Pandas 2.1.0 and mfi indicator

[zakcali]

Which version are you running? The lastest version is on Github. Pip is for major releases.
pandas-ta 0.3.14b
python 3.12
pandas 2.1.0
Do you have TA Lib also installed in your environment?
No

Have you tried the development version? Did it resolve the issue?
no, because is not compatible with python 3.12

Describe the bug
[A clear and concise description of what the bug is.](futurewarning: setting an item of incompatible dtype is deprecated and will raise in a future error of pandas)

To Reproduce

df["MFI"] = ta.mfi(high=df["High"], low=df["Low"], close=df["Close"], volume=df["Volume"], length=14, talib=False)

Expected behavior
pandas 2.0.3 doesn't give that warning

Screenshots
Reason is updated pandas behaviour

Additional context
I find a function from: https://gist.github.com/quantra-go-algo/d3ced509e55ceb15eb0a6bd283e6853f#file-ti_mfi-py
and fine tuned with chatgpt to execute faster than original code and pandas-ta:
I couldn't correct the pandas-ta bug, so I embedded below code to my python program

def mfi(high, low, close, volume, n):
    typical_price = (high + low + close) / 3
    money_flow = typical_price * volume
    mf_sign = np.where(typical_price > typical_price.shift(1), 1, -1)
    signed_mf = money_flow * mf_sign

    # Calculate gain and loss using vectorized operations
    positive_mf = np.where(signed_mf > 0, signed_mf, 0)
    negative_mf = np.where(signed_mf < 0, -signed_mf, 0)

    mf_avg_gain = pd.Series(positive_mf).rolling(n, min_periods=1).sum()
    mf_avg_loss = pd.Series(negative_mf).rolling(n, min_periods=1).sum()

    return (100 - 100 / (1 + mf_avg_gain / mf_avg_loss)).to_numpy()

Thanks for using Pandas TA!

[twopirllc]

Hello @zakcali,

I couldn't correct the pandas-ta bug, so I embedded below code to my python program

Nice! Glad you find a work around and shared the code. 😎

Like many indicators, including mfi, are in need of an update. So this really helps! Thanks 😎

Kind Regards,
KJ

[zakcali]

this one runs faster, by using numpy arrays instead of pandas.series on my 376 items xu100 market. Please ensure that you are not sending an empty array.

Calculate money flow index

def calculate_mfi(high, low, close, volume, period):
    typical_price = (high + low + close) / 3
    money_flow = typical_price * volume
    mf_sign = np.where(typical_price > np.roll(typical_price, shift=1), 1, -1)
    signed_mf = money_flow * mf_sign

    # Calculate gain and loss using vectorized operations
    positive_mf = np.maximum(signed_mf, 0)
    negative_mf = np.maximum(-signed_mf, 0)

    mf_avg_gain = np.convolve(positive_mf, np.ones(period), mode='full')[:len(positive_mf)] / period
    mf_avg_loss = np.convolve(negative_mf, np.ones(period), mode='full')[:len(negative_mf)] / period

    epsilon = 1e-10  # Small epsilon value to avoid division by zero
    mfi = 100 - 100 / (1 + mf_avg_gain / (mf_avg_loss + epsilon))
    return mfi

df = pd.read_csv(file_name)
if len(df.index) == 0: # faster than if df.empty:
    df_is_empty = True
else:
    df_is_empty = False

if df_is_empty: 
    df["MFI"] = np.nan
else:
    df["MFI"] = calculate_mfi(high=df["High"].values, low=df["Low"].values, close=df["Close"].values, volume=df["Volume"].values, period=MFI_PERIOD)

[twopirllc]

Hello @zakcali

Thanks for submitting a faster version. It is now available on the development branch.

KJ

[carissssa]

@twopirllc any idea when this will be in main release? i cannot use dev branches and this is blocking pandas upgrade

[twopirllc]

Hello @carissssa,

A new main release will occur after outstanding PRs are completed and some other bugs and maintenance issues have been taken care of. When, however, is unknowable at this time. 😕

i cannot use dev branches and this is blocking pandas upgrade

What is preventing the use of the dev branch?

Kind Regards,
KJ