df['close_10_sma_xu_close_50_sma'] detects when the 10-period SMA of the close column crosses above the 50-period SMA of the close column.

# cross-over detection
golden_cross = df['close_10_sma_xu_close_50_sma']  # 10 SMA crosses above 50 SMA

Some indicators have default window sizes or parameters that can be changed globally using set_dft_window(indicator_name, parameters).

Note: Changes are global and will not affect existing columns. To re-evaluate indicators with new parameters, you must remove the existing columns first.

Examples:

• Single value (window): set_dft_window('rsi', 14)
• Tuple of values: set_dft_window('macd', (12, 26, 9)) or set_dft_window('kama', (10, 5, 34))

set_dft_window('tema', n)
set_dft_window('kama', (10, 5, 34))
set_dft_window('macd', (12, 26, 9))
set_dft_window('rsi', n)

The library provides built-in access to common financial calculations:

• Typical Price: Accessed via df['middle']. It is the average of high, low, and close. If amount (total cash flow) is available, it uses amount / volume for higher accuracy.
• Log Return: Accessed via df['log-ret']. Calculated as ln(close / last_close).

You can calculate price changes (deltas) and shift columns using string patterns:

Delta (Difference between periods)

• <column>_<window>_d: Difference between current and $n$ periods ago.
• <column>_delta: Shortcut for <column>_-1_d (current vs previous).
• Example: df['close_delta'] or df['high_2_d'] (high price delta between current and 2 days later).

Shift (Moving data forward/backward)

Shift columns backward or forward. Negative values shift forward.

# Example of shifting columns
# df[['close', 'close_-1_s', 'close_2_s']]

df['close_delta']
df['close_-1_d']
df['high_2_d']

The library provides patterns to detect when two columns cross or satisfy comparison logic:

Cross-overs

• <A>_xu_<B>: A crosses up B.
• <A>_xd_<B>: A crosses down B.
• <A>_x_<B>: A crosses B (either direction).

Comparisons

Use standard comparison operators between two columns A and B:

• <A>_le_<B> (Less than or equal)
• <A>_ge_<B> (Greater than or equal)
• <A>_lt_<B> (Less than)
• <A>_gt_<B> (Greater than)
• <A>_eq_<B> (Equal)
• <A>_ne_<B> (Not equal)

kdjk_xu_kdjd
kdjk_xd_kdjd
kdjk_x_kdjd
close_ge_70

To quickly generate a large number of technical indicators (including shortcuts like KDJ, BOLL, and MFI), use the df.init_all() method.

Warning: This operation generates many columns and should be used with caution.

You can count how many times a condition was met within a specific rolling window using the pattern <column>_<window>_c.

Steps:

1. Create a boolean column based on a condition (e.g., df['res'] = df['middle'] > df['close']).
2. Access the count using the pattern on that boolean column.

Example: Count how many times typical price was larger than close in the past 10 periods

tp = df['middle']
df['res'] = df['middle'] > df['close']
# Access the count of True values in the last 10 periods
count_column = df['res_10_c']

Indicators are calculated lazily upon first access. You can access indicators using their name as a column key. For indicators requiring a window (period), you can use the default window by using the indicator name, or specify a custom window using the format {indicator}_{window}.

Example:

• df['rsi'] uses the default 14-period RSI.
• df['rsi_6'] uses a 6-period RSI.
• df['close_20_sma'] calculates a 20-period SMA on the close column.
• df['high_10_ema'] calculates a 10-period EMA on the high column.

# indicators with default windows
rsi = df['rsi']           # 14-period RSI (default)
rsi6 = df['rsi_6']        # 6-period RSI

# moving averages on any column
sma = df['close_20_sma']  # 20-period SMA of close
ema = df['high_10_ema']   # 10-period EMA of high