金融指标技术分析库：  

https://github.com/twopirllc/pandas-ta

zigzag的实现：

"""
reference:
https://github.com/jbn/ZigZag.git
"""
import numpy as np

PEAK = 1
VALLEY = -1


def identify_initial_pivot(X, up_thresh, down_thresh):
    x_0 = X[0]
    x_t = x_0

    max_x = x_0
    min_x = x_0

    max_t = 0
    min_t = 0

    up_thresh += 1
    down_thresh += 1

    for t in range(1, len(X)):
        x_t = X[t]

        if x_t / min_x >= up_thresh:
            return VALLEY if min_t == 0 else PEAK

        if x_t / max_x <= down_thresh:
            return PEAK if max_t == 0 else VALLEY

        if x_t > max_x:
            max_x = x_t
            max_t = t

        if x_t < min_x:
            min_x = x_t
            min_t = t

    t_n = len(X)-1
    return VALLEY if x_0 < X[t_n] else PEAK


def peak_valley_pivots(X, up_thresh, down_thresh):
    """
    Find the peaks and valleys of a series.

    :param X: the series to analyze
    :param up_thresh: minimum relative change necessary to define a peak
    :param down_thesh: minimum relative change necessary to define a valley
    :return: an array with 0 indicating no pivot and -1 and 1 indicating
        valley and peak


    The First and Last Elements
    ---------------------------
    The first and last elements are guaranteed to be annotated as peak or
    valley even if the segments formed do not have the necessary relative
    changes. This is a tradeoff between technical correctness and the
    propensity to make mistakes in data analysis. The possible mistake is
    ignoring data outside the fully realized segments, which may bias
    analysis.
    """
    if down_thresh > 0:
        raise ValueError('The down_thresh must be negative.')

    initial_pivot = identify_initial_pivot(X, up_thresh, down_thresh)
    t_n = len(X)
    pivots = np.zeros(t_n, dtype=np.int_)
    trend = -initial_pivot
    last_pivot_t = 0
    last_pivot_x = X[0]

    pivots[0] = initial_pivot

    # Adding one to the relative change thresholds saves operations. Instead
    # of computing relative change at each point as x_j / x_i - 1, it is
    # computed as x_j / x_1. Then, this value is compared to the threshold + 1.
    # This saves (t_n - 1) subtractions.
    up_thresh += 1
    down_thresh += 1

    for t in range(1, t_n):
        x = X[t]
        r = x / last_pivot_x

        if trend == -1:
            if r >= up_thresh:
                pivots[last_pivot_t] = trend
                trend = PEAK
                last_pivot_x = x
                last_pivot_t = t
            elif x < last_pivot_x:
                last_pivot_x = x
                last_pivot_t = t
        else:
            if r <= down_thresh:
                pivots[last_pivot_t] = trend
                trend = VALLEY
                last_pivot_x = x
                last_pivot_t = t
            elif x > last_pivot_x:
                last_pivot_x = x
                last_pivot_t = t

    if last_pivot_t == t_n-1:
        pivots[last_pivot_t] = trend
    elif pivots[t_n-1] == 0:
        pivots[t_n-1] = -trend

    return pivots


def max_drawdown(X):
    """
    Compute the maximum drawdown of some sequence.

    :return: 0 if the sequence is strictly increasing.
        otherwise the abs value of the maximum drawdown
        of sequence X
    """
    mdd = 0
    peak = X[0]

    for x in X:
        if x > peak:
            peak = x

        dd = (peak - x) / peak

        if dd > mdd:
            mdd = dd

    return mdd if mdd != 0.0 else 0.0


def pivots_to_modes(pivots):
    """
    Translate pivots into trend modes.

    :param pivots: the result of calling ``peak_valley_pivots``
    :return: numpy array of trend modes. That is, between (VALLEY, PEAK] it
    is 1 and between (PEAK, VALLEY] it is -1.
    """

    modes = np.zeros(len(pivots), dtype=np.int_)
    mode = -pivots[0]

    modes[0] = pivots[0]

    for t in range(1, len(pivots)):
        x = pivots[t]
        if x != 0:
            modes[t] = mode
            mode = -x
        else:
            modes[t] = mode

    return modes


def compute_segment_returns(X, pivots):
    """
    :return: numpy array of the pivot-to-pivot returns for each segment."""
    pivot_points = X[pivots != 0]
    return pivot_points[1:] / pivot_points[:-1] - 1.0

使用示例：

import matplotlib
matplotlib.use("TkAgg")
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import sys
import pathlib
sys.path.append("%s/zigzag" % pathlib.Path().absolute())
from zigzag import zigzag


def plot_pivots(X, pivots):
    plt.xlim(0, len(X))
    plt.ylim(X.min()*0.99, X.max()*1.01)
    plt.plot(np.arange(len(X)), X, 'k:', alpha=0.5)
    plt.plot(np.arange(len(X))[pivots != 0], X[pivots != 0], 'k-')
    plt.scatter(np.arange(len(X))[pivots == 1], X[pivots == 1], color='g')
    plt.scatter(np.arange(len(X))[pivots == -1], X[pivots == -1], color='r')


np.random.seed(1997)
X = np.cumprod(1 + np.random.randn(100) * 0.01)
pivots = zigzag.peak_valley_pivots(X, 0.03, -0.03)

plot_pivots(X, pivots)
plt.show()

modes = zigzag.pivots_to_modes(pivots)
print(pd.Series(X).pct_change().groupby(modes).describe().unstack())
print(zigzag.compute_segment_returns(X, pivots))

pandas 的数据输入示例：

from pandas_datareader import get_data_yahoo

X = get_data_yahoo('GOOG')['Adj Close']
pivots = peak_valley_pivots(X.values, 0.2, -0.2)
ts_pivots = pd.Series(X, index=X.index)
ts_pivots = ts_pivots[pivots != 0]
X.plot()
ts_pivots.plot(style='g-o');

———————————————————————————————————————————————————————————-

微信公众号:  共鸣圈

欢迎讨论，邮件:  924948$qq.com       请把$改成@

QQ群：263132197
QQ：    924948

良辰美景补天漏，风雨雷电洗地尘

———————————————————————————————————————————————————————————-