Source code for cistem.protocols.protocol_unblur

# **************************************************************************
# *
# * Authors:     Roberto Marabini (roberto@cnb.csic.es) [1]
# *              Josue Gomez Blanco (josue.gomez-blanco@mcgill.ca) [2]
# *              Grigory Sharov (gsharov@mrc-lmb.cam.ac.uk) [3]
# *
# * [1] Unidad de  Bioinformatica of Centro Nacional de Biotecnologia , CSIC
# * [2] Department of Anatomy and Cell Biology, McGill University
# * [3] MRC Laboratory of Molecular Biology, MRC-LMB
# *
# * This program is free software; you can redistribute it and/or modify
# * it under the terms of the GNU General Public License as published by
# * the Free Software Foundation; either version 3 of the License, or
# * (at your option) any later version.
# *
# * This program is distributed in the hope that it will be useful,
# * but WITHOUT ANY WARRANTY; without even the implied warranty of
# * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# * GNU General Public License for more details.
# *
# * You should have received a copy of the GNU General Public License
# * along with this program; if not, write to the Free Software
# * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
# * 02111-1307  USA
# *
# *  All comments concerning this program package may be sent to the
# *  e-mail address 'scipion@cnb.csic.es'
# *
# **************************************************************************

import time
from math import ceil
from threading import Thread

import pyworkflow.utils as pwutils

from pyworkflow.protocol import STEPS_PARALLEL
from pyworkflow.constants import PROD
import pyworkflow.protocol.params as params
from pyworkflow.gui.plotter import Plotter
from pwem.objects import Image
from pwem.protocols import ProtAlignMovies

from cistem import Plugin
from ..convert import readShiftsMovieAlignment
from ..constants import UNBLUR_BIN


[docs]class CistemProtUnblur(ProtAlignMovies): """ This protocol wraps unblur movie alignment program. """ _label = 'unblur' _devStatus = PROD CONVERT_TO_MRC = 'mrc' def __init__(self, **args): ProtAlignMovies.__init__(self, **args) self.stepsExecutionMode = STEPS_PARALLEL def _getConvertExtension(self, filename): """ Check whether it is needed to convert to .mrc or not """ ext = pwutils.getExt(filename).lower() return None if ext in ['.mrc', '.mrcs', '.tiff', '.tif'] else 'mrc' def _defineAlignmentParams(self, form): form.addHidden('doSaveAveMic', params.BooleanParam, default=True) form.addHidden('useAlignToSum', params.BooleanParam, default=True) group = form.addGroup('Alignment') line = group.addLine('Frames to ALIGN', help='Frames range to ALIGN on each movie. The ' 'first frame is 1. If you set 0 in the final ' 'frame to align, it means that you will ' 'align until the last frame of the movie.') line.addParam('alignFrame0', params.IntParam, default=1, label='from') line.addParam('alignFrameN', params.IntParam, default=0, label='to') group.addParam('binFactor', params.FloatParam, default=1., label='Binning factor', help='1x or 2x. Bin stack before processing.') form.addParam('doComputePSD', params.BooleanParam, default=False, expertLevel=params.LEVEL_ADVANCED, label="Compute PSD?", help="If Yes, the protocol will compute for each " "aligned micrograph the PSD using EMAN2.") form.addParam('doComputeMicThumbnail', params.BooleanParam, expertLevel=params.LEVEL_ADVANCED, default=False, label='Compute micrograph thumbnail?', help='When using this option, we will compute a ' 'micrograph thumbnail with EMAN2 and keep it with the ' 'micrograph object for visualization purposes. ') form.addParam('extraProtocolParams', params.StringParam, default='', expertLevel=params.LEVEL_ADVANCED, label='Additional protocol parameters', help="Here you can provide some extra parameters for the " "protocol, not the underlying unblur program." "You can provide many options separated by space. " "\n\n*Options:* \n\n" "--use_worker_thread \n" " Use an extra thread to compute" " PSD and thumbnail. This will allow requires " "an extra CPU. ") form.addSection(label='Expert Options') line = form.addLine('Shifts (A): ', help='Min and max shifts during alignment.\n\n' 'The minimum shift can be applied ' 'during the initial refinement stage. ' 'Its purpose is to prevent images aligning ' 'to detector artifacts that may be ' 'reinforced in the initial sum which is ' 'used as the first reference. It is ' 'applied only during the first alignment ' 'round, and is ignored after that.\n' 'The maximum shift can be applied in any ' 'single alignment round. Its purpose is ' 'to avoid alignment to spurious noise ' 'peaks by not considering unreasonably ' 'large shifts. This limit is applied ' 'during every alignment round, but only' ' for that round, such that it can be ' 'exceeded over a number of successive rounds.') line.addParam('minShiftInitSearch', params.FloatParam, default='2.0', label='Min shift') line.addParam('OutRadShiftLimit', params.FloatParam, default='40.0', label='Max shift') group = form.addGroup('Exposure filter') group.addParam('doApplyDoseFilter', params.BooleanParam, default=True, label='Exposure filter sums?', help='If selected the resulting aligned movie sums ' 'will be calculated using the exposure filter ' 'as described in Grant and Grigorieff (2015). ' 'Pre-exposure and dose per frame ' 'should be specified during movies import.') group.addParam('doRestoreNoisePwr', params.BooleanParam, default=True, label='Restore power? ', help='If selected, and the exposure filter is used ' 'to calculate the sum then the sum will be ' 'high pass filtered to restore the noise ' 'power. This is essentially the denominator ' 'of Eq. 9 in Grant and Grigorieff (2015).') group = form.addGroup('Convergence') group.addParam('terminShiftThreshold', params.FloatParam, default=1.0, label='Termination threshold (A)', help='The frames will be iteratively aligned ' 'until either the maximum number of ' 'iterations is reached, or if after an ' 'alignment round every frame was shifted ' 'by less than this threshold.') group.addParam('maximumNumberIterations', params.IntParam, default=20, label='Max iterations', help='The maximum number of iterations that ' 'can be run for the movie alignment. ' 'If reached, the alignment will stop ' 'and the current best values will be taken.') group = form.addGroup('Filter') group.addParam('bfactor', params.FloatParam, default=1500., label='B-factor (A^2)', help='This B-Factor is applied to the reference sum ' 'prior to alignment. It is intended to low-pass ' 'filter the images in order to prevent ' 'alignment to spurious noise peaks and ' 'detector artifacts.') line = group.addLine('Mask central cross?', help='If selected, the Fourier transform of ' 'the reference will be masked by a cross ' 'centred on the origin of the transform. ' 'This is intended to reduce the influence ' 'of detector artifacts which often have ' 'considerable power along the central cross.') line.addParam('HWHoriFourMask', params.IntParam, default=1, label='Horiz. mask (px)') line.addParam('HWVertFourMask', params.IntParam, default=1, label='Vert. mask (px)') form.addParallelSection(threads=1, mpi=1) # --------------------------- STEPS functions ----------------------------- def _processMovie(self, movie): inputMovies = self.getInputMovies() self._createTifLink(movie) self._argsUnblur(movie) try: self.runJob(self._getProgram(), self._args, env=Plugin.getEnviron()) def _extraWork(): outMicFn = self._getMicFn(movie) if self.doComputePSD: self._computePSD(outMicFn, outputFn=self._getPsdCorr(movie)) self._saveAlignmentPlots(movie, inputMovies.getSamplingRate()) if self._doComputeMicThumbnail(): self.computeThumbnail(outMicFn, outputFn=self._getOutputMicThumbnail(movie)) if self._useWorkerThread(): thread = Thread(target=_extraWork) thread.start() else: _extraWork() except Exception as e: self.error("ERROR: Unblur has failed for %s. %s" % ( self._getMovieFn(movie), self._getErrorFromUnblurTxt(movie, e))) def _getErrorFromUnblurTxt(self, movie, e): """ Parse output log for errors. :param movie: input movie object :return: the error string """ file = self._getShiftsFn(movie) with open(file, "r") as fh: for line in fh.readlines(): if line.startswith("Error"): return line.replace("Error:", "") return e def _insertFinalSteps(self, deps): stepId = self._insertFunctionStep('waitForThreadStep', prerequisites=deps) return [stepId]
[docs] def waitForThreadStep(self): # Quick and dirty (maybe desperate) way to wait # if the PSD and thumbnail were computed in a thread # If running in streaming this will not be necessary if self._useWorkerThread(): time.sleep(10) # wait 10 sec for the thread to finish
# --------------------------- INFO functions ------------------------------- def _summary(self): summary = [] if hasattr(self, 'outputMicrographs') or \ hasattr(self, 'outputMicrographsDoseWeighted'): summary.append('Aligned %d movies using unblur.' % self.inputMovies.get().getSize()) else: summary.append('Output is not ready') return summary def _citations(self): return ["Campbell2012", "Grant2015b"] def _validate(self): # Check base validation before the specific ones errors = ProtAlignMovies._validate(self) if self.doApplyDoseFilter and self.inputMovies.get(): inputMovies = self.inputMovies.get() doseFrame = inputMovies.getAcquisition().getDosePerFrame() if doseFrame == 0.0 or doseFrame is None: errors.append('Dose per frame for input movies is 0 or not ' 'set. You cannot apply dose filter.') if self.doComputeMicThumbnail or self.doComputePSD: try: from pwem import Domain eman2 = Domain.importFromPlugin('eman2', doRaise=True) except: errors.append("EMAN2 plugin not found!\nComputing thumbnails " "or PSD requires EMAN2 plugin and binaries installed.") return errors # --------------------------- UTILS functions ----------------------------- def _getProgram(self): """ Return program binary. """ return Plugin.getProgram(UNBLUR_BIN) def _argsUnblur(self, movie): """ Format arguments to call unblur program. """ inputMovies = self.getInputMovies() if self.doApplyDoseFilter: preExp, dose = self._getCorrectedDose(inputMovies) else: preExp, dose = 0.0, 0.0 args = {'movieName': self._getMovieFn(movie), 'micFnName': self._getMicFn(movie), 'shiftsFn': self._getShiftsFn(movie), 'samplingRate': self.samplingRate, 'voltage': movie.getAcquisition().getVoltage(), 'bfactor': self.bfactor.get(), 'minShiftInitSearch': self.minShiftInitSearch.get(), 'OutRadShiftLimit': self.OutRadShiftLimit.get(), 'HWVertFourMask': self.HWVertFourMask.get(), 'HWHoriFourMask': self.HWHoriFourMask.get(), 'terminShiftThreshold': self.terminShiftThreshold.get(), 'maximumNumberIterations': self.maximumNumberIterations.get(), 'applyDoseFilter': 'YES' if self.doApplyDoseFilter else 'NO', 'doRestoreNoisePwr': 'YES' if self.doRestoreNoisePwr else 'NO', 'exposurePerFrame': dose, 'binFactor': self.binFactor.get(), 'alignFrame0': self.alignFrame0.get(), 'alignFrameN': self.alignFrameN.get(), 'gainCorrected': 'NO' if inputMovies.getGain() else 'YES', 'gainFn': inputMovies.getGain(), 'preExposureAmount': preExp } argsStr = """ << eof > %(shiftsFn)s %(movieName)s %(micFnName)s %(samplingRate)f %(binFactor)f %(applyDoseFilter)s""" if self.doApplyDoseFilter: argsStr += """ %(voltage)f %(exposurePerFrame)f %(preExposureAmount)f""" argsStr += """ YES %(minShiftInitSearch)f %(OutRadShiftLimit)f %(bfactor)f %(HWVertFourMask)d %(HWHoriFourMask)d %(terminShiftThreshold)f %(maximumNumberIterations)d""" if self.doApplyDoseFilter: argsStr += """ %(doRestoreNoisePwr)s""" if inputMovies.getGain(): argsStr += """ %(gainCorrected)s %(gainFn)s %(alignFrame0)d %(alignFrameN)d NO eof\n """ else: argsStr += """ %(gainCorrected)s %(alignFrame0)d %(alignFrameN)d NO eof\n """ self._args = argsStr % args def _getMovieFn(self, movie): movieFn = movie.getFileName() if movieFn.endswith("tiff"): return pwutils.replaceExt(movieFn, "tif") else: return movieFn def _createTifLink(self, movie): # unblur recognises only tif, not tiff movieFn = movie.getFileName() if movieFn.endswith("tiff"): pwutils.createLink(movieFn, self._getMovieFn(movie)) def _getMicFn(self, movie): if self.doApplyDoseFilter: return self._getExtraPath(self._getOutputMicWtName(movie)) else: return self._getExtraPath(self._getOutputMicName(movie)) def _getShiftsFn(self, movie): return self._getExtraPath(self._getMovieRoot(movie) + '_shifts.txt') def _getMovieShifts(self, movie): """ Returns the x and y shifts for the alignment of this movie. """ pixSize = movie.getSamplingRate() shiftFn = self._getShiftsFn(movie) xShifts, yShifts = readShiftsMovieAlignment(shiftFn) # convert shifts from Angstroms to px xShiftsCorr = [x / pixSize for x in xShifts] yShiftsCorr = [y / pixSize for y in yShifts] return xShiftsCorr, yShiftsCorr def _doComputeMicThumbnail(self): return self.doComputeMicThumbnail def _computePSD(self, inputFn, outputFn, scaleFactor=6): """ Generate a thumbnail of the PSD with EMAN2""" args = "%s %s " % (inputFn, outputFn) args += "--process=math.realtofft --meanshrink %s " % scaleFactor args += "--fixintscaling=sane" from pwem import Domain eman2 = Domain.importFromPlugin('eman2') from pyworkflow.utils.process import runJob runJob(self._log, eman2.Plugin.getProgram('e2proc2d.py'), args, env=eman2.Plugin.getEnviron()) return outputFn def _preprocessOutputMicrograph(self, mic, movie): mic.plotGlobal = Image(location=self._getPlotGlobal(movie)) if self.doComputePSD: mic.psdCorr = Image(location=self._getPsdCorr(movie)) if self._doComputeMicThumbnail(): mic.thumbnail = Image(location=self._getOutputMicThumbnail(movie)) def _getNameExt(self, movie, postFix, ext, extra=False): fn = self._getMovieRoot(movie) + postFix + '.' + ext return self._getExtraPath(fn) if extra else fn def _getPlotGlobal(self, movie): return self._getNameExt(movie, '_global_shifts', 'png', extra=True) def _getPsdCorr(self, movie): return self._getNameExt(movie, '_psd', 'png', extra=True) def _saveAlignmentPlots(self, movie, pixSize): """ Compute alignment shift plots and save to file as png images. """ shiftsX, shiftsY = self._getMovieShifts(movie) first, _ = self._getFrameRange(movie.getNumberOfFrames(), 'align') plotter = createGlobalAlignmentPlot(shiftsX, shiftsY, first, pixSize) plotter.savefig(self._getPlotGlobal(movie)) plotter.close() def _useWorkerThread(self): return '--use_worker_thread' in self.extraProtocolParams.get()
[docs] def getInputMovies(self): return self.inputMovies.get()
def _createOutputMicrographs(self): return not self.doApplyDoseFilter def _createOutputWeightedMicrographs(self): return self.doApplyDoseFilter
[docs]def createGlobalAlignmentPlot(meanX, meanY, first, pixSize): """ Create a plotter with the shift per frame. """ sumMeanX = [] sumMeanY = [] def px_to_ang(apx): y1, y2 = apx.get_ylim() x1, x2 = apx.get_xlim() ax_ang2.set_ylim(y1*pixSize, y2*pixSize) ax_ang.set_xlim(x1*pixSize, x2*pixSize) ax_ang.figure.canvas.draw() ax_ang2.figure.canvas.draw() figureSize = (6, 4) plotter = Plotter(*figureSize) figure = plotter.getFigure() ax_px = figure.add_subplot(111) ax_px.grid() ax_px.set_xlabel('Shift x (px)') ax_px.set_ylabel('Shift y (px)') ax_ang = ax_px.twiny() ax_ang.set_xlabel('Shift x (A)') ax_ang2 = ax_px.twinx() ax_ang2.set_ylabel('Shift y (A)') i = first skipLabels = ceil(len(meanX)/10.0) labelTick = 1 for x, y in zip(meanX, meanY): sumMeanX.append(x) sumMeanY.append(y) if labelTick == 1: ax_px.text(x - 0.02, y + 0.02, str(i)) labelTick = skipLabels else: labelTick -= 1 i += 1 # automatically update lim of ax_ang when lim of ax_px changes. ax_px.callbacks.connect("ylim_changed", px_to_ang) ax_px.callbacks.connect("xlim_changed", px_to_ang) ax_px.plot(sumMeanX, sumMeanY, color='b') ax_px.plot(sumMeanX, sumMeanY, 'yo') ax_px.plot(sumMeanX[0], sumMeanY[0], 'ro', markersize=10, linewidth=0.5) ax_px.set_title('Global frame alignment') plotter.tightLayout() return plotter