# Original author - Krish Subramaniam # $Id: fsutils.py,v 1.4.2.6 2013/01/31 19:23:49 greve Exp $ import os import logging import sys from misc import * from subject_info import * from datastruct_utils import * # logging for both stats2table ch = logging.StreamHandler() #create logger aseglogger = logging.getLogger("asegstats2table") aseglogger.setLevel(logging.INFO) aseglogger.addHandler(ch) aparclogger = logging.getLogger("aparcstats2table") aparclogger.setLevel(logging.INFO) aparclogger.addHandler(ch) tractlogger = logging.getLogger("tractstats2table") tractlogger.setLevel(logging.INFO) tractlogger.addHandler(ch) class BadFileError(Exception): def __init__(self, filename): self.filename = filename def __str__(self): return self.filename """ This is the base class which parses the .stats files. """ class StatsParser: measure_column_map = {} #to be defined in subclass fp = None # filepointer include_structlist = None #parse only these structs exclude_structlist = None #exclude these structs structlist = None #list of structures measurelist = None #list of measures corresponding to the structures # len ( structlist ) must be equal to len (measurelist ) # constructor just needs a .stats filename # load it and report error def __init__(self, filename): self.filename = filename # raise exception if file doesn't exist or # is too small to be a valid stats file if not os.path.exists(filename): raise BadFileError(filename) if os.path.getsize(filename) < 10: raise BadFileError(filename) self.fp = open(filename, 'r') self.include_structlist = [] self.exclude_structlist = [] self.structlist = [] self.measurelist = [] # parse only the following structures def parse_only(self, structlist): # this is simply an Ordered Set # we need this because if inputs repeat, this will take care self.include_structlist = StableDict() for struct in structlist: self.include_structlist[struct] = 1 self.structlist = [] self.measurelist = [] # exclude the following structures def exclude_structs(self, structlist): # this is simply an Ordered Set # we need this because if inputs repeat, this will take care self.exclude_structlist = StableDict() for struct in structlist: self.exclude_structlist[struct] = 1 self.structlist = [] self.measurelist = [] # actual parsing will be done by subclass def parse(self): pass """ aseg.stats parser ( or parser for similarly formatted .stats files ) Derived from StatsParser """ class AsegStatsParser(StatsParser): measure_column_map = {'nvoxels':2,'nvertices':2,'volume':3,'Area_mm2':3, 'mean':5, 'std':6, 'max':8, 'snr':10} maxsegno = None id_name_map = StableDict() # segs > max_seg_no are discarded def set_maxsegno(self, maxsegno): self.maxsegno = int(maxsegno) # we take in the measure we need.. def parse(self, measure): self.id_name_map = StableDict() for line in self.fp: # valid line if line.rfind('#') == -1: strlst = line.split() seg = strlst[1] # segmentation id is in the 2nd field if self.maxsegno and int(seg) > self.maxsegno: continue if self.exclude_structlist and seg in self.exclude_structlist.keys(): continue self.id_name_map[seg] = strlst[4] # Col 4 is Seg Name # Float of the measure we are interested in self.measurelist.append( float(strlst[self.measure_column_map[measure]] )) # if we have a spec which instructs the table to have only specified segs, # we need to make sure the order also has to be maintained # hence we create temp dicts and copy them to the return dicts # If the segmentation name is not present in the current statsfile, # it is simply named as 'Placeholder_Segmentation' in the hope that we'd bump # into this in the future if self.include_structlist: tmp_id_name_map = StableDict() tmp_ml = [] for oseg in self.include_structlist.keys(): idlist = self.id_name_map.keys() if oseg in idlist: corres_meas_index = idlist.index( oseg ) tmp_id_name_map[oseg] = self.id_name_map[oseg] tmp_ml.append( self.measurelist[corres_meas_index]) else: tmp_id_name_map[oseg] ='Placeholder_Segmentation' tmp_ml.append( 0.0 ) self.id_name_map = tmp_id_name_map self.measurelist = tmp_ml[:] # edge case in case of measure = volume self.fp.seek(0) for line in self.fp: # volume measures.. which are found at the beginning of files and the # values are in the 4th column ( comma separated ) # segidlist is not useful here, all the values of segidlist will be # 'Placeholder_ID' # this is because len of all 3 lists -- structlist, measurelist and segidlists should be same if measure == 'volume': beg_struct_tuple = ( ('# Measure lhCortex, lhCortexVol,', 'lhCortexVol'), ('# Measure rhCortex, rhCortexVol,', 'rhCortexVol'), ('# Measure Cortex, CortexVol,', 'CortexVol'), ('# Measure lhCorticalWhiteMatter, lhCorticalWhiteMatterVol,','lhCorticalWhiteMatterVol'), ('# Measure rhCorticalWhiteMatter, rhCorticalWhiteMatterVol,','rhCorticalWhiteMatterVol'), ('# Measure CorticalWhiteMatter, CorticalWhiteMatterVol,','CorticalWhiteMatterVol'), ('# Measure SubCortGray, SubCortGrayVol,','SubCortGrayVol'), ('# Measure TotalGray, TotalGrayVol,','TotalGrayVol'), ('# Measure SuperTentorial, SuperTentorialVol,','SuperTentorialVol'), ('# Measure SupraTentorial, SupraTentorialVol,','SupraTentorialVol'), ('# Measure SupraTentorialNotVent, SupraTentorialVolNotVent,','SupraTentorialVolNotVent'), ('# Measure SupraTentorialNotVentVox, SupraTentorialVolNotVentVox,','SupraTentorialVolNotVentVox'), ('# Measure IntraCranialVol, ICV,','IntraCranialVol'), ('# Measure EstimatedTotalIntraCranialVol, eTIV,','EstimatedTotalIntraCranialVol'), ('# Measure Mask, MaskVol,','MaskVol'), ('# Measure BrainVol-to-eTIV, BrainVol-to-eTIV,','BrainVol-to-eTIV'), ('# Measure BrainSegVol-to-eTIV, BrainSegVol-to-eTIV,','BrainSegVol-to-eTIV'), ('# Measure MaskVol-to-eTIV, MaskVol-to-eTIV,','MaskVol-to-eTIV'), ('# Measure lhSurfaceHoles, lhSurfaceHoles,','lhSurfaceHoles'), ('# Measure rhSurfaceHoles, rhSurfaceHoles,','rhSurfaceHoles'), ('# Measure SurfaceHoles, SurfaceHoles,','SurfaceHoles'), ('# Measure BrainSeg, BrainSegVol,','BrainSegVol'), ('# Measure BrainSegNotVent, BrainSegVolNotVent,','BrainSegVolNotVent'), ('# Measure BrainSegNotVentSurf, BrainSegVolNotVentSurf,','BrainSegVolNotVentSurf'),) c = 0 for start, structn in beg_struct_tuple: c = c + 1 if line.startswith(start) : strlst = line.split(',') self.id_name_map['Placeholder_ID'+str(c)] = structn self.measurelist.append( float(strlst[3])) # check sanity assert( len(self.id_name_map) == len(self.measurelist)) # return return self.id_name_map, self.measurelist """ ?h.aparc*.stats parser ( or parser for similarly formatted .stats files ) Derived from StatsParser """ class AparcStatsParser(StatsParser): # this is a map of measure requested and its corresponding column# in ?h.aparc*.stats measure_column_map = {'area':2, 'volume':3, 'thickness':4, 'thicknessstd':5, 'meancurv':6, 'gauscurv':7, 'foldind':8, 'curvind':9 } parc_measure_map = StableDict() # we take in the measure we need.. def parse(self, measure): self.parc_measure_map = StableDict() for line in self.fp: # a valid line is a line without a '#' if line.rfind('#') == -1: strlist = line.split() # for every parcellation parcid = strlist[0] val = float(strlist[self.measure_column_map[measure]]) self.parc_measure_map[parcid] = val # if we have a spec which instructs the table to have only specified parcs, # we need to make sure the order has to be maintained if self.include_structlist: tmp_parc_measure_map = StableDict() for oparc in self.include_structlist.keys(): parclist = self.parc_measure_map.keys() if oparc in parclist: tmp_parc_measure_map[oparc] = self.parc_measure_map[oparc] else: tmp_parc_measure_map[oparc] = 0.0 self.parc_measure_map = tmp_parc_measure_map # measures which are found at the beginning of files. self.fp.seek(0) for line in self.fp: if measure == 'area': beg_struct_tuple = ( ('# Measure Cortex, WhiteSurfArea,', 'WhiteSurfArea'), ) for start, structn in beg_struct_tuple: if line.startswith(start): strlst = line.split(',') self.parc_measure_map[structn] = float( strlst[3]) if measure == 'thickness': beg_struct_tuple = ( ('# Measure Cortex, MeanThickness,', 'MeanThickness'), ) for start, structn in beg_struct_tuple: if line.startswith(start): strlst = line.split(',') self.parc_measure_map[structn] = float( strlst[3]) return self.parc_measure_map """ tracula overall.stats file parser ( or parser for similarly formatted .stats files ) Derived from StatsParser """ class TractOverallStatsParser(StatsParser): # this is a map of measure and its corresponding value in tract.overall.stats measure_value_map = StableDict() # we take in the measure we need.. def parse(self): self.measure_value_map = StableDict() pre_tuple = ( ('# pathwayname','pathway'), ('# subjectname','subject'), ) pthwy_subj = StableDict() for line in self.fp: # a valid line is a line without a '#' if line.rfind('#') == -1: strlist = line.split() # for every measure assign value meas = strlist[0] # if measure is in excluded measure list, we are not interested if self.exclude_structlist and meas in self.exclude_structlist.keys(): continue val = float(strlist[1]) self.measure_value_map[meas] = val else: # lookfor subjectname and pathwayname for start, what in pre_tuple: if line.startswith(start): strlst = line.split(' ') name = strlst[2].strip() pthwy_subj[what] = name # if we have a spec which instructs the table to have only specified measures, # we need to make sure the order has to be maintained if self.include_structlist: tmp_measure_value_map = StableDict() for omeasure in self.include_structlist.keys(): measurelist = self.measure_value_map.keys() if omeasure in measurelist: tmp_measure_value_map[omeasure] = self.measure_value_map[omeasure] else: print 'ERROR: The measure you requested: '+ omeasure + ' is not valid' print 'It is not found one of the overall statistics file' sys.exit(1) self.measure_value_map = tmp_measure_value_map return pthwy_subj, self.measure_value_map """ tracula byvoxel.stats file parser ( or parser for similarly formatted .stats files ) Derived from StatsParser """ class TractByvoxelStatsParser(StatsParser): # this is a map of measure and its corresponding value in tract.byvoxel.stats measure_value_map = StableDict() # map of measures needed and their corresponding column in byvoxel.stats file measure_column_map = {'AD':3, 'RD':4, 'MD':5, 'FA':6 } # we take in the measure we need.. def parse(self, measure): self.measure_value_map = StableDict() pre_tuple = ( ('# pathwayname','pathway'), ('# subjectname','subject'), ) pthwy_subj = StableDict() # scroll to the start of data. find subjectname # and pathwayname meanwhile line = self.fp.readline() while not line.startswith('# pathway start'): # lookfor subjectname and pathwayname for start, what in pre_tuple: if line.startswith(start): strlst = line.split(' ') name = strlst[2].strip() pthwy_subj[what] = name line = self.fp.readline() line = self.fp.readline() # skip to the 'x y z AD RD MD FA line line = self.fp.readline() # skip to the actual data # scroll until the end of data collecting the spec-ed measure count = 0 while not line.startswith('# pathway end'): strlst = line.split() val = float(strlst[ self.measure_column_map[measure] ]) count = count + 1 self.measure_value_map[str(count)] = val line = self.fp.readline() return pthwy_subj, self.measure_value_map