#779 #625 #775 BUG ENH pvi update with numba MAINT use to_numpy()

pandas_ta/candles/cdl_inside.py

@@ -62,7 +62,7 @@ def cdl_inside(
     offset = v_offset(offset)
 
     # Calculate
-    np_high, np_low = high.values, low.values
+    np_high, np_low = high.to_numpy(), low.to_numpy()
     np_inside = np_cdl_inside(np_high, np_low)
     inside = Series(np_inside, index=close.index, dtype=bool)
 

pandas_ta/candles/ha.py

@@ -65,8 +65,8 @@ def ha(
         return
 
     # Calculate
-    np_open, np_high = open_.values, high.values
-    np_low, np_close = low.values, close.values
+    np_open, np_high = open_.to_numpy(), high.to_numpy()
+    np_low, np_close = low.to_numpy(), close.to_numpy()
     ha_open, ha_high, ha_low, ha_close = np_ha(np_open, np_high, np_low, np_close)
     df = DataFrame({
         "HA_open": ha_open,

pandas_ta/core.py

@@ -1806,10 +1806,10 @@ def obv(self, offset=None, **kwargs: DictLike):
         result = obv(close=close, volume=volume, offset=offset, **kwargs)
         return self._post_process(result, **kwargs)
 
-    def pvi(self, length=None, initial=None, mamode=None, offset=None, **kwargs: DictLike):
+    def pvi(self, length=None, initial=None, mamode=None, overlay=None, offset=None, **kwargs: DictLike):
         close = self._get_column(kwargs.pop("close", "close"))
         volume = self._get_column(kwargs.pop("volume", "volume"))
-        result = pvi(close=close, volume=volume, length=length, initial=initial, mamode=mamode, offset=offset, **kwargs)
+        result = pvi(close=close, volume=volume, length=length, initial=initial, mamode=mamode, overlay=overlay, offset=offset, **kwargs)
         return self._post_process(result, **kwargs)
 
     def pvo(self, fast=None, slow=None, signal=None, scalar=None, offset=None, **kwargs: DictLike):

pandas_ta/cycles/reflex.py

@@ -93,7 +93,7 @@ def reflex(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     result = np_reflex(np_close, length, smooth, alpha, pi, sqrt2)
     result[:length] = nan
     result = Series(result, index=close.index)

pandas_ta/momentum/crsi.py

@@ -145,7 +145,7 @@ def crsi(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     streak = Series(consecutive_streak(np_close), index=close.index)
 
     if Imports["talib"] and mode_tal:

pandas_ta/momentum/rvgi.py

@@ -63,7 +63,7 @@ def rvgi(
     rvgi = numerator / denominator
     signal = swma(rvgi, length=swma_length)
 
-    if all(isnan(signal.values)):
+    if all(isnan(signal.to_numpy())):
         return  # Emergency Break
 
     # Offset

pandas_ta/overlap/alma.py

@@ -55,7 +55,7 @@ def alma(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     x = arange(length)
     k = floor(offset_ * (length - 1))
     weights = exp(-0.5 * ((sigma / length) * (x - k)) ** 2)

pandas_ta/overlap/dema.py

@@ -52,7 +52,7 @@ def dema(
         ema2 = ema(close=ema1, length=length, talib=mode_tal)
         dema = 2 * ema1 - ema2
 
-    if all(isnan(dema.values)):
+    if all(isnan(dema.to_numpy())):
         return  # Emergency Break
 
     # Offset

pandas_ta/overlap/hl2.py

@@ -35,7 +35,7 @@ def hl2(
         return
 
     # Calculate
-    avg = 0.5 * (high.values + low.values)
+    avg = 0.5 * (high.to_numpy() + low.to_numpy())
     hl2 = Series(avg, index=high.index)
 
     # Offset

pandas_ta/overlap/hlc3.py

@@ -43,7 +43,7 @@ def hlc3(
         from talib import TYPPRICE
         hlc3 = TYPPRICE(high, low, close)
     else:
-        avg = (high.values + low.values + close.values) / 3.0
+        avg = (high.to_numpy() + low.to_numpy() + close.to_numpy()) / 3.0
         hlc3 = Series(avg, index=close.index)
 
     # Offset

pandas_ta/overlap/linreg.py

@@ -70,7 +70,7 @@ def linreg(
     tsf = kwargs.pop("tsf", False)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
 
     if Imports["talib"] and mode_tal and not r:
         from talib import LINEARREG, LINEARREG_ANGLE, LINEARREG_INTERCEPT, LINEARREG_SLOPE, TSF

pandas_ta/overlap/mama.py

@@ -143,7 +143,7 @@ def mama(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     if Imports["talib"] and mode_tal:
         from talib import MAMA
         mama, fama = MAMA(np_close, fastlimit, slowlimit)

pandas_ta/overlap/ohlc4.py

@@ -36,7 +36,7 @@ def ohlc4(
     offset = v_offset(offset)
 
     # Calculate
-    avg = 0.25 * (open_.values + high.values + low.values + close.values)
+    avg = 0.25 * (open_.to_numpy() + high.to_numpy() + low.to_numpy() + close.to_numpy())
     ohlc4 = Series(avg, index=close.index)
 
     # Offset

pandas_ta/overlap/pivots.py

@@ -190,10 +190,10 @@ def pivots(
             index=dt_index
         )
 
-    np_open = df.open.values
-    np_high = df.high.values
-    np_low = df.low.values
-    np_close = df.close.values
+    np_open = df.open.to_numpy()
+    np_high = df.high.to_numpy()
+    np_low = df.low.to_numpy()
+    np_close = df.close.to_numpy()
 
     # Create nan arrays for "demark" and "fibonacci" pivots
     _nan_array = zeros_like(np_close)

pandas_ta/overlap/sma.py

@@ -80,7 +80,7 @@ def sma(
         from talib import SMA
         sma = SMA(close, length)
     else:
-        np_close = close.values
+        np_close = close.to_numpy()
         sma = np_sma(np_close, length)
         sma = Series(sma, index=close.index)
 

pandas_ta/overlap/ssf.py

@@ -95,7 +95,7 @@ def ssf(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     if everget:
         ssf = np_ssf_everget(np_close, length, pi, sqrt2)
     else:

pandas_ta/overlap/ssf3.py

@@ -77,7 +77,7 @@ def ssf3(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     ssf = np_ssf3(np_close, length, pi, sqrt3)
     ssf = Series(ssf, index=close.index)
 

pandas_ta/overlap/wcp.py

@@ -49,7 +49,7 @@ def wcp(
         from talib import WCLPRICE
         wcp = WCLPRICE(high, low, close)
     else:
-        weight = high.values + low.values + 2 * close.values
+        weight = high.to_numpy() + low.to_numpy() + 2 * close.to_numpy()
         wcp = Series(weight, index=close.index)
 
     # Offset

pandas_ta/overlap/wma.py

@@ -76,7 +76,7 @@ def wma(
         from talib import WMA
         wma = WMA(close, length)
     else:
-        np_close = close.values
+        np_close = close.to_numpy()
         wma_ = np_wma(np_close, length, asc, True)
         wma = Series(wma_, index=close.index)
 

pandas_ta/performance/log_return.py

@@ -42,7 +42,7 @@ def log_return(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     if cumulative:
         r = np_close / np_close[0]
     else:

pandas_ta/performance/percent_return.py

@@ -42,7 +42,7 @@ def percent_return(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     if cumulative:
         pr = (np_close / np_close[0]) - 1
     else:

pandas_ta/transform/ifisher.py

@@ -55,14 +55,14 @@ def ifisher(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     is_remapped = logical_and(np_close >= -1, np_close <= 1)
     if not all(is_remapped):
         np_max, np_min = max(np_close), min(np_close)
         close_map = remap(close, from_min=np_min, from_max=np_max, to_min=-1, to_max=1)
-        if close_map is None or all(isnan(close_map.values)):
+        if close_map is None or all(isnan(close_map.to_numpy())):
             return  # Emergency Break
-        np_close = close_map.values
+        np_close = close_map.to_numpy()
     amped = exp(amp * np_close)
     result = (amped - 1) / (amped + 1)
 

pandas_ta/transform/remap.py

@@ -50,7 +50,7 @@ def remap(
     frange, trange = from_max - from_min, to_max - to_min
     if frange <= 0 or trange <= 0:
         return
-    result = to_min + (trange / frange) * (close.values - from_min)
+    result = to_min + (trange / frange) * (close.to_numpy() - from_min)
     result = Series(result, index=close.index)
 
     # Offset

pandas_ta/trend/alphatrend.py

@@ -126,9 +126,9 @@ def alphatrend(
     if momo is None:
         return
 
-    np_upper_atr, np_lower_atr = upper_atr.values, lower_atr.values
+    np_upper_atr, np_lower_atr = upper_atr.to_numpy(), lower_atr.to_numpy()
 
-    at = np_alpha(np_lower_atr, np_upper_atr, momo.values >= threshold)
+    at = np_alpha(np_lower_atr, np_upper_atr, momo.to_numpy() >= threshold)
     at = Series(at, index=close.index)
 
     atl = at.shift(lag)

pandas_ta/trend/decay.py

@@ -71,7 +71,7 @@ def decay(
     offset = v_offset(offset)
 
     # Calculate
-    _mode, np_close = "L", close.values
+    _mode, np_close = "L", close.to_numpy()
 
     if mode in ["exp", "exponential"]:
         _mode = "EXP"

pandas_ta/trend/ht_trendline.py

@@ -123,7 +123,7 @@ def ht_trendline(
         from talib import HT_TRENDLINE
         tl = HT_TRENDLINE(close)
     else:
-        np_close = close.values
+        np_close = close.to_numpy()
         np_tl = np_ht_trendline(np_close)
 
         if prenan > 0:

pandas_ta/trend/trendflex.py

@@ -92,7 +92,7 @@ def trendflex(
     offset = v_offset(offset)
 
     # Calculate
-    np_close = close.values
+    np_close = close.to_numpy()
     result = np_trendflex(np_close, length, smooth, alpha, pi, sqrt2)
     result[:length] = nan
     result = Series(result, index=close.index)

pandas_ta/trend/xsignals.py

@@ -93,7 +93,7 @@ def xsignals(
 
     # Modify trades to fill gaps for trends
     trades.replace({0: nan}, inplace=True)
-    trades.interpolate(method="pad", inplace=True)
+    trades.ffill(limit_area="inside", inplace=True) # or trades.bfill(limit_area="inside", inplace=True)
     trades.fillna(0, inplace=True)
 
     trends = (trades > 0).astype(int)

pandas_ta/trend/zigzag.py

@@ -60,7 +60,7 @@ def zigzag(
 
     if close is not None:
         close = v_series(close, _length + 1)
-        np_close = close.values
+        np_close = close.to_numpy()
         if close is None:
             return
 
@@ -70,7 +70,7 @@ def zigzag(
     offset = v_offset(offset)
 
     # Calculation
-    np_high, np_low = high.values, low.values
+    np_high, np_low = high.to_numpy(), low.to_numpy()
     highest_high = high.rolling(window=pivot_leg, center=True, min_periods=0).max()
     lowest_low = low.rolling(window=pivot_leg, center=True, min_periods=0).min()
 

pandas_ta/utils/_math.py

@@ -104,7 +104,7 @@ def geometric_mean(series: Series) -> Float:
     if n < 1:
         return series.iloc[0]
 
-    has_zeros = 0 in series.values
+    has_zeros = 0 in series.to_numpy()
     if has_zeros:
         series = series.fillna(0) + 1
     if all(series > 0):

pandas_ta/volatility/atrts.py

@@ -122,7 +122,7 @@ def atrts(
     atr_ *= multiplier
     ma_ = _ma(mamode, close, length=ma_length, talib=mode_tal)
 
-    np_close, np_ma, np_atr = close.values, ma_.values, atr_.values
+    np_close, np_ma, np_atr = close.to_numpy(), ma_.to_numpy(), atr_.to_numpy()
     np_atrts_, _, _ = np_atrts(np_close, np_ma, np_atr, length, ma_length)
 
     percent = kwargs.pop("percent", False)

pandas_ta/volume/kvo.py

@@ -69,11 +69,11 @@ def kvo(
     sv = signed_volume.loc[signed_volume.first_valid_index():, ]
 
     kvo = ma(mamode, sv, length=fast) - ma(mamode, sv, length=slow)
-    if kvo is None or all(isnan(kvo.values)):
+    if kvo is None or all(isnan(kvo.to_numpy())):
         return  # Emergency Break
 
     kvo_signal = ma(mamode, kvo.loc[kvo.first_valid_index():, ], length=signal)
-    if kvo_signal is None or all(isnan(kvo_signal.values)):
+    if kvo_signal is None or all(isnan(kvo_signal.to_numpy())):
         return  # Emergency Break
 
     # Offset

pandas_ta/volume/mfi.py

@@ -68,8 +68,8 @@ def mfi(
     else:
         m, _ones = close.size, ones(length)
 
-        tp = (high.values + low.values + close.values) / 3.0
-        smf = tp * volume.values * where(tp > roll(tp, shift=drift), 1, -1)
+        tp = (high.to_numpy() + low.to_numpy() + close.to_numpy()) / 3.0
+        smf = tp * volume.to_numpy() * where(tp > roll(tp, shift=drift), 1, -1)
 
         pos, neg = maximum(smf, 0), maximum(-smf, 0)
         avg_gain, avg_loss = convolve(pos, _ones)[:m], convolve(neg, _ones)[:m]