Python numpy.fabs() Examples

def find_right_intersect(vec, target_val, start_index=0):
        nearest_index = start_index
        next_index = start_index

        size = len(vec) - 1
        if next_index == size:
            return size

        next_val = vec[next_index]
        best_distance = np.abs(next_val - target_val)
        while (next_index < size):
            next_index += 1
            next_val = vec[next_index]
            dist = np.fabs(next_val - target_val)  # pylint: disable=assignment-from-no-return
            if dist < best_distance:
                best_distance = dist
                nearest_index = next_index
            if next_index == size or next_val < target_val:
                break
        return nearest_index 

def absdiff(self, constant_value, separate_re_im=False):
        """
        Returns a ReportableQty that is the (element-wise in the vector case)
        difference between `constant_value` and this one given by:

        `abs(self - constant_value)`.
        """
        if separate_re_im:
            re_v = _np.fabs(_np.real(self.value) - _np.real(constant_value))
            im_v = _np.fabs(_np.imag(self.value) - _np.imag(constant_value))
            if self.has_eb():
                return (ReportableQty(re_v, _np.fabs(_np.real(self.errbar)), self.nonMarkovianEBs),
                        ReportableQty(im_v, _np.fabs(_np.imag(self.errbar)), self.nonMarkovianEBs))
            else:
                return ReportableQty(re_v), ReportableQty(im_v)

        else:
            v = _np.absolute(self.value - constant_value)
            if self.has_eb():
                return ReportableQty(v, _np.absolute(self.errbar), self.nonMarkovianEBs)
            else:
                return ReportableQty(v) 

def test_irreg_hf(self):
        idx = date_range('2012-6-22 21:59:51', freq='S', periods=100)
        df = DataFrame(np.random.randn(len(idx), 2), idx)

        irreg = df.iloc[[0, 1, 3, 4]]
        _, ax = self.plt.subplots()
        irreg.plot(ax=ax)
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()

        sec = 1. / 24 / 60 / 60
        assert (np.fabs(diffs[1:] - [sec, sec * 2, sec]) < 1e-8).all()

        _, ax = self.plt.subplots()
        df2 = df.copy()
        df2.index = df.index.astype(object)
        df2.plot(ax=ax)
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()
        assert (np.fabs(diffs[1:] - sec) < 1e-8).all() 

def rotipp(acceleration_x, time_step_x, acceleration_y, time_step_y, periods,
        percentile, damping=0.05, units="cm/s/s", method="Nigam-Jennings"):
    """
    Returns the rotationally independent spectrum RotIpp as defined by
    Boore (2010)
    """
    if np.fabs(time_step_x - time_step_y) > 1E-10:
        raise ValueError("Record pair must have the same time-step!")
    acceleration_x, acceleration_y = equalise_series(acceleration_x,
                                                     acceleration_y)
    target, rota, rotv, rotd, angles = rotdpp(acceleration_x, time_step_x,
                                              acceleration_y, time_step_y,
                                              periods, percentile, damping,
                                              units, method)
    locn, penalty = _get_gmrotd_penalty(
        np.hstack([target["PGA"],target["Pseudo-Acceleration"]]),
        rota)
    target_theta = np.radians(angles[locn])
    arotpp = acceleration_x * np.cos(target_theta) +\
        acceleration_y * np.sin(target_theta)
    spec = get_response_spectrum(arotpp, time_step_x, periods, damping, units,
        method)[0]
    spec["GMRot{:2.0f}".format(percentile)] = target
    return spec 

def nextPos(self, x, x_b, r1, r2, r3, r4, task):
		r"""Move individual to new position in search space.

		Args:
			x (numpy.ndarray): Individual represented with components.
			x_b (nmppy.ndarray): Best individual represented with components.
			r1 (float): Number dependent on algorithm iteration/generations.
			r2 (float): Random number in range of 0 and 2 * PI.
			r3 (float): Random number in range [Rmin, Rmax].
			r4 (float): Random number in range [0, 1].
			task (Task): Optimization task.

		Returns:
			numpy.ndarray: New individual that is moved based on individual ``x``.
		"""
		return task.repair(x + r1 * (sin(r2) if r4 < 0.5 else cos(r2)) * fabs(r3 * x_b - x), self.Rand) 

def rho(self, z):
        r"""
        The robust criterion function for Huber's t.

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        rho : array
            rho(z) = .5*z**2            for \|z\| <= t

            rho(z) = \|z\|*t - .5*t**2    for \|z\| > t
        """
        z = np.asarray(z)
        test = self._subset(z)
        return (test * 0.5 * z**2 +
                (1 - test) * (np.fabs(z) * self.t - 0.5 * self.t**2)) 

def rho(self, z):
        r"""
        The robust criterion function for Ramsay's Ea.

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        rho : array
            rho(z) = a**-2 * (1 - exp(-a*\|z\|)*(1 + a*\|z\|))
        """
        z = np.asarray(z)
        return (1 - np.exp(-self.a * np.fabs(z)) *
                (1 + self.a * np.fabs(z))) / self.a**2 

def psi(self, z):
        """
        The psi function for Ramsay's Ea estimator

        The analytic derivative of rho

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        psi : array
            psi(z) = z*exp(-a*\|z\|)
        """
        z = np.asarray(z)
        return z * np.exp(-self.a * np.fabs(z)) 

def weights(self, z):
        """
        Ramsay's Ea weighting function for the IRLS algorithm

        The psi function scaled by z

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        weights : array
            weights(z) = exp(-a*\|z\|)
        """

        z = np.asarray(z)
        return np.exp(-self.a * np.fabs(z)) 

def cont(self, ML=False, rtol=1.0e-05, params_rtol=1e-5, params_atol=1e-4):
        '''convergence check for iterative estimation

        '''

        self.dev, old = self.deviance(ML=ML), self.dev

        #self.history.append(np.hstack((self.dev, self.a)))
        self.history['llf'].append(self.dev)
        self.history['params'].append(self.a.copy())
        self.history['D'].append(self.D.copy())

        if np.fabs((self.dev - old) / self.dev) < rtol:   #why is there times `*`?
            #print np.fabs((self.dev - old)), self.dev, old
            self.termination = 'llf'
            return False

        #break if parameters converged
        #TODO: check termination conditions, OR or AND
        if np.all(np.abs(self.a - self._a_old) < (params_rtol * self.a + params_atol)):
            self.termination = 'params'
            return False

        self._a_old =  self.a.copy()
        return True 

def test_irreg_hf(self):
        idx = date_range('2012-6-22 21:59:51', freq='S', periods=100)
        df = DataFrame(np.random.randn(len(idx), 2), idx)

        irreg = df.iloc[[0, 1, 3, 4]]
        _, ax = self.plt.subplots()
        irreg.plot(ax=ax)
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()

        sec = 1. / 24 / 60 / 60
        assert (np.fabs(diffs[1:] - [sec, sec * 2, sec]) < 1e-8).all()

        _, ax = self.plt.subplots()
        df2 = df.copy()
        df2.index = df.index.astype(object)
        df2.plot(ax=ax)
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()
        assert (np.fabs(diffs[1:] - sec) < 1e-8).all() 

def _wrap_results(result, dtype):
    """ wrap our results if needed """

    if is_datetime64_dtype(dtype):
        if not isinstance(result, np.ndarray):
            result = tslib.Timestamp(result)
        else:
            result = result.view(dtype)
    elif is_timedelta64_dtype(dtype):
        if not isinstance(result, np.ndarray):

            # raise if we have a timedelta64[ns] which is too large
            if np.fabs(result) > _int64_max:
                raise ValueError("overflow in timedelta operation")

            result = tslib.Timedelta(result, unit='ns')
        else:
            result = result.astype('i8').view(dtype)

    return result 

def lidar_to_img(points, img_size):
    # pdb.set_trace()
    lidar_data = np.array(points[:, :2])
    lidar_data *= 9.9999
    lidar_data -= (0.5 * img_size, 0.5 * img_size)
    lidar_data = np.fabs(lidar_data)
    lidar_data = lidar_data.astype(np.int32)
    lidar_data = np.reshape(lidar_data, (-1, 2))
    lidar_img = np.zeros((img_size, img_size))
    lidar_img[tuple(lidar_data.T)] = 255
    return torch.tensor(lidar_img).cuda()


# def lidar_to_img(points, img_size):
#     # pdb.set_trace()
#     lidar_data = points[:, :2]
#     lidar_data *= 9.9999
#     lidar_data -= torch.tensor((0.5 * img_size, 0.5 * img_size)).cuda()
#     lidar_data = torch.abs(lidar_data)
#     lidar_data = torch.floor(lidar_data).long()
#     lidar_data = lidar_data.view(-1, 2)
#     lidar_img = torch.zeros((img_size, img_size)).cuda()
#     lidar_img[lidar_data.permute(1,0)] = 255
#     return lidar_img 

def lidar_to_heightmap(points, img_size):
    # pdb.set_trace()
    lidar_data = np.array(points[:, :2])
    height_data = np.array(points[:,2])
    height_data *= 255/2
    height_data[height_data < 0] = 0
    height_data[height_data > 255] = 255
    height_data = np.fabs(height_data)
    height_data = height_data.astype(np.int32)


    lidar_data *= 9.9999
    lidar_data -= (0.5 * img_size, 0.5 * img_size)
    lidar_data = np.fabs(lidar_data)
    lidar_data = lidar_data.astype(np.int32)
    lidar_data = np.reshape(lidar_data, (-1, 2))
    lidar_img = np.zeros((img_size, img_size))
    lidar_img[tuple(lidar_data.T)] = height_data # TODO: sort the point wrt height first lex sort
    return lidar_img 

def _max_error(arrays1, arrays2):
  """Computes maximum elementwise gap between two lists of ndarrays.

  Computes the maximum elementwise gap between two lists with the same length,
  of arrays with the same shape.

  Args:
    arrays1: a lists of np.ndarrays.
    arrays2: a lists of np.ndarrays of the same shape as arrays1.

  Returns:
    The maximum elementwise absolute difference between the two lists of arrays.
  """
  error = 0
  for array1, array2 in zip(arrays1, arrays2):
    if array1.size or array2.size:  # Handle zero size ndarrays correctly
      error = np.maximum(error, np.fabs(array1 - array2).max())
  return error 

def __call__(self, df_resid, nobs, resid):
        h = (df_resid)/nobs*(self.d**2 + (1-self.d**2)*\
                    Gaussian.cdf(self.d)-.5 - self.d/(np.sqrt(2*np.pi))*\
                    np.exp(-.5*self.d**2))
        s = mad(resid)
        subset = lambda x: np.less(np.fabs(resid/x),self.d)
        chi = lambda s: subset(s)*(resid/s)**2/2+(1-subset(s))*(self.d**2/2)
        scalehist = [np.inf,s]
        niter = 1
        while (np.abs(scalehist[niter-1] - scalehist[niter])>self.tol \
                and niter < self.maxiter):
            nscale = np.sqrt(1/(nobs*h)*np.sum(chi(scalehist[-1]))*\
                    scalehist[-1]**2)
            scalehist.append(nscale)
            niter += 1
            #if niter == self.maxiter:
            #    raise ValueError("Huber's scale failed to converge")
        return scalehist[-1] 

def weights(self, z):
        """
        Huber's t weighting function for the IRLS algorithm

        The psi function scaled by z

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        weights : array
            weights(z) = 1          for \|z\| <= t

            weights(z) = t/\|z\|      for \|z\| > t
        """
        z = np.asarray(z)
        test = self._subset(z)
        absz = np.fabs(z)
        absz[test] = 1.0
        return test + (1 - test) * self.t / absz 

def test_irreg_hf(self):
        import matplotlib.pyplot as plt
        fig = plt.gcf()
        plt.clf()
        fig.add_subplot(111)

        idx = date_range('2012-6-22 21:59:51', freq='S', periods=100)
        df = DataFrame(np.random.randn(len(idx), 2), idx)

        irreg = df.ix[[0, 1, 3, 4]]
        ax = irreg.plot()
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()

        sec = 1. / 24 / 60 / 60
        self.assert_((np.fabs(diffs[1:] - [sec, sec * 2, sec]) < 1e-8).all())

        plt.clf()
        fig.add_subplot(111)
        df2 = df.copy()
        df2.index = df.index.asobject
        ax = df2.plot()
        diffs = Series(ax.get_lines()[0].get_xydata()[:, 0]).diff()
        self.assert_((np.fabs(diffs[1:] - sec) < 1e-8).all()) 

def weights(self, z):
        """
        Hampel weighting function for the IRLS algorithm

        The psi function scaled by z

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        weights : array
            weights(z) = 1                            for \|z\| <= a

            weights(z) = a/\|z\|                        for a < \|z\| <= b

            weights(z) = a*(c - \|z\|)/(\|z\|*(c-b))      for b < \|z\| <= c

            weights(z) = 0                            for \|z\| > c

        """
        z = np.asarray(z)
        a = self.a; b = self.b; c = self.c
        t1, t2, t3 = self._subset(z)
        v = (t1 +
            t2 * a/np.fabs(z) +
            t3 * a*(c-np.fabs(z))/(np.fabs(z)*(c-b)))
        v[np.where(np.isnan(v))]=1. # for some reason 0 returns a nan?
        return v 

def mad(a, c=Gaussian.ppf(3/4.), axis=0, center=np.median):
    # c \approx .6745
    """
    The Median Absolute Deviation along given axis of an array

    Parameters
    ----------
    a : array-like
        Input array.
    c : float, optional
        The normalization constant.  Defined as scipy.stats.norm.ppf(3/4.),
        which is approximately .6745.
    axis : int, optional
        The defaul is 0. Can also be None.
    center : callable or float
        If a callable is provided, such as the default `np.median` then it
        is expected to be called center(a). The axis argument will be applied
        via np.apply_over_axes. Otherwise, provide a float.

    Returns
    -------
    mad : float
        `mad` = median(abs(`a` - center))/`c`
    """
    a = np.asarray(a)
    if callable(center):
        center = np.apply_over_axes(center, a, axis)
    return np.median((np.fabs(a-center))/c, axis=axis) 

def rho(self, z):
        r"""
        The robust criterion function for Hampel's estimator

        Parameters
        ----------
        z : array-like
            1d array

        Returns
        -------
        rho : array
            rho(z) = (1/2.)*z**2                    for \|z\| <= a

            rho(z) = a*\|z\| - 1/2.*a**2              for a < \|z\| <= b

            rho(z) = a*(c*\|z\|-(1/2.)*z**2)/(c-b)    for b < \|z\| <= c

            rho(z) = a*(b + c - a)                  for \|z\| > c
        """

        z = np.fabs(z)
        a = self.a; b = self.b; c = self.c
        t1, t2, t3 = self._subset(z)
        v = (t1 * z**2 * 0.5 +
             t2 * (a * z - a**2 * 0.5) +
             t3 * (a * (c * z - z**2 * 0.5) / (c - b) - 7 * a**2 / 6.) +
             (1 - t1 + t2 + t3) * a * (b + c - a))
        return v 

def _subset(self, z):
        """
        Hampel's function is defined piecewise over the range of z
        """
        z = np.fabs(np.asarray(z))
        t1 = np.less_equal(z, self.a)
        t2 = np.less_equal(z, self.b) * np.greater(z, self.a)
        t3 = np.less_equal(z, self.c) * np.greater(z, self.b)
        return t1, t2, t3 

def get_peak_measures(time_step, acceleration, get_vel=False, 
    get_disp=False):
    """
    Returns the peak measures from acceleration, velocity and displacement
    time-series
    :param float time_step:
        Time step of acceleration time series in s
    :param numpy.ndarray acceleration:
        Acceleration time series
    :param bool get_vel:
        Choose to return (and therefore calculate) velocity (True) or otherwise
        (false)
    :returns:
        * pga - Peak Ground Acceleration
        * pgv - Peak Ground Velocity
        * pgd - Peak Ground Displacement
        * velocity - Velocity Time Series
        * dispalcement - Displacement Time series
    """
    pga = np.max(np.fabs(acceleration))
    velocity = None
    displacement = None
    # If displacement is not required then do not integrate to get
    # displacement time series
    if get_disp:
        get_vel = True
    if get_vel:
        velocity = time_step * cumtrapz(acceleration, initial=0.)
        pgv = np.max(np.fabs(velocity))
    else:
        pgv = None
    if get_disp:
        displacement = time_step * cumtrapz(velocity, initial=0.)
        pgd = np.max(np.fabs(displacement))
    else:
        pgd = None
    return pga, pgv, pgd, velocity, displacement 

def cont(self):
        '''condition to continue iteration loop

        Parameters
        ----------
        tol

        Returns
        -------
        cont : bool
            If true, then iteration should be continued.

        '''
        self.iter += 1 #moved here to always count, not necessary
        if DEBUG:
            print(self.iter, self.results.Y.shape)
            print(self.results.predict(self.exog).shape, self.weights.shape)
        curdev = (((self.results.Y - self.results.predict(self.exog))**2) * self.weights).sum()

        if self.iter > self.maxiter: #kill it, no max iterationoption
            return False
        if np.fabs((self.dev - curdev) / curdev) < self.rtol:
            self.dev = curdev
            return False

        #self.iter += 1
        self.dev = curdev
        return True 

def _transformation_param_deceptive(y, A=0.35, B=0.001, C=0.05):
    tmp1 = np.floor(y - A + B) * (1.0 - C + (A - B) / B) / (A - B)
    tmp2 = np.floor(A + B - y) * (1.0 - C + (1.0 - A - B) / B) / (1.0 - A - B)
    ret = 1.0 + (np.fabs(y - A) - B) * (tmp1 + tmp2 + 1.0 / B)
    return correct_to_01(ret)


# ---------------------------------------------------------------------------------------------------------
# REDUCTION
# --------------------------------------------------------------------------------------------------------- 

def test_replace2(self):
        N = 100
        ser = pd.Series(np.fabs(np.random.randn(N)), tm.makeDateIndex(N),
                        dtype=object)
        ser[:5] = np.nan
        ser[6:10] = 'foo'
        ser[20:30] = 'bar'

        # replace list with a single value
        rs = ser.replace([np.nan, 'foo', 'bar'], -1)

        assert (rs[:5] == -1).all()
        assert (rs[6:10] == -1).all()
        assert (rs[20:30] == -1).all()
        assert (pd.isna(ser[:5])).all()

        # replace with different values
        rs = ser.replace({np.nan: -1, 'foo': -2, 'bar': -3})

        assert (rs[:5] == -1).all()
        assert (rs[6:10] == -2).all()
        assert (rs[20:30] == -3).all()
        assert (pd.isna(ser[:5])).all()

        # replace with different values with 2 lists
        rs2 = ser.replace([np.nan, 'foo', 'bar'], [-1, -2, -3])
        tm.assert_series_equal(rs, rs2)

        # replace inplace
        ser.replace([np.nan, 'foo', 'bar'], -1, inplace=True)
        assert (ser[:5] == -1).all()
        assert (ser[6:10] == -1).all()
        assert (ser[20:30] == -1).all() 

def _transformation_param_dependent(y, y_deg, A=0.98 / 49.98, B=0.02, C=50.0):
    aux = A - (1.0 - 2.0 * y_deg) * np.fabs(np.floor(0.5 - y_deg) + A)
    ret = np.power(y, B + (C - B) * aux)
    return correct_to_01(ret) 

def angle_between_vectors(v0, v1, directed=True, axis=0):
    """Return angle between vectors.

    If directed is False, the input vectors are interpreted as undirected axes,
    i.e. the maximum angle is pi/2.

    >>> a = angle_between_vectors([1, -2, 3], [-1, 2, -3])
    >>> numpy.allclose(a, math.pi)
    True
    >>> a = angle_between_vectors([1, -2, 3], [-1, 2, -3], directed=False)
    >>> numpy.allclose(a, 0)
    True
    >>> v0 = [[2, 0, 0, 2], [0, 2, 0, 2], [0, 0, 2, 2]]
    >>> v1 = [[3], [0], [0]]
    >>> a = angle_between_vectors(v0, v1)
    >>> numpy.allclose(a, [0, 1.5708, 1.5708, 0.95532])
    True
    >>> v0 = [[2, 0, 0], [2, 0, 0], [0, 2, 0], [2, 0, 0]]
    >>> v1 = [[0, 3, 0], [0, 0, 3], [0, 0, 3], [3, 3, 3]]
    >>> a = angle_between_vectors(v0, v1, axis=1)
    >>> numpy.allclose(a, [1.5708, 1.5708, 1.5708, 0.95532])
    True

    """
    v0 = numpy.array(v0, dtype=numpy.float64, copy=False)
    v1 = numpy.array(v1, dtype=numpy.float64, copy=False)
    dot = numpy.sum(v0 * v1, axis=axis)
    dot /= vector_norm(v0, axis=axis) * vector_norm(v1, axis=axis)
    return numpy.arccos(dot if directed else numpy.fabs(dot)) 

def _transformation_shift_multi_modal(y, A, B, C):
    tmp1 = np.fabs(y - C) / (2.0 * (np.floor(C - y) + C))
    tmp2 = (4.0 * A + 2.0) * np.pi * (0.5 - tmp1)
    ret = (1.0 + np.cos(tmp2) + 4.0 * B * np.power(tmp1, 2.0)) / (B + 2.0)
    return correct_to_01(ret) 

def angle_between_vectors(v0, v1, directed=True, axis=0):
    """Return angle between vectors.

    If directed is False, the input vectors are interpreted as undirected axes,
    i.e. the maximum angle is pi/2.

    >>> a = angle_between_vectors([1, -2, 3], [-1, 2, -3])
    >>> numpy.allclose(a, math.pi)
    True
    >>> a = angle_between_vectors([1, -2, 3], [-1, 2, -3], directed=False)
    >>> numpy.allclose(a, 0)
    True
    >>> v0 = [[2, 0, 0, 2], [0, 2, 0, 2], [0, 0, 2, 2]]
    >>> v1 = [[3], [0], [0]]
    >>> a = angle_between_vectors(v0, v1)
    >>> numpy.allclose(a, [0, 1.5708, 1.5708, 0.95532])
    True
    >>> v0 = [[2, 0, 0], [2, 0, 0], [0, 2, 0], [2, 0, 0]]
    >>> v1 = [[0, 3, 0], [0, 0, 3], [0, 0, 3], [3, 3, 3]]
    >>> a = angle_between_vectors(v0, v1, axis=1)
    >>> numpy.allclose(a, [1.5708, 1.5708, 1.5708, 0.95532])
    True

    """
    v0 = numpy.array(v0, dtype=numpy.float64, copy=False)
    v1 = numpy.array(v1, dtype=numpy.float64, copy=False)
    dot = numpy.sum(v0 * v1, axis=axis)
    dot /= vector_norm(v0, axis=axis) * vector_norm(v1, axis=axis)
    return numpy.arccos(dot if directed else numpy.fabs(dot))