Suggestion: ADX different values for neg and pos

[Brankhos]

Can you add new _window for neg and pos? Default will fetch from default adx but if you typed a number(neg, pos or both) it will fetch it. And can you do this for other functions as well? Its hard to add it for me

[Brankhos]

wrong//

class ADXIndicator2(IndicatorMixin2):

def __init__(
    self,
    high: pd.Series,
    low: pd.Series,
    close: pd.Series,
    window: int = 14,
    fillna: bool = False,
    window_p: int = 14,
    window_n: int = 14,
):
    self._high = high
    self._low = low
    self._close = close
    self._window = window
    self._fillna = fillna
    self._window_p = window_p
    self._window_n = window_n

    self._run()

def _run(self):
    if self._window == 0:
        raise ValueError("window may not be 0")

    close_shift = self._close.shift(1)
    pdm = _get_min_max(self._high, close_shift, "max")
    pdn = _get_min_max(self._low, close_shift, "min")
    diff_directional_movement = pdm - pdn

    self._trs_initial = np.zeros(self._window - 1)
    self._trs = np.zeros(len(self._close) - (self._window - 1))
    self._trs[0] = diff_directional_movement.dropna()[0 : self._window].sum()
    diff_directional_movement = diff_directional_movement.reset_index(drop=True)

    for i in range(1, len(self._trs) - 1):
        self._trs[i] = (
            self._trs[i - 1]
            - (self._trs[i - 1] / float(self._window))
            + diff_directional_movement[self._window + i]
        )

    diff_up = self._high - self._high.shift(1)
    diff_down = self._low.shift(1) - self._low
    pos = abs(((diff_up > diff_down) & (diff_up > 0)) * diff_up)
    neg = abs(((diff_down > diff_up) & (diff_down > 0)) * diff_down)

    self._dip = np.zeros(len(self._close) - (self._window_p - 1))
    self._dip[0] = pos.dropna()[0 : self._window_p].sum()

    pos = pos.reset_index(drop=True)

    for i in range(1, len(self._dip) - 1):
        self._dip[i] = (
            self._dip[i - 1]
            - (self._dip[i - 1] / float(self._window_p))
            + pos[self._window_p + i]
        )

    self._din = np.zeros(len(self._close) - (self._window_n - 1))
    self._din[0] = neg.dropna()[0 : self._window_n].sum()

    neg = neg.reset_index(drop=True)

    for i in range(1, len(self._din) - 1):
        self._din[i] = (
            self._din[i - 1]
            - (self._din[i - 1] / float(self._window_n))
            + neg[self._window_n + i]
        )

def adx(self) -> pd.Series:
    """Average Directional Index (ADX)
    Returns:
        pandas.Series: New feature generated.tr
    """
    dip = np.zeros(len(self._trs))
    for i in range(len(self._trs)):
        dip[i] = 100 * (self._dip[i] / self._trs[i])

    din = np.zeros(len(self._trs))
    for i in range(len(self._trs)):
        din[i] = 100 * (self._din[i] / self._trs[i])

    directional_index = 100 * np.abs((dip - din) / (dip + din))

    adx_series = np.zeros(len(self._trs))
    adx_series[self._window] = directional_index[0 : self._window].mean()

    for i in range(self._window + 1, len(adx_series)):
        adx_series[i] = (
            (adx_series[i - 1] * (self._window - 1)) + directional_index[i - 1]
        ) / float(self._window)

    adx_series = np.concatenate((self._trs_initial, adx_series), axis=0)
    adx_series = pd.Series(data=adx_series, index=self._close.index)

    adx_series = self._check_fillna(adx_series, value=20)
    return pd.Series(adx_series, name="adx")

def adx_pos(self) -> pd.Series:
    """Plus Directional Indicator (+DI)
    Returns:
        pandas.Series: New feature generated.
    """
    dip = np.zeros(len(self._close))
    for i in range(1, len(self._trs) - 1):
        dip[i + self._window_p] = 100 * (self._dip[i] / self._trs[i])

    adx_pos_series = self._check_fillna(
        pd.Series(dip, index=self._close.index), value=20
    )
    return pd.Series(adx_pos_series, name="adx_pos")

def adx_neg(self) -> pd.Series:
    """Minus Directional Indicator (-DI)
    Returns:
        pandas.Series: New feature generated.
    """
    din = np.zeros(len(self._close))
    for i in range(1, len(self._trs) - 1):
        din[i + self._window_n] = 100 * (self._din[i] / self._trs[i])

    adx_neg_series = self._check_fillna(
        pd.Series(din, index=self._close.index), value=20
    )
    return pd.Series(adx_neg_series, name="adx_neg")

[sasan-saqaeeyan]

hello, i got different ADX value using ta library from meta trader5 adx indicator ? help me to solve my problem?