我们从Python开源项目中，提取了以下16个代码示例，用于说明如何使用scipy.signal.argrelextrema()。
def peaks(spectra,frequency,number=3,thresh=0.01): """ Return the peaks from the Fourier transform Variables: number: integer. number of peaks to print. thresh: float. Threshhold intensity for printing. Returns: Energy (eV), Intensity (depends on type of spectra) """ from scipy.signal import argrelextrema as pks # find all peak indices [idx], and remove those below thresh [jdx] idx = pks(np.abs(spectra),np.greater,order=3) jdx = np.where((np.abs(spectra[idx]) >= thresh)) kdx = idx[0][jdx[0]] # indices of peaks matching criteria if number > len(kdx): number = len(kdx) print("First "+str(number)+" peaks (eV) found: ") for i in xrange(number): print("{0:.4f}".format(frequency[kdx][i]*27.2114), "{0:.4f}".format(spectra[kdx][i]))
def get_local_maxima(x, y): """ This function ... :param x: :param y: :return: """ m = argrelextrema(y, np.greater)[0].tolist() # Find the index of the absolute maximum (should also be included, is not for example when it is at the edge) index = np.argmax(y) if index not in m: m.append(index) x_maxima = [x[i] for i in m] y_maxima = [y[i] for i in m] return x_maxima, y_maxima # -----------------------------------------------------------------
def get_local_minima(x, y): """ This function ... :param x: :param y: :return: """ m = argrelextrema(y, np.less)[0].tolist() # Find the indx of the absolute minimum (should also be included, is not for example when it is at the edge) index = np.argmin(y) if index not in m: m.append(index) x_minima = [x[i] for i in m] y_minima = [y[i] for i in m] return x_minima, y_minima # -----------------------------------------------------------------
def detect_peaks(hist, count=2): hist_copy = hist peaks = len(argrelextrema(hist_copy, np.greater, mode="wrap")[0]) sigma = log1p(peaks) print(peaks, sigma) while (peaks > count): new_hist = gaussian_filter(hist_copy, sigma=sigma) peaks = len(argrelextrema(new_hist, np.greater, mode="wrap")[0]) if peaks < count: peaks = count + 1 sigma = sigma * 0.5 continue hist_copy = new_hist sigma = log1p(peaks) print(peaks, sigma) return argrelextrema(hist_copy, np.greater, mode="wrap")[0]
def plot_kde(data): bw = 1.06 * st.stdev(data) / (len(data) ** .2) kde = KernelDensity(kernel='gaussian', bandwidth=bw).fit( np.array(data).reshape(-1, 1)) s = np.linspace(0, 1) e = kde.score_samples(s.reshape(-1, 1)) plt.plot(s, e) mi, ma = argrelextrema(e, np.less)[0], argrelextrema(e, np.greater)[0] logger.info("Minima: %s" % s[mi]) logger.info("Maxima: %s" % s[ma]) plt.plot(s[:mi[0] + 1], e[:mi[0] + 1], 'r', s[mi[0]:mi[1] + 1], e[mi[0]:mi[1] + 1], 'g', s[mi[1]:], e[mi[1]:], 'b', s[ma], e[ma], 'go', s[mi], e[mi], 'ro') plt.xlabel('Probability')
def find_hist_peaks(img, drop_first_max=False): '''Finds the first two interior peaks in the histogram; usually a good binary threshold lies halfway between these two peaks.''' import scipy.signal as sig array = sitk.GetArrayFromImage(img) hist = np.histogram(array, 16) hist_counts = np.array(hist[0]) hist_centers = np.array(hist[1]) peaks = sig.argrelextrema(hist_counts, np.greater)[0] if hist_counts[peaks[0]] < (0.01 * (hist_counts[peaks].mean())): # Don't allow a "noise" peak at the beginning. peaks = peaks[1:] if len(peaks) < 2: # Only one peak is found if the other peak is really at the edge. peaks = np.array([0, peaks[0]]) elif hist_counts[0] > hist_counts[peaks[0]]: if not drop_first_max: peaks = np.array([0, peaks[0]]) elif drop_first_max: peaks = np.array([peaks[1], peaks[2]]) i_L = hist_centers[peaks[0]] i_FM = hist_centers[peaks[1]] return i_L, i_FM
def spikes(data, start_sample, threshs, pad_len, order): for col_i in range(data.shape[1]): column = data[:, col_i] rel_extremes, = argrelextrema( column, lambda x, y: thres_extrema(x, y, threshs[col_i]), order=order) mask = (rel_extremes >= pad_len) & ( rel_extremes < data.shape[0] - pad_len) extreme_samples = rel_extremes[mask] + start_sample - pad_len yield from [(col_i, sample) for sample in extreme_samples]
def test_thres_extrema(): rs, = argrelextrema(data1[:, 0], lambda x, y: thres_extrema(x, y, 0.1)) assert len(rs) == 1 assert rs == 1
def _significant_periods(periods, pgram): # Order ascending periods = periods[::-1] pgram = pgram[::-1] # Scale and extract significant pgram = (pgram - pgram.min()) / pgram.ptp() significant = argrelextrema(pgram, np.greater, order=5) return periods[significant], pgram[significant]
def _significant_acf(corr, has_confint): if has_confint: corr, confint = corr periods = argrelextrema(np.abs(corr), np.greater, order=3)[0] corr = corr[periods] if has_confint: confint = confint[periods] result = np.column_stack((periods, corr)) if has_confint: result = (result, np.column_stack((periods, confint))) return result
def find_max_min(prices): prices_ = prices.copy() prices_.index = linspace(1., len(prices_), len(prices_)) #kr = KernelReg([prices_.values], [prices_.index.values], var_type='c', bw=[1.8, 1]) kr = KernelReg([prices_.values], [prices_.index.values], var_type='c', bw=[2]) # ????????????? ? # Either a user-specified bandwidth or the method for bandwidth selection. # If a string, valid values are ‘cv_ls’ (least-squares cross-validation) and ‘aic’ (AIC Hurvich bandwidth estimation). # Default is ‘cv_ls’. f = kr.fit([prices_.index.values]) smooth_prices = pd.Series(data=f[0], index=prices.index) local_max = argrelextrema(smooth_prices.values, np.greater)[0] local_min = argrelextrema(smooth_prices.values, np.less)[0] price_local_max_dt = [] for i in local_max: if (i > 1) and (i < len(prices) - 1): price_local_max_dt.append(prices.iloc[i - 2:i + 2].argmax()) price_local_min_dt = [] for i in local_min: if (i > 1) and (i < len(prices) - 1): price_local_min_dt.append(prices.iloc[i - 2:i + 2].argmin()) prices.name = 'price' maxima = pd.DataFrame(prices.loc[price_local_max_dt]) minima = pd.DataFrame(prices.loc[price_local_min_dt]) max_min = pd.concat([maxima, minima]).sort_index() max_min.index.name = 'date' max_min = max_min.reset_index() max_min = max_min[~max_min.date.duplicated()] p = prices.reset_index() max_min['day_num'] = p[p['index'].isin(max_min.date)].index.values max_min = max_min.set_index('day_num').price return max_min
def find_pred_onsets(scores, window): if window.shape[0] > 0: onset_function = np.convolve(scores, window, mode='same') else: onset_function = scores # see page 592 of "Universal onset detection with bidirectional long short-term memory neural networks" maxima = argrelextrema(onset_function, np.greater_equal, order=1)[0] return set(list(maxima))
def idxsOfRelativeExtrema(x, maxima=True, allowEq=False, axis=0): """ >>> idxsOfRelativeExtrema([2,1,5]) array([0, 2]) >>> idxsOfRelativeExtrema([2,1]) array([0]) >>> idxsOfRelativeExtrema([1,2]) array([1]) >>> idxsOfRelativeExtrema([2,1,5], maxima=False) array([1]) >>> idxsOfRelativeExtrema([1,1,1], allowEq=False) array([], dtype=int64) >>> idxsOfRelativeExtrema([0,1,1,1,0], allowEq=False) array([], dtype=int64) >>> idxsOfRelativeExtrema([1,1,1], allowEq=True) array([0, 1, 2]) """ if len(x) == 0: return np.empty(1) # [] if len(x) == 1: return np.zeros(1) # [0] x = np.asarray(x) pad = -np.inf if maxima else np.inf if maxima: if allowEq: func = np.greater_equal else: func = np.greater else: if allowEq: func = np.less_equal else: func = np.less if len(x.shape) == 1: x = np.r_[x, pad] # combine with wrap to check endpoints return signal.argrelextrema(x, func, mode='wrap')[0] elif axis == 0: pad = np.zeros((1, x.shape[1])) + pad x = np.vstack((x, pad)) return signal.argrelextrema(x, func, mode='wrap', axis=axis) elif axis == 1: pad = np.zeros((x.shape[0], 1)) + pad x = np.hstack((x, pad)) return signal.argrelextrema(x, func, mode='wrap', axis=axis) else: raise NotImplementedError("only supports axis={0, 1}!")
