QSpaceConverter.py 44.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
# coding: utf-8
# /*##########################################################################
#
# 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.
#
# ###########################################################################*/

26 27
from __future__ import absolute_import

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84
__authors__ = ["D. Naudet"]
__license__ = "MIT"
__date__ = "01/03/2016"


import os
import time
import ctypes
from threading import Thread
import multiprocessing as mp
import multiprocessing.sharedctypes as mp_sharedctypes

import numpy as np
import xrayutilities as xu

# from scipy.signal import medfilt2d

from ...util.filt_utils import medfilt2D
from ...util.histogramnd_lut import histogramnd_get_lut, histogramnd_from_lut
# from silx.math import histogramnd
from ...io import XsocsH5, QSpaceH5

disp_times = False


class QSpaceConverter(object):
    (READY, RUNNING, DONE,
     ERROR, CANCELED, UNKNOWN) = __STATUSES = range(6)
    """ Available status codes """

    status = property(lambda self: self.__status)
    """ Current status code of this instance """

    status_msg = property(lambda self: self.__status_msg)
    """ Status message if any, or None """

    results = property(lambda self: self.__results)
    """ Parse results. KmapParseResults instance. """

    xsocsH5_f = property(lambda self: self.__xsocsH5_f)
    """ Input file name. """

    output_f = property(lambda self: self.__output_f)
    """ Output file name. """

    qspace_dims = property(lambda self: self.__params['qspace_dims'])
    """ dimensions of the Q Space (i.e : number of bins). """

    image_binning = property(lambda self: self.__params['image_binning'])
    """ Binning applied to the images before conversion. """

    sample_indices = property(lambda self: self.__params['sample_indices'])
    """ Indices of sample positions that will be converted. """

    n_proc = property(lambda self: self.__n_proc)
    """ Number of processes to use. Will use cpu_count() if None or 0. """

Damien Naudet's avatar
Damien Naudet committed
85 86
    roi = property(lambda self: self.__params['roi'])
    """ Selected ROI in sample coordinates : [xmin, xmax, ymin, ymax] """
87 88 89 90 91 92

    def __init__(self,
                 xsocsH5_f,
                 qspace_dims=None,
                 img_binning=None,
                 output_f=None,
Damien Naudet's avatar
Damien Naudet committed
93
                 roi=None,
94
                 entries=None,
95 96 97 98 99 100
                 callback=None):
        """
        Merger for the Kmap SPEC and EDF files. This loads a spech5 file,
             converts it to HDF5 and then tries to match scans and edf image
             files.
        :param xsocsH5_f: path to the input XsocsH5 file.
101 102 103
        :param qspace_dims: dimensions of the qspace volume
        :param img_binning: binning to apply to the images before conversion.
            Default : (1, 1)
104
        :param output_f: path to the output file that will be created.
105 106 107
        :param roi: Roi in sample coordinates (xMin, xMax, yMin, yMax)
        :param entries: a list of entry names to convert to qspace. If None,
            all entries found in the xsocsH5File will be used.
108 109 110 111 112 113 114 115 116 117 118
        :param callback: callback to call when the parsing is done.
        """
        super(QSpaceConverter, self).__init__()

        self.__status = None

        self.__set_status(self.UNKNOWN, 'Init')

        self.__xsocsH5_f = xsocsH5_f
        self.__output_f = output_f

119 120 121 122 123 124 125 126 127 128 129
        xsocsH5 = XsocsH5.XsocsH5(xsocsH5_f)
        # checking entries
        if entries is None:
            entries = xsocsH5.entries()
        else:
            diff = set(entries) - set(xsocsH5.entries())
            if len(diff) > 0:
                raise ValueError('The following entries were not found in '
                                 'the input file :\n - {0}'
                                 ''.format('\n -'.join(diff)))

130 131 132
        self.__params = {'qspace_dims': None,
                         'image_binning': None,
                         'sample_indices': None,
133 134
                         'roi': None,
                         'entries': sorted(entries)}
135 136 137 138 139 140 141 142 143 144 145 146 147

        self.__callback = callback
        self.__n_proc = None
        self.__overwrite = False

        self.__shared_progress = None
        self.__results = None
        self.__term_evt = None

        self.__thread = None

        self.image_binning = img_binning
        self.qspace_dims = qspace_dims
Damien Naudet's avatar
Damien Naudet committed
148
        self.roi = roi
149 150 151 152 153

        self.__set_status(self.READY)

    def __get_scans(self):
        """
154
        Returns the entries that will be converted.
155
        """
156
        return self.__params['entries']
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287

    scans = property(__get_scans)
    """ Returns the scans found in the input file. """

    def __set_status(self, status, msg=None):
        """
        Sets the status of this instance.
        :param status:
        :param msg:
        :return:
        """
        assert status in self.__STATUSES
        self.__status = status
        self.__status_msg = msg

    def convert(self,
                overwrite=False,
                blocking=True,
                callback=None,
                check_consistency=True):
        """
        Starts the conversion.
        :param overwrite: if False raises an exception if some files already
        exist.
        :param blocking: if False, the merge will be done in a separate
         thread and this method will return immediately.
        :param callback: callback that will be called when the merging is done.
        It overwrites the one passed the constructor.
        :param check_consistency: set to False to ignore any incensitencies
        in the input entries (e.g : different counters, ...).
        :return:
        """

        if self.is_running():
            raise RuntimeError('This QSpaceConverter instance is already '
                               'parsing.')

        self.__set_status(self.RUNNING)

        errors = self.check_parameters()

        if len(errors) > 0:
            msg = 'Invalid parameters.\n{0}'.format('\n'.join(errors))
            raise ValueError(msg)

        errors = self.check_consistency()

        if len(errors) > 0:
            msg = 'Inconsistent input data.\n{0}'.format('\n'.join(errors))

            if check_consistency:
                raise ValueError(msg)
            else:
                print('==============.')
                print('==============.')
                print('WARNING.')
                print(msg)

        output_f = self.__output_f
        if output_f is None:
            self.__set_status(self.ERROR)
            raise ValueError('Output file name (output_f) has not been set.')

        output_dir = os.path.dirname(output_f)
        if not os.path.exists(output_dir):
            os.makedirs(output_dir)

        if not overwrite:
            if len(self.check_overwrite()):
                    self.__set_status(self.ERROR)
                    raise RuntimeError('Some files already exist. Use the '
                                       'overwrite keyword to ignore this '
                                       'warning.')

        self.__results = None
        self.__overwrite = overwrite

        if callback is not None:
            self.__callback = callback

        if blocking:
            self.__run_convert()
        else:
            thread = self.__thread = Thread(target=self.__run_convert)
            thread.start()

    @qspace_dims.setter
    def qspace_dims(self, qspace_dims):
        """
        Sets the dimensions of the qspace volume (i.e. number of bins).
        """

        if qspace_dims is None or None in qspace_dims:
            self.__params['qspace_dims'] = None
            return

        qspace_dims = np.array(qspace_dims, ndmin=1).astype(np.int32)

        if qspace_dims.ndim != 1 or qspace_dims.size != 3:
            raise ValueError('qspace_dims must be a three elements array.')

        if not np.all(qspace_dims > 1):
            raise ValueError('<qspace_dims> values must be strictly'
                             ' greater than one.')
        self.__params['qspace_dims'] = qspace_dims

    @image_binning.setter
    def image_binning(self, image_binning):
        """
        Binning applied to the image before converting to qspace
        """
        err = False
        if image_binning is None:
            self.__params['image_binning'] = (1, 1)
            return

        image_binning_int = None
        if len(image_binning) != 2:
            raise ValueError('image_binning must be a two elements array.')
        if None in image_binning:
            err = True
        else:
            image_binning_int = [int(image_binning[0]), int(image_binning[1])]
            if min(image_binning_int) <= 0:
                err = True
        if err:
            raise ValueError('<image_binning> values must be strictly'
                             ' positive integers.')
        self.__params['image_binning'] = np.array(image_binning_int,
                                                  dtype=np.int32)

288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
    # @sample_indices.setter
    # def sample_indices(self, sample_indices):
    #     """
    #     Binning applied to the image before converting to qspace
    #     """
    #     if sample_indices is None:
    #         self.__params['sample_indices'] = None
    #         return
    #
    #     sample_indices = np.array(sample_indices, ndmin=1).astype(np.long)
    #
    #     if sample_indices.ndim != 1:
    #         raise ValueError('sample_indices must be a 1D array.')
    #
    #     if len(sample_indices) == 0:
    #         self.__params['sample_indices'] = None
    #         return
    #
    #     # TODO : check values
    #     self.__params['sample_indices'] = np.array(sample_indices,
    #                                                dtype=np.int32)
309

Damien Naudet's avatar
Damien Naudet committed
310 311 312 313 314 315 316 317 318 319
    @roi.setter
    def roi(self, roi):
        """
        Sets the roi. Set to None to unset it. To change an already set roi
        the previous one has to be unset first.
        :param roi: roi coordinates in sample coordinates.
            Four elements array : (xmin, xmax, ymin, ymax)
        :return:
        """
        if self.roi is False:
320 321
            raise ValueError('Cannot set a rectangular ROI, pos_indices are '
                             'already set, remove them first.')
Damien Naudet's avatar
Damien Naudet committed
322
        self.__params['roi'] = roi
323 324 325 326 327 328 329 330 331 332 333
        self.__params['sample_indices'] = self.__indices_from_roi()

    def __indices_from_roi(self):
        # TODO : check all positions
        # at the moment using only the first scan's positions
        with XsocsH5.XsocsH5(self.__xsocsH5_f) as xsocsH5:
            entries = xsocsH5.entries()
            positions = xsocsH5.scan_positions(entries[0])
            x_pos = positions.pos_0
            y_pos = positions.pos_1

Damien Naudet's avatar
Damien Naudet committed
334 335
        roi = self.roi
        if self.roi is None:
336 337
            return np.arange(len(x_pos))

Damien Naudet's avatar
Damien Naudet committed
338 339 340 341
        x_min = roi[0]
        x_max = roi[1]
        y_min = roi[2]
        y_max = roi[3]
342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437

        # we cant do this because the points arent perfectly aligned!
        # we could end up with non rectangular rois
        pos_indices = np.where((x_pos >= x_min) & (x_pos <= x_max) &
                               (y_pos >= y_min) & (y_pos <= y_max))[0]
        # # TODO : rework this
        # n_x = scan_params['motor_0_steps']
        # n_y = scan_params['motor_1_steps']
        # steps_0 = scan_params['motor_0_steps']
        # steps_1 = scan_params['motor_1_steps']
        # x = np.linspace(scan_params['motor_0_start'],
        #                 scan_params['motor_0_end'], steps_0, endpoint=False)
        # y = np.linspace(scan_params['motor_1_start'],
        #                 scan_params['motor_1_end'], steps_1, endpoint=False)


        # x_pos = x_pos[]
        #
        # x_pos.shape = (n_y, n_x)
        # y_pos.shape = (n_y, n_x)
        # pos_indices_2d = np.where((x_pos >= x_min) & (x_pos <= x_max) &
        #                           (y_pos >= y_min) & (y_pos <= y_max))[0]
        return pos_indices  # pos_indices_2d.shape

    def check_overwrite(self):
        """
        Checks if the output file(s) already exist(s).
        """
        output_f = self.__output_f
        if output_f is not None and os.path.exists(output_f):
            return [self.__output_f]
        return []

    def summary(self):
        """
        Gives an summary of what will be done.
        """
        # TODO : finish
        files = [self.output_f]
        return files

    def check_parameters(self):
        """
        Checks if the RecipSpaceConverter parameters are valid.
        Returns a list of strings describing those errors, if any,
        or an empty list.
        """
        errors = []
        image_binning = self.image_binning
        qspace_dims = self.qspace_dims

        if (image_binning is None
                or None in image_binning
                or len(image_binning) != 2
                or image_binning.min() <= 0):
            errors.append('- "image binning" : must be an array of two'
                          ' strictly positive integers.')

        if (qspace_dims is None
                or None in qspace_dims
                or len(qspace_dims) != 3
                or qspace_dims.min() <= 0):
            errors.append('- "qspace size" must be an array of three'
                          ' strictly positive integers.')
        return errors

    def check_consistency(self):
        """
        Check if all entries have the same values plus some other
        MINIMAL checks.
        This does not check if the parameter values are valid.
        Returns a list of strings describing those errors, if any,
        or an empty list.
        """
        errors = []

        params = _get_all_params(self.__xsocsH5_f)

        def check_values(dic, key, description):
            values = [dic[scan][key] for scan in sorted(dic.keys())]
            if isinstance(values[0], (list, tuple)):
                values = [tuple(val) for val in values]
            values_set = set(values)
            if len(values_set) != 1:
                errors.append('Parameter inconsistency : '
                              '"{0}" : {1}.'
                              ''.format(description, '; '.join(str(m)
                                        for m in values_set)))

        check_values(params, 'n_images', 'Number of images')
        check_values(params, 'n_positions', 'Number of X/Y positions')
        check_values(params, 'img_size', 'Images size')
        check_values(params, 'beam_energy', 'Beam energy')
        check_values(params, 'chan_per_deg', 'Chan. per deg.')
        check_values(params, 'center_chan', 'Center channel')

Damien Naudet's avatar
Damien Naudet committed
438 439 440
        keys = list(params.keys())
        n_images = params[keys[0]]['n_images']
        n_positions = params[keys[0]]['n_positions']
441 442 443 444 445 446 447 448
        if n_images != n_positions:
            errors.append('number of images != number of X/Y coordinates '
                          'on sample : '
                          '{0} != {1}'.format(n_images, n_positions))

        return errors

    def scan_params(self, scan):
Damien Naudet's avatar
Damien Naudet committed
449
        """ Returns the scan parameters (filled during acquisition). """
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
        params = _get_all_params(self.__xsocsH5_f)
        return params[scan]

    def __run_convert(self):
        """
        Performs the conversion.
        :return:
        """

        self.__set_status(self.RUNNING)

        image_binning = self.image_binning
        qspace_dims = self.qspace_dims
        xsocsH5_f = self.xsocsH5_f
        output_f = self.output_f
465
        sample_roi = self.__params['roi']
466 467 468 469 470 471

        try:
            ta = time.time()

            params = _get_all_params(xsocsH5_f)

472
            entries = self.__get_scans()
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560
            n_entries = len(entries)

            first_param = params[entries[0]]

            beam_energy = first_param['beam_energy']
            if beam_energy is None:
                raise ValueError('Invalid/missing beam energy : {0}.'
                                 ''.format(beam_energy))

            chan_per_deg = first_param['chan_per_deg']
            if beam_energy is None or len(chan_per_deg) != 2:
                raise ValueError('Invalid/missing chan_per_deg value : {0}.'
                                 ''.format(chan_per_deg))

            center_chan = first_param['center_chan']
            if beam_energy is None or len(center_chan) != 2:
                raise ValueError('Invalid/missing center_chan value : {0}.'
                                 ''.format(center_chan))

            n_images = first_param['n_images']
            if n_images is None or n_images == 0:
                raise ValueError(
                    'Data does not contain any images (n_images={0}).'
                    ''.format(n_images))

            img_size = first_param['img_size']
            if img_size is None or 0 in img_size:
                raise ValueError('Invalid image size (img_size={0}).'
                                 ''.format(img_size))

            # TODO value testing
            sample_indices = self.sample_indices
            if sample_indices is None:
                sample_indices = np.arange(n_images)
            else:
                n_images = len(sample_indices)

            n_xy = len(sample_indices)

            print('Parameters :')
            print('\t- beam energy  : {0}'.format(beam_energy))
            print('\t- center chan  : {0}'.format(center_chan))
            print('\t- chan per deg : {0}'.format(chan_per_deg))
            print('\t- img binning : {0}'.format(image_binning))
            print('\t- qspace size : {0}'.format(qspace_dims))

            # TODO : make this editable?
            nx, ny, nz = qspace_dims
            qconv = xu.experiment.QConversion(['y-', 'z-'],
                                              ['z-', 'y-'],
                                              [1, 0, 0])

            # convention for coordinate system:
            # x downstream
            # z upwards
            # y to the "outside"
            # (righthanded)
            hxrd = xu.HXRD([1, 0, 0],
                           [0, 0, 1],
                           en=beam_energy,
                           qconv=qconv)

            hxrd.Ang2Q.init_area('z-',
                                 'y+',
                                 cch1=center_chan[0],
                                 cch2=center_chan[1],
                                 Nch1=img_size[0],
                                 Nch2=img_size[1],
                                 chpdeg1=chan_per_deg[0],
                                 chpdeg2=chan_per_deg[1],
                                 Nav=image_binning)

            # shape of the array that will store the qx/qy/qz for all
            # rocking angles
            q_shape = (n_entries,
                       (img_size[0] // image_binning[0]) * (
                           img_size[1] // image_binning[1]),
                       3)

            # then the array
            q_ar = np.zeros(q_shape, dtype=np.float64)

            img_dtype = None

            with XsocsH5.XsocsH5(xsocsH5_f, mode='r') as master_h5:

                entry_files = []

561 562
                all_entries = set(master_h5.entries())

563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
                positions = master_h5.scan_positions(entries[0])
                sample_x = positions.pos_0
                sample_y = positions.pos_1

                for entry_idx, entry in enumerate(entries):
                    entry_file = master_h5.entry_filename(entry)
                    if not os.path.isabs(entry_file):
                        base_dir = os.path.dirname(xsocsH5_f)
                        entry_file = os.path.abspath(os.path.join(base_dir,
                                                                  entry_file))
                    entry_files.append(entry_file)

                    phi = np.float64(master_h5.positioner(entry, 'phi'))
                    eta = np.float64(master_h5.positioner(entry, 'eta'))
                    nu = np.float64(master_h5.positioner(entry, 'nu'))
                    delta = np.float64(master_h5.positioner(entry, 'del'))

                    qx, qy, qz = hxrd.Ang2Q.area(eta, phi, nu, delta)
                    q_ar[entry_idx, :, 0] = qx.reshape(-1)
                    q_ar[entry_idx, :, 1] = qy.reshape(-1)
                    q_ar[entry_idx, :, 2] = qz.reshape(-1)

                    entry_dtype = master_h5.image_dtype(entry=entry)

                    if img_dtype is None:
                        img_dtype = entry_dtype
                    elif img_dtype != entry_dtype:
                        raise TypeError(
                            'All images in the input HDF5 files should '
                            'be of the same type. Found {0} and {1}.'
                            ''.format(img_dtype, entry_dtype))

            # custom bins range to have the same histo as xrayutilities.gridder3d
            # bins centered around the qx, qy, qz
            # bins will be like :
            # bin_1 = [min - step/2, min + step/2[
            # bin_2 = [min - step/2, min + 3*step/2]
            # ...
            # bin_N = [max - step/2, max + step/2]
            qx_min = q_ar[:, :, 0].min()
            qy_min = q_ar[:, :, 1].min()
            qz_min = q_ar[:, :, 2].min()
            qx_max = q_ar[:, :, 0].max()
            qy_max = q_ar[:, :, 1].max()
            qz_max = q_ar[:, :, 2].max()

            step_x = (qx_max - qx_min) / (nx - 1.)
            step_y = (qy_max - qy_min) / (ny - 1.)
            step_z = (qz_max - qz_min) / (nz - 1.)

            bins_rng_x = ([qx_min - step_x / 2., qx_min +
                           (qx_max - qx_min + step_x) - step_x / 2.])
            bins_rng_y = ([qy_min - step_y / 2., qy_min +
                           (qy_max - qy_min + step_y) - step_y / 2.])
            bins_rng_z = ([qz_min - step_z / 2., qz_min +
                           (qz_max - qz_min + step_z) - step_z / 2.])
            bins_rng = [bins_rng_x, bins_rng_y, bins_rng_z]

            qx_idx = qx_min + step_x * np.arange(0, nx, dtype=np.float64)
            qy_idx = qy_min + step_y * np.arange(0, ny, dtype=np.float64)
            qz_idx = qz_min + step_z * np.arange(0, nz, dtype=np.float64)

            # TODO : on windows we may be forced to use shared memory
            # TODO : find why we use more memory when using shared arrays
            #        this shouldnt be the case
            #        (use the same amount as non shared mem)
            # on linux apparently we dont because when fork() is called data is
            # only copied on write.
            # shared histo used by all processes
            # histo_shared = mp_sharedctypes.RawArray(ctypes.c_int32,
            #                                         nx * ny * nz)
            # histo = np.frombuffer(histo_shared, dtype='int32')
            # histo.shape = nx, ny, nz
            # histo[:] = 0
            histo = np.zeros(qspace_dims, dtype=np.int32)

            # shared LUT used by all processes
            # h_lut = None
            # h_lut_shared = None
            h_lut = []
            lut = None

            for h_idx in range(n_entries):
                lut = histogramnd_get_lut(q_ar[h_idx, ...],
                                          bins_rng,
                                          [nx, ny, nz],
                                          last_bin_closed=True)

                # if h_lut_shared is None:
                #     lut_dtype = lut[0].dtype
                #     if lut_dtype == np.int16:
                #         lut_ctype = ctypes.c_int16
                #     elif lut_dtype == np.int32:
                #         lut_ctype = ctypes.c_int32
                #     elif lut_dtype == np.int64:
                #         lut_ctype == ctypes.c_int64
                #     else:
                #         raise TypeError('Unknown type returned by '
                #                         'histogramnd_get_lut : {0}.'
                #                         ''.format(lut.dtype))
                #     h_lut_shared = mp_sharedctypes.RawArray(lut_ctype,
                #                                       n_images * lut[0].size)
                #     h_lut = np.frombuffer(h_lut_shared, dtype=lut_dtype)
                #     h_lut.shape = (n_images, -1)
                #
                # h_lut[h_idx, ...] = lut[0]
                h_lut.append(lut[0])
                histo += lut[1]

            del lut
            del q_ar

            # TODO : split the output file into several files? speedup?
            output_shape = histo.shape

            chunks = (1,
                      max(output_shape[0] // 4, 1),
                      max(output_shape[1] // 4, 1),
                      max(output_shape[2] // 4, 1),)
            qspace_sum_chunks = max(n_images // 10, 1),

684 685
            discarded_entries = sorted(all_entries - set(entries))

686 687
            _create_result_file(output_f,
                                output_shape,
688 689
                                image_binning,
                                sample_roi,
690 691 692 693 694 695
                                sample_x[sample_indices],
                                sample_y[sample_indices],
                                qx_idx,
                                qy_idx,
                                qz_idx,
                                histo,
696 697
                                selected_entries=entries,
                                discarded_entries=discarded_entries,
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
                                compression='lzf',
                                qspace_chunks=chunks,
                                qspace_sum_chunks=qspace_sum_chunks,
                                overwrite=self.__overwrite)

            manager = mp.Manager()
            self.__term_evt = term_evt = manager.Event()

            write_lock = manager.Lock()
            idx_queue = manager.Queue()

            n_proc = self.n_proc
            if n_proc is None or n_proc <= 0:
                n_proc = mp.cpu_count()

            self.__shared_progress = mp_sharedctypes.RawArray(ctypes.c_int32,
                                                              n_proc)
            np.frombuffer(self.__shared_progress, dtype='int32')[:] = 0

            pool = mp.Pool(n_proc,
                           initializer=_init_thread,
                           initargs=(idx_queue,
                                     write_lock,
                                     bins_rng,
                                     qspace_dims,
                                     h_lut,  # _shared,
                                     None,  # lut_dtype,
                                     n_xy,
                                     histo,  # _shared,))
                                     self.__shared_progress,
                                     term_evt,))

            if disp_times:
                class myTimes(object):
                    def __init__(self):
                        self.t_histo = 0.
                        self.t_sum = 0.
                        self.t_mask = 0.
                        self.t_read = 0.
                        self.t_context = 0.
                        self.t_dnsamp = 0.
                        self.t_medfilt = 0.
                        self.t_write = 0.
                        self.t_w_lock = 0.

                    def update(self, arg):
                        (t_read_, t_context_, t_dnsamp_, t_medfilt_, t_histo_,
                         t_mask_, t_sum_, t_write_, t_w_lock_) = arg[2]
                        self.t_histo += t_histo_
                        self.t_sum += t_sum_
                        self.t_mask += t_mask_
                        self.t_read += t_read_
                        self.t_context += t_context_
                        self.t_dnsamp += t_dnsamp_
                        self.t_medfilt += t_medfilt_
                        self.t_write += t_write_
                        self.t_w_lock += t_w_lock_

                res_times = myTimes()
                callback = res_times.update
            else:
                callback = None

            # creating the processes
            results = []
            for th_idx in range(n_proc):
                arg_list = (th_idx,
                            entry_files,
                            entries,
                            img_size,
                            output_f,
                            image_binning,
                            img_dtype)
                res = pool.apply_async(_to_qspace, args=arg_list,
                                       callback=callback)
                results.append(res)

            # sending the image indices
            for result_idx, pos_idx in enumerate(sample_indices):
                idx_queue.put((result_idx, pos_idx))

Damien Naudet's avatar
Damien Naudet committed
779 780
            # sending the None value to let the threads know that they
            # should return
781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842
            for th_idx in range(n_proc):
                idx_queue.put(None)

            pool.close()
            pool.join()

            tb = time.time()

            if disp_times:
                print('TOTAL {0}'.format(tb - ta))
                print('Read {0}'.format(res_times.t_read))
                print('Context {0}'.format(res_times.t_context))
                print('Dn Sample {0}'.format(res_times.t_dnsamp))
                print('Medfilt {0}'.format(res_times.t_medfilt))
                print('Histo {0}'.format(res_times.t_histo))
                print('Mask {0}'.format(res_times.t_mask))
                print('Sum {0}'.format(res_times.t_sum))
                print('Write {0}'.format(res_times.t_write))
                print('(lock : {0})'.format(res_times.t_w_lock))

            proc_results = [result.get() for result in results]
            proc_codes = np.array([proc_result[0]
                                   for proc_result in proc_results])

            rc = self.DONE
            if not np.all(proc_codes == self.DONE):
                if self.ERROR in proc_codes:
                    rc = self.ERROR
                elif self.CANCELED in proc_codes:
                    rc = self.CANCELED
                else:
                    raise ValueError('Unknown return code.')

            if rc != self.DONE:
                errMsg = 'Conversion failed. Process status :'
                for th_idx, result in enumerate(proc_results):
                    errMsg += ('\n- Proc {0} : rc={1}; {2}'
                               ''.format(th_idx, result[0], result[1]))
                self.__set_status(rc, errMsg)
            else:
                self.__set_status(rc)

        except Exception as ex:
            self.__set_status(self.ERROR, str(ex))
        else:
            self.__results = self.output_f

        # TODO : catch exception?
        if self.__callback:
            self.__callback()

        return self.__results

    def wait(self):
        """
        Waits until parsing is done, or returns if it is not running.
        :return:
        """
        if self.__thread:
            self.__thread.join()

    def __running_exception(self):
Damien Naudet's avatar
Damien Naudet committed
843
        """ Raises an exception if a conversion is in progress. """
844 845 846 847 848
        if self.is_running():
            raise RuntimeError('Operation not permitted while '
                               'a parse or merge in running.')

    def is_running(self):
Damien Naudet's avatar
Damien Naudet committed
849
        """ Returns True if a conversion is in progress. """
850 851 852 853 854
        return self.status == QSpaceConverter.RUNNING
        #self.__thread and self.__thread.is_alive()

    @output_f.setter
    def output_f(self, output_f):
Damien Naudet's avatar
Damien Naudet committed
855
        """ Sets the output file. """
856 857 858 859 860 861 862
        if not isinstance(output_f, str):
            raise TypeError('output_f must be a string. Received {0}'
                            ''.format(type(output_f)))
        self.__output_f = output_f

    @n_proc.setter
    def n_proc(self, n_proc):
Damien Naudet's avatar
Damien Naudet committed
863 864 865 866
        """ Sets the number of processes to use. If None or 0 the number of
            processes used will be the number returned by
            multiprocessing.cpu_count.
        """
867 868 869 870 871 872 873 874 875 876 877
        if n_proc is None:
            self.__n_proc = None
            return

        n_proc = int(n_proc)
        if n_proc <= 0:
            self.__n_proc = None
        else:
            self.__n_proc = n_proc

    def abort(self, wait=True):
Damien Naudet's avatar
Damien Naudet committed
878 879 880 881 882 883
        """
        Aborts the current conversion, if any.
        :param wait: set to False to return immediatly without waiting for the
        processes to return.
        :return:
        """
884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933
        if self.is_running():
            self.__term_evt.set()
            if wait:
                self.wait()

    def progress(self):
        """
        Returns the progress of the conversion.
        :return:
        """
        if self.__shared_progress:
            progress = np.frombuffer(self.__shared_progress, dtype='int32')
            return progress.max()
        return 0


def _init_thread(idx_queue_,
                 write_lock_,
                 bins_rng_,
                 qspace_size_,
                 h_lut_shared_,
                 h_lut_dtype_,
                 n_xy_,
                 histo_shared_,
                 shared_progress_,
                 term_evt_):
    global idx_queue, \
        write_lock, \
        bins_rng, \
        qspace_size, \
        h_lut_shared, \
        h_lut_dtype, \
        n_xy, \
        histo_shared, \
        shared_progress, \
        term_evt

    idx_queue = idx_queue_
    write_lock = write_lock_
    bins_rng = bins_rng_
    qspace_size = qspace_size_
    h_lut_shared = h_lut_shared_
    h_lut_dtype = h_lut_dtype_
    n_xy = n_xy_
    histo_shared = histo_shared_
    shared_progress = shared_progress_
    term_evt = term_evt_


def _create_result_file(h5_fn,
934 935
                        qspace_dims,
                        image_binning,
936
                        sample_roi,
937 938
                        pos_x,
                        pos_y,
939 940 941
                        q_x,
                        q_y,
                        q_z,
942
                        histo,
943 944
                        selected_entries,
                        discarded_entries=None,
945 946 947 948
                        compression='lzf',
                        qspace_chunks=None,
                        qspace_sum_chunks=None,
                        overwrite=False):
949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968
    """
    Initializes the output file.
    :param h5_fn: name of the file to initialize
    :param qspace_dims: dimensions of the q space
    :param image_binning: binning applied to the images
    :param pos_x: sample X positions (one for each qspace cube)
    :param pos_y: sample Y positions (one for each qspace cube)
    :param q_x: X coordinates of the qspace cube
    :param q_y: Y coordinates of the qspace cube
    :param q_z: Z coordinates of the qspace cube
    :param histo: histogram (number of hits per element of the qspace elements)
    :param selected_entries: list of input entries used for the conversion
    :param discarded_entries: list of input entries discarded, or None
    :param compression: datasets compression
    :param qspace_chunks: qspace chunking
    :param qspace_sum_chunks:
    :param overwrite: True to force overwriting the file if it already exists.
    :return:
    """

969 970 971 972 973 974 975 976 977 978 979
    if not overwrite:
        mode = 'w-'
    else:
        mode = 'w'

    dir_name = os.path.dirname(h5_fn)
    if len(dir_name) > 0 and not os.path.exists(dir_name):
        os.makedirs(dir_name)

    qspace_h5 = QSpaceH5.QSpaceH5Writer(h5_fn, mode=mode)
    qspace_h5.init_file(len(pos_x),
980
                        qspace_dims,
981 982 983 984 985 986
                        qspace_chunks=qspace_chunks,
                        qspace_sum_chunks=qspace_sum_chunks,
                        compression=compression)
    qspace_h5.set_histo(histo)
    qspace_h5.set_sample_x(pos_x)
    qspace_h5.set_sample_y(pos_y)
987 988 989 990 991
    qspace_h5.set_qx(q_x)
    qspace_h5.set_qy(q_y)
    qspace_h5.set_qz(q_z)
    qspace_h5.set_entries(selected_entries, discarded=discarded_entries)
    qspace_h5.set_image_binning(image_binning)
992
    qspace_h5.set_sample_roi(sample_roi)
993 994 995 996 997 998 999 1000 1001


def _to_qspace(th_idx,
               entry_files,
               entries,
               img_size,
               output_fn,
               image_binning,
               img_dtype):
Damien Naudet's avatar
Damien Naudet committed
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
    """
    Fonction running in a process. Performs the conversion.
    :param th_idx:
    :param entry_files:
    :param entries:
    :param img_size:
    :param output_fn:
    :param image_binning:
    :param img_dtype:
    :return:
    """
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
    print('Thread {0} started.'.format(th_idx))

    t_histo = 0.
    t_mask = 0.
    t_sum = 0.
    t_read = 0.
    t_dnsamp = 0.
    t_medfilt = 0.
    t_write = 0.
    t_w_lock = 0.
    t_context = 0.

    output_h5 = QSpaceH5.QSpaceH5Writer(output_fn, mode='r+')

    if shared_progress is not None:
        progress_np = np.frombuffer(shared_progress, dtype='int32')
        progress_np[th_idx] = 0
    else:
        progress_np = None

    # histo = np.frombuffer(histo_shared, dtype='int32')
    # histo.shape = qspace_size
    histo = histo_shared
    mask = histo > 0

    # h_lut = np.frombuffer(h_lut_shared, dtype=h_lut_dtype)
    # h_lut.shape = (n_xy, -1)
    h_lut = h_lut_shared

    img = np.ascontiguousarray(np.zeros(img_size), dtype=img_dtype)

    # TODO : handle case when nav is not a multiple of img_size!!
    # TODO : find why the first version is faster than the second one
    img_shape_1 = img_size[0] // image_binning[0], image_binning[0], img_size[1]
    img_shape_2 = (img_shape_1[0], img_shape_1[2] // image_binning[1],
                   image_binning[1])
    sum_axis_1 = 1
    sum_axis_2 = 2
    # img_shape_1 = img_size[0], img_size[1]/nav[1], nav[1]
    # img_shape_2 = img_size[0]//nav[0], nav[0], img_shape_1[1]
    # sum_axis_1 = 2
    # sum_axis_2 = 1
    avg_weight = 1. / (image_binning[0] * image_binning[1])

    rc = None
    errMsg = None
    try:
        while True:
            if term_evt.is_set():  # noqa
                rc = QSpaceConverter.CANCELED
                raise Exception('conversion aborted')

            next_data = idx_queue.get()
            if next_data is None:
                rc = QSpaceConverter.DONE
                break

            result_idx, image_idx = next_data
            if result_idx % 100 == 0:
                print('#{0}/{1}'.format(result_idx, n_xy))

            cumul = None
            # histo = None

            for entry_idx, entry in enumerate(entries):

                t0 = time.time()

                try:
                    # TODO : there s room for improvement here maybe
                    # (recreating a XsocsH5 instance each time slows down
                    # slows down things a big, not much tho)
                    # TODO : add a lock on the files if there is no SWMR
                    # test if it slows down things much
                    with XsocsH5.XsocsH5(entry_files[entry_idx],
                                         mode='r').image_dset_ctx(entry) \
                            as img_data:  # noqa
                        t1 = time.time()
                        img_data.read_direct(img,
                                             source_sel=np.s_[image_idx],
                                             dest_sel=None)
                        t_context = time.time() - t1
                        # img = img_data[image_idx].astype(np.float64)
                except Exception as ex:
                    raise RuntimeError('Error in proc {0} while reading '
                                       'img {1} from entry {2} ({3}) : {4}.'
                                       ''.format(th_idx, image_idx, entry_idx,
                                                 entry, ex))

                t_read += time.time() - t0
                t0 = time.time()

                if image_binning[0] != 1 or image_binning[1] != 1:
                    intensity = img.reshape(img_shape_1). \
                                    sum(axis=sum_axis_1,
                                        dtype=np.uint32).reshape(img_shape_2). \
                                    sum(axis=sum_axis_2, dtype=np.uint32) * \
                                avg_weight
                    # intensity = xu.blockAverage2D(img, nav[0], nav[1], roi=roi)
                else:
                    intensity = img

                t_dnsamp += time.time() - t0
                t0 = time.time()

                # intensity = medfilt2d(intensity, 3)
                intensity = medfilt2D(intensity, kernel=[3, 3], n_threads=None)

                t_medfilt += time.time() - t0
                t0 = time.time()

                try:
                    cumul = histogramnd_from_lut(intensity.reshape(-1),
                                                 h_lut[entry_idx],
                                                 shape=qspace_size,
                                                 weighted_histo=cumul,
                                                 dtype=np.float64)
                except Exception as ex:
                    print('EX2 {0}'.format(str(ex)))
                    raise ex

                t_histo += time.time() - t0

            t0 = time.time()
            cumul_sum = cumul.sum(dtype=np.float64)
            t_sum += time.time() - t0

            t0 = time.time()
            # cumul[mask] = cumul[mask]/histo[mask]
            t_mask += time.time() - t0

            t0 = time.time()
            write_lock.acquire()
            t_w_lock += time.time() - t0
            t0 = time.time()
            try:
                output_h5.set_position_data(result_idx, cumul, cumul_sum)
            except Exception as ex:
                raise RuntimeError('Error in proc {0} while writing result '
                                   'for img {1} (idx = {3}) : {2}.)'
                                   ''.format(th_idx, image_idx, ex, result_idx))
            write_lock.release()

            if progress_np is not None:
                progress_np[th_idx] = round(100. * (result_idx + 1.) / n_xy)

            t_write += time.time() - t0
    except Exception as ex:
        if rc is None:
            rc = QSpaceConverter.ERROR
        errMsg = 'In thread {0} : {1}.'.format(th_idx, str(ex))
        term_evt.set()

    if rc is None:
        rc = QSpaceConverter.DONE

    if disp_times:
        print('Thread {0} is done. Times={1}'
              ''.format(th_idx, (t_read, t_context, t_dnsamp,
                                 t_medfilt, t_histo,
                                 t_mask, t_sum, t_write, t_w_lock)))
    return [rc, errMsg, (t_read, t_context, t_dnsamp,
                         t_medfilt, t_histo,
                         t_mask, t_sum, t_write, t_w_lock,)]


def _get_all_params(data_h5f):
    """
    Read the whole data and returns the parameters for each entry.
    Returns a dictionary will the scans as keys and the following fields :
    n_images, n_positions, img_size, beam_energy, chan_per_deg,
    center_chan
    Each of those fields are N elements arrays, where N is the number of
    scans found in the file.
    """
    n_images = []
    n_positions = []
    img_sizes = []
    beam_energies = []
    center_chans = []
    chan_per_degs = []
    angles = []

    with XsocsH5.XsocsH5(data_h5f, mode='r') as master_h5:
        entries = master_h5.entries()

        for entry_idx, entry in enumerate(entries):
            n_image = master_h5.n_images(entry=entry)
            img_size = master_h5.image_size(entry=entry)

            imgnr = master_h5.measurement(entry, 'imgnr')
            n_position = len(imgnr) if imgnr is not None else None

            beam_energy = master_h5.beam_energy(entry=entry)
            chan_per_deg = master_h5.chan_per_deg(entry=entry)
            center_chan = master_h5.direct_beam(entry=entry)

            angle = master_h5.positioner(entry, 'eta')

            n_images.append(n_image)
            n_positions.append(n_position)
            img_sizes.append(img_size)
            beam_energies.append(beam_energy)
            chan_per_degs.append(chan_per_deg)
            center_chans.append(center_chan)
            angles.append(angle)

    result = dict([(scan, dict(scans=entries[idx],
                               n_images=n_images[idx],
                               n_positions=n_positions[idx],
                               img_size=img_sizes[idx],
                               beam_energy=beam_energies[idx],
                               chan_per_deg=chan_per_degs[idx],
                               center_chan=center_chans[idx],
                               angle=angles[idx]))
                   for idx, scan in enumerate(entries)])
    return result


if __name__ == '__main__':
    pass