def make_labels(self, ilines):
"""parse a list of the form
LABEL : VALUES ...
and return a LABEL list (with no trailing separator (':'))
initialize maxcounts, subjcounts here
"""
llist = []
for lind, lstr in enumerate(ilines):
# get label and value list
rv, label, vals = self.get_label_vals(lstr)
if rv < 1: continue
nvals = len(vals)
# label = self.find_parent_label(label)
if self.verb > 2: print '++ label: %s, %d val(s)' % (label, nvals)
llist.append(label)
self.maxcounts[label] = nvals
self.subjcounts[label] = 0
if not UTIL.vals_are_unique(llist):
print '** warning: labels are not unique, will use only last values'
llist = UTIL.get_unique_sublist(llist)
return 0, llist
def parse_infile_names(self):
"""try to get subject and possibly group names from infiles
fill self.snames and self.gnames, if possible
1. get SID
- if files look like out.ss_review.SID.txt, that is a good start
- else, look for varying part of filename
2. get GID
- replace SID in infile names and for varying group name
"""
rv, slist = UTIL.list_minus_pref_suf(self.infiles,'out.ss_review.','.txt')
if rv < 0: return
if rv > 0:
if self.verb > 1: print '++ trying to get SID from glob form'
slist = UTIL.list_minus_glob_form(self.infiles, strip='dir')
else:
if self.verb > 1: print "++ have SIDs from 'out.ss_reiview' form"
if len(slist) == 0:
if self.verb > 1: print "-- empty SID list"
return
# make sure names are unique and not empty
if not UTIL.vals_are_unique(slist):
if self.verb > 1: print '-- SIDs not detected: not unique'
return
minlen = min([len(ss) for ss in slist])
if minlen < 1:
if self.verb > 1: print '-- SIDs not detected: some would be empty'
return
# we have a subject list
self.snames = slist
# now go for GID, start by replacing SIDs in infiles
newfiles = [fname.replace(slist[ind], 'SUBJ') for ind, fname in
enumerate(self.infiles)]
if UTIL.vals_are_constant(newfiles):
print '-- no groups detected from filenames'
return
# okay, try to make a group list
glist = UTIL.list_minus_glob_form(newfiles)
# cannot have dirs in result
for gid in glist:
if gid.find('/') >= 0:
if self.verb>1: print '-- no GIDs, dirs vary in multiple places'
return
minlen = min([len(ss) for ss in glist])
if minlen < 1:
if self.verb > 1: print '-- GIDs not detected: some would be empty'
return
if self.verb > 1: print "++ have GIDs from infiles"
self.gnames = glist
def make_labels(self, ilines):
"""parse a list of the form
LABEL : VALUES ...
and return a LABEL list (with no trailing separator (':'))
initialize maxcounts, subjcounts here
"""
llist = []
for lind, lstr in enumerate(ilines):
# get label and value list
rv, label, vals = self.get_label_vals(lstr)
if rv < 1: continue
nvals = len(vals)
# label = self.find_parent_label(label)
if self.verb > 2: print '++ label: %s, %d val(s)' % (label, nvals)
llist.append(label)
self.maxcounts[label] = nvals
self.subjcounts[label] = 0
if not UTIL.vals_are_unique(llist):
print '** warning: labels are not unique, will use only last values'
llist = UTIL.get_unique_sublist(llist)
return 0, llist
def parse_infile_names(self):
"""try to get subject and possibly group names from infiles
fill self.snames and self.gnames, if possible
1. get SID
- if files look like out.ss_review.SID.txt, that is a good start
- else, look for varying part of filename
2. get GID
- replace SID in infile names and for varying group name
"""
rv, slist = UTIL.list_minus_pref_suf(self.infiles,'out.ss_review.','.txt')
if rv < 0: return
if rv > 0:
if self.verb > 1: print '++ trying to get SID from glob form'
slist = UTIL.list_minus_glob_form(self.infiles, strip='dir')
else:
if self.verb > 1: print "++ have SIDs from 'out.ss_reiview' form"
if len(slist) == 0:
if self.verb > 1: print "-- empty SID list"
return
# make sure names are unique and not empty
if not UTIL.vals_are_unique(slist):
if self.verb > 1: print '-- SIDs not detected: not unique'
return
minlen = min([len(ss) for ss in slist])
if minlen < 1:
if self.verb > 1: print '-- SIDs not detected: some would be empty'
return
# we have a subject list
self.snames = slist
# now go for GID, start by replacing SIDs in infiles
newfiles = [fname.replace(slist[ind], 'SUBJ') for ind, fname in
enumerate(self.infiles)]
if UTIL.vals_are_constant(newfiles):
print '-- no groups detected from filenames'
return
# okay, try to make a group list
glist = UTIL.list_minus_glob_form(newfiles)
# cannot have dirs in result
for gid in glist:
if gid.find('/') >= 0:
if self.verb>1: print '-- no GIDs, dirs vary in multiple places'
return
minlen = min([len(ss) for ss in glist])
if minlen < 1:
if self.verb > 1: print '-- GIDs not detected: some would be empty'
return
if self.verb > 1: print "++ have GIDs from infiles"
self.gnames = glist
def set_ids_from_dsets(self, prefix='', suffix='', hpad=0, tpad=0, dpre=0):
"""use the varying part of the dataset names for subject IDs
If hpad > 0 or tpad > 0, expand into the head or tail of the dsets.
If prefix or suffix is passed, apply them.
return 0 on success, 1 on error
"""
if hpad < 0 or tpad < 0:
print('** set_ids_from_dsets: will not apply negative padding')
return 1
# try filenames without paths, first
dlist = [s.dset.split('/')[-1] for s in self.subjects]
if UTIL.vals_are_constant(dlist):
print('** constant dataset names (%s)' % dlist[0])
print(' trying directories...')
dlist = [s.dset for s in self.subjects]
slist = UTIL.list_minus_glob_form(dlist,
hpad,
tpad,
keep_dent_pre=dpre)
# in the case of diretories, check for success
# (maybe we can try to skip past them, that might be okay)
for index in range(len(slist)):
if '/' in slist[index]:
posn = slist[index].rfind('/')
slist[index] = slist[index][posn + 1:]
if len(slist[index]) < 1:
print(
'** failed to extract subject IDs from directory list')
print(' (directories do not vary at single level)')
return 1
if len(slist) != len(self.subjects):
print('** failed to set SIDs from dset names\n' \
' dsets = %s\n' \
' slist = %s' % (dlist, slist))
return 1
if not UTIL.vals_are_unique(slist):
print('** cannot set IDs from dsets, labels not unique: %s' %
slist)
print('-- labels come from dsets: %s' % dlist)
return 1
for ind, subj in enumerate(self.subjects):
subj.sid = '%s%s%s' % (prefix, slist[ind], suffix)
return 0
def check_inputs(self):
"""check for required inputs: anat, epi (check existence?)"""
if self.uvars.is_empty('anat'):
self.errors.append('** unspecified anatomical dataset')
if self.uvars.is_empty('epi'):
self.errors.append('** unspecified EPI dataset')
if len(self.uvars.cost_list) < 1:
self.errors.append('** unspecified cost function(s)')
if not UTIL.vals_are_unique(self.uvars.cost_list):
self.errors.append('** cost functions are not unique')
return len(self.errors)
def check_inputs(self):
"""check for required inputs: anat, epi (check existence?)"""
if self.uvars.is_empty('anat'):
self.errors.append('** unspecified anatomical dataset')
if self.uvars.is_empty('epi'):
self.errors.append('** unspecified EPI dataset')
if len(self.uvars.cost_list) < 1:
self.errors.append('** unspecified cost function(s)')
if not UTIL.vals_are_unique(self.uvars.cost_list):
self.errors.append('** cost functions are not unique')
return len(self.errors)
def set_ids_from_dsets(self, prefix='', suffix='', hpad=0, tpad=0, dpre=0):
"""use the varying part of the dataset names for subject IDs
If hpad > 0 or tpad > 0, expand into the head or tail of the dsets.
If prefix or suffix is passed, apply them.
return 0 on success, 1 on error
"""
if hpad < 0 or tpad < 0:
print '** set_ids_from_dsets: will not apply negative padding'
return 1
# try filenames without paths, first
dlist = [s.dset.split('/')[-1] for s in self.subjects]
if UTIL.vals_are_constant(dlist):
print '** constant dataset names (%s)' % dlist[0]
print ' trying directories...'
dlist = [s.dset for s in self.subjects]
slist = UTIL.list_minus_glob_form(dlist, hpad, tpad, keep_dent_pre=dpre)
# in the case of diretories, check for success
# (maybe we can try to skip past them, that might be okay)
for index in range(len(slist)):
if '/' in slist[index]:
posn = slist[index].rfind('/')
slist[index] = slist[index][posn+1:]
if len(slist[index]) < 1:
print '** failed to extract subject IDs from directory list'
print ' (directories do not vary at single level)'
return 1
if len(slist) != len(self.subjects):
print '** failed to set SIDs from dset names\n' \
' dsets = %s\n' \
' slist = %s' % (dlist, slist)
return 1
if not UTIL.vals_are_unique(slist):
print '** cannot set IDs from dsets, labels not unique: %s' % slist
print '-- labels come from dsets: %s' % dlist
return 1
for ind, subj in enumerate(self.subjects):
subj.sid = '%s%s%s' % (prefix, slist[ind], suffix)
return 0
def restrict_ids_to_dsets(self, valid_ids=[], require=1):
"""restrict subject IDs to those in valid_ids list
require all valid_ids to exist, or fail
return 0 on success
"""
# bail if either list is empty
if len(self.subjects) == 0: return 0
if len(valid_ids) == 0: return 0
# check that valid_ids are unique
if not UTIL.vals_are_unique(valid_ids):
print('** restrict_ids: ids are not unique')
return 1
# check that all valid_ids exist, and generate new subject list
all_ids = [subj.sid for subj in self.subjects]
new_subjs = []
missing = 0
missed_id = '' # example of missing ID
for sid in valid_ids:
if sid in all_ids:
old_index = all_ids.index(sid)
new_subjs.append(self.subjects[old_index])
else:
if self.verb > 1:
print(
"** restrict_ids: cannot restrict to missing ID '%s'" %
sid)
missed_id = sid
missing += 1
if missing:
print("** restrict_ids: missing %d of %d IDs" \
% (missing,len(valid_ids)))
print(" IDs look like: %s" % ' '.join(all_ids[:3]))
print(" missing IDs look like: %s" % missed_id)
if require:
return 1
else:
print("-- restrict_ids: allowing %d missing IDs..." % missing)
# apply restricted list
self.subjects = new_subjs
return 0
def remove_ids_from_dsets(self, remove_ids=[], require=1):
"""restrict subject IDs to those not in remove_ids list
if require: require all remove_ids to exist, or fail
return 0 on success
"""
# bail if either list is empty
if len(self.subjects) == 0: return 0
if len(remove_ids) == 0: return 0
# check that remove_ids are unique
if not UTIL.vals_are_unique(remove_ids):
print('** remove_ids: ids are not unique')
return 1
# check that all remove_ids exist, and fail if not
all_ids = [subj.sid for subj in self.subjects]
missing = 0
for sid in remove_ids:
if sid not in all_ids:
if self.verb > 1:
print("** remove_ids: cannot remove missing ID '%s'" % sid)
missed_id = sid
missing += 1
if missing and (require or self.verb > 1):
print("** remove_ids: missing %d of %d IDs" \
% (missing,len(remove_ids)))
print(" IDs look like: %s" % ' '.join(all_ids[:3]))
print(" missing IDs look like: %s" % missed_id)
# if required, this is fatal
if require:
return 1
else:
print("-- remove_ids: allowing %d missing IDs..." % missing)
# generate a new subject list
new_subjs = []
for sindex, sid in enumerate(all_ids):
if sid not in remove_ids:
new_subjs.append(self.subjects[sindex])
# apply remove list
self.subjects = new_subjs
return 0
def check_inputs(self):
"""check for required inputs: anat, epi (check existence?)"""
if self.uvars.is_empty('anat'):
self.errors.append('** unspecified anatomical dataset')
if self.uvars.is_empty('epi'):
self.errors.append('** unspecified EPI dataset')
if len(self.uvars.cost_list) < 1:
self.errors.append('** unspecified cost function(s)')
if not UTIL.vals_are_unique(self.uvars.cost_list):
self.errors.append('** cost functions are not unique')
if self.uvars.align_centers == 'yes' and self.uvars.giant_move == 'no':
self.warnings.append( \
"** 'align centers' without 'giant move' is dangerous,\n" \
" consider adding 'giant move'")
return len(self.errors)
def check_inputs(self):
"""check for required inputs: anat, epi (check existence?)"""
if self.uvars.is_empty('anat'):
self.errors.append('** unspecified anatomical dataset')
if self.uvars.is_empty('epi'):
self.errors.append('** unspecified EPI dataset')
if len(self.uvars.cost_list) < 1:
self.errors.append('** unspecified cost function(s)')
if not UTIL.vals_are_unique(self.uvars.cost_list):
self.errors.append('** cost functions are not unique')
if self.uvars.align_centers == 'yes' and self.uvars.giant_move == 'no':
self.warnings.append( \
"** 'align centers' without 'giant move' is dangerous,\n" \
" consider adding 'giant move'")
return len(self.errors)
def create_review_table(self, labels=[]):
"""fill value lines
for each infile
for each label
if dict[label]: append values
append any needed empty spaces
"""
passed_labels = 0
if labels == []:
labels = self.labels
else:
passed_labels = 1
# be sure each label is in proper list
for label in labels:
if not label in self.labels:
print("** review_table: given label '%s' not in list" %
label)
return 1, []
if len(labels) < 1: return 1, []
# labels, starting with input files
RT = []
tline = []
nfiles = len(self.infiles)
# --------------------
# labels, starting with input files
# start with subject, if possible
dosubj = len(self.snames) == len(self.infiles)
dogrp = len(self.gnames) == len(self.infiles)
doinfiles = (not dogrp and not dosubj)
# if "subject ID" is label[0], and they are unique,
# then clear dosubj and doinfiles
if labels[0] == 'subject ID':
# if subject IDs are uniq and fir
label = labels[0]
sid_list = [self.ldict[ind][label] for ind in range(nfiles)]
if UTIL.vals_are_unique(sid_list):
dosubj = 0
doinfiles = 0
# allow user to force inclusion
if self.show_infiles:
doinfiles = 1
# ------------------------------------------------------------
# first 2 lines, fill header lines
# --------------------
# main header of labels
tline = []
if doinfiles: tline.append('infile')
if dogrp: tline.append('group')
if dosubj: tline.append('subject')
# add value positions
for label in labels:
nf = self.maxcounts[label] - 1
tline.append('%s' % label)
tline.extend([''] * nf)
RT.append(tline)
tline = []
# --------------------
# and value labels
if doinfiles: tline.append('value')
if dogrp: tline.append('value')
if dosubj: tline.append('value')
for label in labels:
nf = self.maxcounts[label]
for ind in range(nf):
tline.append('value_%d' % (ind + 1))
RT.append(tline)
# ------------------------------------------------------------
# add value lines, one per input file
for ind, infile in enumerate(self.infiles):
tline = [] # current line to add to table
# first is infile, if requested or nothing else to show
if doinfiles:
if infile == '-': tline.append('stdin')
else: tline.append('%s' % infile)
# then possibly group
if dogrp: tline.append('%s' % self.gnames[ind])
# then possibly subject
if dosubj: tline.append('%s' % self.snames[ind])
for label in labels:
nf = self.maxcounts[label]
try:
vals = self.ldict[ind][label]
except:
if self.verb > 2:
print('** infile %s missing key %s' % (infile, label))
vals = []
nv = len(vals)
if nv > 0: tline.extend(vals)
if nf > nv: tline.extend([''] * (nf - nv))
RT.append(tline)
return 0, RT
