peak_fit.py

#!/usr/bin/python
# coding: utf8
# /*##########################################################################
#
# Copyright (c) 2015-2016 European Synchrotron Radiation Facility
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
# THE SOFTWARE.
#
# ###########################################################################*/

from __future__ import absolute_import, division

__authors__ = ["D. Naudet"]
__date__ = "01/06/2016"
__license__ = "MIT"

import logging
import functools
import multiprocessing

import numpy as np
from scipy.optimize import leastsq

from silx.math.fit import snip1d

from ... import config
from ...io import QSpaceH5
from ...io.FitH5 import BackgroundTypes
from ...util import gaussian, project
from .fitresults import FitResult


_logger = logging.getLogger(__name__)


def background_estimation(mode, data):
    """Estimates a background of mode kind for data

    :param BackgroundTypes mode: The kind of background to compute
    :param numpy.ndarray data: The data array to process
    :return: The estimated background as an array with same shape as input
    :rtype: numpy.ndarray
    :raises ValueError: In case mode is not known
    """
    # Background subtraction
    if mode == BackgroundTypes.CONSTANT:
        # Shift data so that smallest value is 0
        return np.ones_like(data) * np.nanmin(data)

    elif mode == BackgroundTypes.LINEAR:
        # Simple linear background
        return np.linspace(data[0], data[-1], num=len(data), endpoint=True)

    elif mode == BackgroundTypes.SNIP:
        # Using snip background
        return snip1d(data, snip_width=len(data))

    elif mode == BackgroundTypes.NONE:
        return np.zeros_like(data)

    else:
        raise ValueError("Unsupported background mode")


class FitTypes(object):
    """Kind of fit available"""
    ALLOWED = range(2)
    GAUSSIAN, CENTROID = ALLOWED


class PeakFitter(object):
    """Class performing fit/com processing

    :param str qspace_f: path to the HDF5 file containing the qspace cubes
    :param FitTypes fit_type:
    :param Union[numpy.ndarray,None] indices:
        indices of the cubes (in the input HDF5 dataset) for which
        the dim0/dim1/dim2 peaks coordinates will be computed. E.g : if the array
        [1, 2, 3] is provided, only those cubes will be fitted.
    :param Union[int,None] n_proc:
        Number of process to use. If None, the config value is used.
    :param Union[List[List[int]],None] roi_indices: QSpace ROI to process
    :param BackgroundTypes background: The background subtraction to perform
    """

    READY, RUNNING, DONE, ERROR, CANCELED = __STATUSES = range(5)

    def __init__(self,
                 qspace_f,
                 fit_type=FitTypes.GAUSSIAN,
                 indices=None,
                 n_proc=None,
                 roi_indices=None,
                 background=BackgroundTypes.NONE):
        super(PeakFitter, self).__init__()

        self.__results = None
        self.__status = self.READY

        self.__qspace_f = qspace_f
        self.__fit_type = fit_type
        self.__background = background

        self.__n_proc = n_proc if n_proc else config.DEFAULT_PROCESS_NUMBER

        if roi_indices is not None:
            self.__roi_indices = np.array(roi_indices[:])
        else:
            self.__roi_indices = None

        if fit_type not in FitTypes.ALLOWED:
            self.__set_status(self.ERROR)
            raise ValueError('Unknown fit type: {0}'.format(fit_type))

        if background not in BackgroundTypes.ALLOWED:
            self.__set_status(self.ERROR)
            raise ValueError('Unknown background type: {}'.format(background))

        try:
            with QSpaceH5.QSpaceH5(qspace_f) as qspace_h5:
                with qspace_h5.qspace_dset_ctx() as dset:
                    qdata_shape = dset.shape
        except IOError:
            self.__set_status(self.ERROR)
            raise

        if indices is None:
            n_points = qdata_shape[0]
            self.__indices = np.arange(n_points)
        else:
            self.__indices = np.array(indices, copy=True)

    def __set_status(self, status):
        assert status in self.__STATUSES
        self.__status = status

    status = property(lambda self: self.__status)

    results = property(lambda self: self.__results)

    def peak_fit(self, blocking=True):
        """Run fit/com processing

        :param bool blocking:
            True for blocking until the end of the processing,
            False for yielding progress during the processing.
        """
        self.__results = None

        if blocking:
            for _ in self.__peak_fit():
                pass
        else:
            for progress, _ in enumerate(self.__peak_fit()):
                yield progress / len(self.__indices)

    def __peak_fit(self):
        self.__set_status(self.RUNNING)

        pool = multiprocessing.Pool(self.__n_proc)

        fit_results = []
        for result in pool.imap(
                functools.partial(_fit_process,
                                  qspace_f=self.__qspace_f,
                                  fit_type=self.__fit_type,
                                  background_type=self.__background,
                                  roiIndices=self.__roi_indices),
                self.__indices):
            fit_results.append(result)
            yield result
        pool.close()
        pool.join()

        # Prepare FitResult object
        with QSpaceH5.QSpaceH5(self.__qspace_f) as qspace_h5:
            x_pos = qspace_h5.sample_x[self.__indices]
            y_pos = qspace_h5.sample_y[self.__indices]
            q_dim0, q_dim1, q_dim2 = qspace_h5.qspace_dimension_values

        if self.__roi_indices is not None:
            q_dim0 = q_dim0[self.__roi_indices[0][0]:self.__roi_indices[0][1]]
            q_dim1 = q_dim1[self.__roi_indices[1][0]:self.__roi_indices[1][1]]
            q_dim2 = q_dim2[self.__roi_indices[2][0]:self.__roi_indices[2][1]]

        if self.__fit_type == FitTypes.GAUSSIAN:
            fit_name = 'Gaussian'
            result_name = 'gauss_0'
            result_dtype = [('Area', np.float64),
                            ('Center', np.float64),
                            ('Sigma', np.float64),
                            ('Status', np.bool_)]

        elif self.__fit_type == FitTypes.CENTROID:
            fit_name = 'Centroid'
            result_name = 'centroid'
            result_dtype = [('COM', np.float64),
                            ('I_sum', np.float64),
                            ('I_max', np.float64),
                            ('Pos_max', np.float64),
                            ('Status', np.bool_)]

        else:
            raise RuntimeError('Unknown Fit Type')

        results = FitResult(entry=fit_name,
                            sample_x=x_pos,
                            sample_y=y_pos,
                            q_x=q_dim0,
                            q_y=q_dim1,
                            q_z=q_dim2,
                            background_mode=self.__background)
        fit_results = np.array(fit_results, dtype=result_dtype)
        # From points x axes to axes x points
        fit_results = np.transpose(fit_results)
        for axis_index, array in enumerate(fit_results):
            for name, _ in result_dtype[:-1]:
                results._add_axis_result(result_name, axis_index, name, array[name])
            results._set_axis_status(axis_index, array['Status'])

        self.__results = results
        self.__set_status(self.DONE)


# Per process cache for fit process
_per_process_cache = None


def _fit_process(index,
                 qspace_f,
                 fit_type=FitTypes.GAUSSIAN,
                 background_type=BackgroundTypes.NONE,
                 roiIndices=None):
    """Run fit processing.

    It loads a QSpace, extracts a ROI from it, project to axes,
    and then for each axis, it subtracts a background and performs a fit/COM.

    This function is run through a multiprocessing.Pool

    :param int index: The index of the QSpace to process
    :param str qspace_f: Filename of the hdf5 file containing QSpace
    :param FitTypes fit_type: The kind of fit to perform
    :param BackgroundTypes background_type:
        The kind of background subtraction to perform
    :param Union[List[List[int]],None] roiIndices:
        Optional QSpace ROI start:end in the 3 dimensions
    :return: Fit results as a list of results for dim0, dim1 and dim2
    :rtype: List[List[Union[float,bool]]]
    """
    global _per_process_cache
    if _per_process_cache is None:  # Initialize per process cache
        qspace_h5 = QSpaceH5.QSpaceH5(qspace_f)
        _per_process_cache = (
            qspace_h5, qspace_h5.qspace_dimension_values, qspace_h5.histo)

    # Retrieve data/file from cache
    qspace_h5, axes, hits = _per_process_cache

    # Load qspace
    qspace = qspace_h5.qspace_slice(index)

    # apply Qspace ROI
    if roiIndices is not None:
        dim0Slice = slice(roiIndices[0][0], roiIndices[0][1], 1)
        dim1Slice = slice(roiIndices[1][0], roiIndices[1][1], 1)
        dim2Slice = slice(roiIndices[2][0], roiIndices[2][1], 1)

        axes = [axis[roi] for axis, roi in
                zip(axes, (dim0Slice, dim1Slice, dim2Slice))]
        hits = hits[dim0Slice, dim1Slice, dim2Slice]
        qspace = qspace[dim0Slice, dim1Slice, dim2Slice]

    # Normalize with hits and project to axes
    projections = project(qspace, hits)

    # Background subtraction
    if background_type != BackgroundTypes.NONE:
        for array in projections:
            array -= background_estimation(background_type, array)

    # Fit/COM
    fit = {FitTypes.CENTROID: centroid,
           FitTypes.GAUSSIAN: gaussian_fit}[fit_type]
    result = [fit(axis, values) for axis, values in zip(axes, projections)]

    return result


# Center of mass

def centroid(axis, signal):
    """Returns Center of mass and maximum information

    :param numpy.ndarray axis: 1D x data
    :param numpy.ndarray signal: 1D y data
    :return: Center of mass, sum of signal, max of signal, position of max and status
        ('COM', 'I_sum', 'I_max', 'Pos_max', 'status')
    :rtype: List[Union[float,bool]]
    """
    signal_sum = signal.sum()
    if signal_sum == 0:
        return float('nan'), float('nan'), float('nan'), float('nan'), False

    else:
        max_idx = signal.argmax()
        return (float(np.dot(axis, signal) / signal_sum),
                float(signal_sum),
                float(signal[max_idx]),
                float(axis[max_idx]),
                True)


# Gaussian fit

def _gaussian_err(parameters, axis, signal):
    """Returns difference between signal and given gaussian

    :param List[float] parameters: area, center, sigma
    :param numpy.ndarray axis: 1D x data
    :param numpy.ndarray signal: 1D y data
    :return:
    """
    area, center, sigma = parameters
    return gaussian(axis, area, center, sigma) - signal


_SQRT_2_PI = np.sqrt(2 * np.pi)


def gaussian_fit(axis, signal):
    """Returns gaussian fitting information

    parameters: (a, c, s)
    and f(x) = (a / (sqrt(2 * pi) * s)) * exp(- 0.5 * ((x - c) / s)^2)

    :param axis: 1D x data
    :param signal: 1D y data
    :return: Area, center, sigma, status
        ('Area', 'Center', 'Sigma', 'status)
    :rtype: List[Union[float,bool]]
    """
    # compute guess
    area = signal.sum() * (axis[-1] - axis[0]) / len(axis)
    center = axis[signal.argmax()]
    sigma = area / (signal.max() * _SQRT_2_PI)

    # Fit a gaussian
    result = leastsq(_gaussian_err,
                     x0=(area, center, sigma),
                     args=(axis, signal),
                     maxfev=100000,
                     full_output=True)

    if result[4] not in [1, 2, 3, 4]:
        return float('nan'), float('nan'), float('nan'), False

    else:
        area, center, sigma = result[0]
        return float(area), float(center), float(sigma), True