Python - AttributeError: module 'pandas' has no attribute 'ewm'

Question

I'm trying to run the below code but I'm getting an error.

line 43, in <module> ups_avg = pd.ewm(ups, span=RSI_N)[-1] 
AttributeError: module 'pandas' has no attribute 'ewm

I'm using the latest pandas module. How do I fix line 43 please? Thanks

from binance.client import Client    
import numpy as np    
import pandas as pd    
import smtplib    
import time    
import yaml    

CONFIG = yaml.load(open('./CONFIG.yml'))    
API_KEY = CONFIG['binance_api']['key']    
API_SECRET = CONFIG['binance_api']['secret']    
user = CONFIG['gmail']['user']    
passwd = CONFIG['gmail']['password']    

client = Client(API_KEY, API_SECRET)

# against ETH    
SYMBOLS = ('ADA', 'ADX', 'BAT', 'BCC', 'DASH', 'EOS', 'IOTA',

        'LTC', 'NEO', 'OMG', 'STORJ', 'XLM', 'NANO', 'XRP', 'XVG', 'ZEC')    
RSI_N = 14    
RSI_THRESHOLD = 8    
RUN_INTERVAL_MINS = 30

def send_email(rsi_values):    
    if len(rsi_values) > 0:        
        message = '\n'.join('{0:>8} {1:.2f}'.format(symbol, rsi) for (symbol, rsi) in rsi_values)    
        email_text = 'From: {0}\nTo: {1}\nSubject: Stock Recommendations\n\n{2}'.format(user, user, message)

        try:    
            server = smtplib.SMTP_SSL('smtp.gmail.com', 465)    
            server.ehlo()    
            server.login(user, passwd)    
            server.sendmail(user, user, email_text)    
            server.close()    
        except:    
            pass

while True:    
    rsi_values = []    
    for SYMBOL in SYMBOLS:    
        klines = client.get_historical_klines(SYMBOL + 'ETH', Client.KLINE_INTERVAL_30MINUTE, '{} hours ago UTC'.format((RSI_N + 3) // 2))    
        closings = np.asarray(klines, dtype=np.float)[-RSI_N - 1:, 4]        
        diffs = np.diff(closings)    
        ups = diffs.clip(min=0)    
        downs = diffs.clip(max=0)    
        ups_avg = pd.ewma(ups, span=RSI_N)[-1]    
        downs_avg = -pd.ewma(downs, span=RSI_N)[-1]    
        rs = ups_avg / downs_avg    
        rsi = 100 - 100 / (1 + rs)    
        rsi_values.append((SYMBOL, rsi))

    print('\n'.join('{0:>8} {1:.2f}'.format(symbol, rsi) for (symbol, rsi) in rsi_values))    
    rsi_values = list(filter(lambda x: x[1] < RSI_THRESHOLD, rsi_values))        
    send_email(rsi_values)        
    time.sleep(60 * RUN_INTERVAL_MINS)

Answer 1

use object dataframe.ewm().mean()

ups.ewm(span=RSI_N)
Eg.
df = DataFrame({'B': [0, 1, 2, np.nan, 4]})
df.ewm(com=0.5).mean()

Answer 2

I had a similar issue with trying to calculate the exponentially weighted moving average on a pandas dataframe. I am assuming you are trying to do the same with "ups_avg" being an exponentially weighted average of "ups".

This solved the problem for me, illustrating with a dummy dataframe:

df_test = pd.DataFrame(data = np.arange(0,100))
df_test_ewma = df_test.ewm(span=2).mean()