def createStarFile(self, starfile, mrcStackFile):
'''
Create a file with the required constant strings in it
'''
star = starFile.StarFile(starfile)
labels = [
'_rlnImageName',
'_rlnMicrographName',
'_rlnDefocusU',
'_rlnDefocusV',
'_rlnDefocusAngle',
'_rlnVoltage',
'_rlnSphericalAberration',
'_rlnAmplitudeContrast',
]
valueSets = [] #list of strings for star file
partParamsList = self.getStackParticleParams()
stackPath = os.path.join(self.params['rundir'], mrcStackFile)
for partParams in partParamsList:
relionDataLine = ("%d@%s %d %.6f %.6f %.6f %d %.6f %.6f" % (
partParams['ptclnum'],
stackPath,
partParams['filmNum'],
partParams['defocus2'],
partParams['defocus1'],
partParams['angle_astigmatism'],
partParams['kv'],
self.params['cs'],
partParams['amplitude_contrast'],
))
valueSets.append(relionDataLine)
star = starFile.StarFile(starfile)
star.buildLoopFile("data_", labels, valueSets)
star.write()
def parseParticleDataIterationFile(self,paramfile,test=False):
'''
parse data.star file from Relion 1.2:
'''
if not os.path.isfile(paramfile):
apDisplay.printError("Relion data.star file does not exist: %s" % (paramfile))
apDisplay.printMsg("Parsing parameter file: %s" % (paramfile))
# Use the star file class to read the parameter file
f = starFile.StarFile( paramfile )
f.read()
dataBlock = f.getDataBlock("data_images")
loopDict = dataBlock.getLoopDict() # there is only one loop in the data_images block
partnum = 1 # partnum starts with 1, not 0
partdict = {}
for valueSet in loopDict:
paramdict = {
'partnum' : partnum,
'angleRot' : float(valueSet["_rlnAngleRot"]),
'angleTilt' : float(valueSet["_rlnAngleTilt"]),
'anglePsi' : float(valueSet["_rlnAnglePsi"]),
'originX' : float(valueSet["_rlnOriginX"]),
'originY' : float(valueSet["_rlnOriginY"]),
}
partdict[paramdict['partnum']] = paramdict
partnum = partnum + 1
if len(partdict) < 2:
apDisplay.printError("No particles found in particle data file %s" % (paramfile))
apDisplay.printMsg("Processed %d particles" % (len(partdict)))
return partdict
def convertFSCFileForIteration(self, iteration):
''' RElion FSC info is in recon_it002_model.star and the final recon_model.star file, convert this to 3DEM format
For the final iteration, we need to use recon_model.star rather than the iteration specific file to get the actual values.
'''
lastiter = self.findLastCompletedIteration()
if iteration == lastiter:
fscfile = os.path.join(self.projmatchpath, "recon_model.star"%(iteration))
else:
fscfile = os.path.join(self.projmatchpath, "recon_it%.3d_half1_model.star"%(iteration))
f = starFile.StarFile( fscfile )
try:
f.read()
except e:
if iteration == lastiter:
apDisplay.printWarning("%s file could not be opened, data will NOT be inserted into the database. Relion did NOT run to completion." % fscfile)
else:
apDisplay.printWarning("%s file could not be opened, data will NOT be inserted into the database" % fscfile)
return False
dataBlock = f.getDataBlock("data_model_class_1")
loopDict = dataBlock.getLoopDict()
newfscfile = open(os.path.join(self.resultspath, "recon_%s_it%.3d_vol001.fsc" % (self.params['timestamp'], iteration)), "w")
newfscfile.write("### column (1) inverse Angstroms, column (2) Fourier Shell Correlation (FSC)")
for valueSet in loopDict:
resolution = float(valueSet["_rlnResolution"]) # already in inverse angstrom
fsc = float(valueSet["_rlnGoldStandardFsc"]) # should go from a value of 1 to 0
newfscfile.write("%.6f\t%.6f\n" % (resolution, fsc))
newfscfile.close()
return True
def readParamsFromOptimiserFile(self, iternum):
"""
read the recon_it000_optimiser.star file to get iteration parameters,
such as symmetry and phase residual cutoff
"""
iterparams = {'iternum': iternum,}
#recondir = os.path.join(self.params['rundir'], "recon")
recondir = self.params['rundir']
optimiserfile = "recon_it%03d_optimiser.star"%(iternum)
optimiserfilepath = os.path.join(recondir, optimiserfile)
# Create a dictionary of parameters found in the optimiser file
paramdict = {}
### parse optimiser file to populate dictionary
f = starFile.StarFile( optimiserfilepath )
f.read()
dataBlock = f.getDataBlock("data_optimiser_general")
paramdict = dataBlock.getLabelDict()
header = f.getHeader()
# Parse the header for the command line options used to run relion
sheader = header.strip()
bits = sheader.split("--")
for bit in bits:
params = bit.split()
key = params[0]
if len(params) > 1 :
value = params[1]
paramdict[key] = value
else:
paramdict[key] = True
if 'ini_high' not in paramdict:
paramdict['ini_high'] = 0
iterparams['ctf'] = self.convertBool(paramdict['_rlnDoCorrectCtf']) if paramdict['_rlnDoCorrectCtf'] else False
iterparams['ctf_intact_first_peak'] = self.convertBool(paramdict['_rlnDoIgnoreCtfUntilFirstPeak']) if paramdict['_rlnDoIgnoreCtfUntilFirstPeak'] else False
iterparams['ctf_corrected_ref'] = self.convertBool(paramdict['_rlnRefsAreCtfCorrected']) if paramdict['_rlnRefsAreCtfCorrected'] else False
iterparams['offset_step'] = int(paramdict['offset_step'])
iterparams['auto_local_healpix_order'] = int(paramdict['_rlnAutoLocalSearchesHealpixOrder'])
iterparams['healpix_order'] = int(paramdict['healpix_order'])
iterparams['offset_range'] = int(paramdict['offset_range'])
iterparams['ini_high'] = int(paramdict['ini_high'])
iterparams['sym'] = paramdict['sym']
### convert symetry to Appion
symtext = apFrealign.convertFrealignSymToAppionSym(iterparams['sym'])
symmdata = apSymmetry.findSymmetry(symtext)
apDisplay.printMsg("Found symmetry %s with id %s"%(symmdata['eman_name'], symmdata.dbid))
iterparams['symmdata'] = symmdata
return iterparams
def writeLoopDictToStarFile(loopDict, datablockname, starfilename): labels = loopDict[0].keys() valueSets = [] for line in loopDict: params = "" for p in line.values(): params += " %s" % p valueSets.append(params) star = starFile.StarFile(starfilename) star.buildLoopFile(datablockname, labels, valueSets) star.write()
def excludeClassesFromRelionDataFile2(starfile, outlist, *classlist):
### input is Relion data star file, e.g. run1_it025_data.star
### outputs a list without the particles belonging to excluded classes
### list is EMAN-style, particle numbering starts with 0 (note that in Relion particle numbering starts with 1)
### *classlist can be as many classes as you wish separated by comma, e.g. 1 or 1,3,6; classes are numbered Relion format, starting with 1
star = starFile.StarFile(starfile)
star.read()
dataBlock = star.getDataBlock("data_images")
loopDict = dataBlock.getLoopDict()
badparts = []
goodparts = []
for i in range(len(loopDict)):
rlnpart = float(loopDict[i]['_rlnImageName'].split("@")[0]) # starts with 1
rlnclass = float(loopDict[i]['_rlnClassNumber'])
bad = False
badpart = 0
for c in classlist:
if float(c) == rlnclass:
badpart = int(rlnpart)
bad = True
if bad is True:
badparts.append(badpart)
for i in range(len(loopDict)):
if (i+1) not in badparts:
goodparts.append(i+1)
f = open(outlist, "w")
for gp in goodparts:
f.write("%d\n" % (gp-1)) # EMAN-style numbering, starts with 0
f.close()
loopDictNew = starFile.LoopBlock
# loopDictNew = []
print dir(loopDict), "loop Dict"
print dir(loopDictNew), "loop Dict New"
print len(loopDict), "length of loop Dict"
print loopDict[0]
for gp in goodparts:
loopDictNew.append(loopDict[gp-1])
print len(loopDictNew), "length of loop Dict New"
dataBlockNew = starFile.DataBlock("data_images")
dataBlockNew.setLoopBlocks([loopDictNew])
print dir(dataBlockNew)
print len(dataBlockNew.loopBlocks)
star.setDataBlock(dataBlockNew,"data_images")
dbnew = star.getDataBlock("data_images")
print dbnew.name, "****"
print dir(dbnew)
print len(dbnew.loopBlocks)
loopDict = dbnew.getLoopDict()
print len(loopDict)
star.write("destination.star")
def listFileFromRelionStarFile(starfile, outlist, datablock="data_images"):
### write out an EMAN-style list file (numbering starts with 0) based on a relion star file with images and image numbers
star = starFile.StarFile(starfile)
star.read()
dataBlock = star.getDataBlock(datablock)
loopDict = dataBlock.getLoopDict()
allparts = []
for i in range(len(loopDict)):
allparts.append(int(loopDict[i]['_rlnImageName'].split("@")[0])) # starts with 1
allparts.sort()
f = open(outlist, "w")
for p in allparts:
f.write("%d\n" % (p-1)) # EMAN-style numbering, starts with 0
f.close()
def sortRelionStarFileByParticleNumber(instarfile, outstarfile, datablock="data_images"):
star = starFile.StarFile(instarfile)
star.read()
dataBlock = star.getDataBlock(datablock)
loopDict = dataBlock.getLoopDict()
partdict = {}
for i in range(len(loopDict)):
rlnpart = (int(loopDict[i]['_rlnImageName'].split("@")[0])) # starts with 1
partdict[rlnpart] = loopDict[i]
l = sorted(partdict, key=lambda key: partdict[key])
l.sort()
# write sorted star file
loopDictNew = []
for val in l:
loopDictNew.append(partdict[val])
writeLoopDictToStarFile(loopDictNew, datablock, outstarfile)
def excludeClassesFromRelionDataFile(instarfile, datablock, outlist, outstarfile, *classlist):
### input is Relion data star file, e.g. run1_it025_data.star
### datablock: The start of a data block is defined by the keyword "data_" followed by an optional string for identification (e.g., "data_images")
### outputs a list without the particles belonging to excluded classes
### list is EMAN-style, particle numbering starts with 0 (note that in Relion particle numbering starts with 1)
### outstarfile is the output star file
### *classlist can be as many classes as you wish separated by comma, e.g. 1 or 1,3,6; classes are numbered Relion format, starting with 1
star = starFile.StarFile(instarfile)
star.read()
dataBlock = star.getDataBlock(datablock)
loopDict = dataBlock.getLoopDict()
badparts = []
goodparts = []
loopDictNew = []
for i in range(len(loopDict)):
rlnpart = float(loopDict[i]['_rlnImageName'].split("@")[0]) # starts with 1
rlnclass = float(loopDict[i]['_rlnClassNumber'])
bad = False
badpart = 0
for c in classlist:
if float(c) == rlnclass:
badpart = int(rlnpart)
bad = True
if bad is True:
badparts.append(badpart)
else:
goodparts.append(int(rlnpart))
loopDictNew.append(loopDict[i])
badparts.sort()
goodparts.sort()
f = open(outlist, "w")
for gp in goodparts:
f.write("%d\n" % (gp-1)) # EMAN-style numbering, starts with 0
f.close()
# selected star file parameters, write star file with excluded images
writeLoopDictToStarFile(loopDictNew, datablock, outstarfile)
def start(self):
"""
for each particle in the stack, get the information that RELION needs
"""
stackPartList = apStack.getStackParticlesFromId(self.params['stackid'])
nptcls = len(stackPartList)
currentImageId = stackPartList[0]['particle']['image'].dbid
count = 0
imagenumber = 1
partParamsList = []
sys.stderr.write("reading stack particle data\n")
#create list of dictionaries that will be passed to starFile maker later
for stackPart in stackPartList:
count += 1
if count % 100 == 0:
sys.stderr.write(".")
if count % 10000 == 0:
sys.stderr.write("\nparticle %d of %d\n" % (count, nptcls))
# extra particle number information not read by Relion
if count != stackPart['particleNumber']:
apDisplay.printWarning(
"particle number in database is not in sync")
partParams = {}
partParams['ptclnum'] = count
### get image data
imagedata = stackPart['particle']['image']
if imagedata.dbid != currentImageId:
imagenumber += 1
currentImageId = imagedata.dbid
partParams['filmNum'] = imagenumber
#print partParams['filmNum']
partParams['kv'] = imagedata['scope']['high tension'] / 1000.0
partParams['cs'] = imagedata['scope']['tem']['cs'] * 1000
### get CTF data from image
ctfdata = ctfdb.getBestCtfValue(imagedata,
msg=False,
sortType='maxconf')
if ctfdata is not None:
# use defocus & astigmatism values
partParams['defocus1'] = abs(ctfdata['defocus1'] * 1e10)
partParams['defocus2'] = abs(ctfdata['defocus2'] * 1e10)
partParams['angle_astigmatism'] = ctfdata['angle_astigmatism']
partParams['amplitude_contrast'] = ctfdata[
'amplitude_contrast']
else:
apDisplay.printError(
"No ctf information for particle %d in image %s" %
(count, imagedata['filename']))
partParamsList.append(partParams)
###now make star file
#first make header
star = starFile.StarFile(self.params['outstar'])
labels = [
'_rlnImageName',
'_rlnMicrographName',
'_rlnDefocusU',
'_rlnDefocusV',
'_rlnDefocusAngle',
'_rlnVoltage',
'_rlnSphericalAberration',
'_rlnAmplitudeContrast',
]
valueSets = [] #list of strings for star file
###now make particle data
for partParams in partParamsList:
relionDataLine = (
"%[email protected] mic%d %.6f %.6f %.6f %d %.6f %.6f" % (
partParams['ptclnum'],
partParams['filmNum'],
partParams['defocus2'],
partParams['defocus1'],
partParams['angle_astigmatism'],
partParams['kv'],
partParams['cs'],
partParams['amplitude_contrast'],
))
valueSets.append(relionDataLine)
star.buildLoopFile("data_", labels, valueSets)
star.write()
def readStarFileDataBlock(starfile, datablock): star = starFile.StarFile(starfile) star.read() dataBlock = star.getDataBlock(datablock) loopDict = dataBlock.getLoopDict() return loopDict
