def v_file_init_alt1_pre(filein,fileout,tensecode,voicecode=None):
nin = 0
nout = 0
f = codecs.open(filein,"r","utf-8")
fout = codecs.open(fileout,"w","utf-8")
roots = {}
for line in f:
line = line.rstrip('\r\n')
nin = nin + 1
if line == '':
continue
# Example. line = aMS:10P
m = re.search(r'^(.*?):([0-9]+)([AP]?)$',line)
if not m:
print "PROBLEM WITH INPUT LINE:",line
exit(1)
(root,theclass,pada) = (m.group(1),m.group(2),m.group(3))
if (theclass == '0') or (pada == ''):
print "skipping ",line
continue
dbg=False
dtype=None
if (pada == 'P'):
voice = "a"
else:
# pada == 'A'
voice = "m"
tenses=[tensecode]
for tense0 in tenses:
if tense0 == 'pop':
tense = 'opt'
else:
tense = tense0
try:
outputs = v_file_init_alt1_pre_helper(root,theclass,voice,tense,dtype=dtype,dbg=dbg)
except (NameError,MyException) as err:
print "\ncase=",nin,"line=",line
print err
continue
if isinstance(outputs,list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# adjustment for fact that Elisp uses 'pop' instead of 'opt'
if tense == 'opt':
output = output.replace(' opt ',' pop ',1) # replace only 1st occurrence
if voice == 'p':
output = re.sub(r'[0-9]+p:','p:',output)
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)",root,theclass,voice,tense
f.close()
fout.close()
def v_file_init_alt1_pre(intab, indir, outtab, outdir, n1, n2, tensecode,
voicecode):
nin = 0
nout = 0
filein = "%s/%s" % (indir, intab)
fileout = "%s/%s" % (outdir, outtab)
f = codecs.open(filein, "r", "utf-8")
fout = codecs.open(fileout, "w", "utf-8")
roots = {}
for line in f:
line = line.rstrip('\r\n')
nin = nin + 1
if line == '':
continue
if not ((n1 <= nin) and (nin <= n2)):
continue
# Example. line = (aMS 10 P <MW=aMS,17,1>)
m = re.search(r'\((.*?)<', line)
if not m:
print "PROBLEM WITH INPUT LINE:", line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words = temp.split(' ')
#print "words=",words
(root, theclass, pada) = words
dbg = False
dtype = None
if voicecode == 'p':
voice = 'p'
if root in roots:
continue # avoid duplicates
roots[root] = True
if preserve_elisp_errors:
#no good reason for this choice of Elisp code.
#In fact, 'theclass' IS used in constructing the passive base,
# for example 'aMS 10 p pre' 3s of '(aMSApyate aMSyate)'
# while aMS 0 p pre -> aMSyate
theclass = "0"
elif (pada == 'P'):
voice = "a"
else:
# pada == 'A'
voice = "m"
if tensecode == 'pre':
tenses = ['pre', 'ipf', 'ipv', 'pop'] # 4 special tenses
else:
print "UNKNOWN tensecode=", tensecode
exit(1)
for tense0 in tenses:
if tense0 == 'pop':
tense = 'opt'
else:
tense = tense0
try:
outputs = v_file_init_alt1_pre_helper(root,
theclass,
voice,
tense,
dtype=dtype,
dbg=dbg)
except (NameError, MyException) as err:
print "\ncase=", nin, "line=", line
print err
continue
if isinstance(outputs, list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# adjustment for fact that Elisp uses 'pop' instead of 'opt'
if tense == 'opt':
output = output.replace(' opt ', ' pop ',
1) # replace only 1st occurrence
if voice == 'p':
output = re.sub(r'[0-9]+p:', 'p:', output)
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)", root, theclass, voice, tense
f.close()
fout.close()
def v_file_init_alt1_fut(intab, indir, outtab, outdir, n1, n2, tensecode,
voicecode):
""" A separate routine is needed to facilitate comparison to elisp
v-file-init-alt1-fut, since there is arrangement of 'non-duplicate'
cases prior to the actual conjugation generation.
"""
nin = 0
nout = 0
filein = "%s/%s" % (indir, intab)
fileout = "%s/%s" % (outdir, outtab)
f = codecs.open(filein, "r", "utf-8")
fout = codecs.open(fileout, "w", "utf-8")
# phase 1
roots = {}
rootlist = []
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<', line)
if not m:
print "PROBLEM WITH INPUT LINE:", line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words = temp.split(' ')
#print "words=",words
(root, theclass, pada) = words
# pada is not used
if root not in roots:
roots[root] = [] # value is a list of classes
rootlist.append(root)
if theclass not in roots[root]:
roots[root].append(theclass)
f.close()
# phase 2
dbg = False
dbg1 = True
dbg1 = False
for root in rootlist:
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
multiclass = None
class10 = None
class0 = None
theclasses = roots[root]
if dbg1:
if root != 'aMh':
continue
print root, theclasses
cvs = []
for theclass in theclasses:
if theclass == '10':
class10 = theclass
elif class0 == None:
class0 = theclass
if (theclass not in cvs) and (theclass != "10"):
if multiclass:
multiclass = multiclass + "/" + theclass
else:
multiclass = theclass
cvs.append(theclass)
# reset cvs artificially
cvs = []
if class0 != None:
cvs.insert(0, [class0, multiclass])
if class10 != None:
cvs.insert(0, [class10, class10])
if dbg1:
print cvs
for (theclass, classid) in cvs:
tenses = ['fut', 'pft', 'con', 'ben']
# 'p' (passive) is same as 'm' (middle) for 'fut pft con ben'
voices = ['a', 'm']
for voice in voices:
for tense in tenses:
dtype = None
try:
# note classid is not used, except for format of output in helper
outputs = v_file_init_alt1_pre_helper(root,
classid,
voice,
tense,
dtype=dtype,
dbg=dbg)
except (NameError, MyException) as err:
print "\ncase=", nin, "line=", line
print err
continue
if isinstance(outputs, list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# adjustment for fact that Elisp uses 'pop' instead of 'opt'
if tense == 'opt':
output = output.replace(
' opt ', ' pop ',
1) # replace only 1st occurrence
if voice == 'p':
output = re.sub(r'[0-9]+p:', 'p:', output)
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)", root, theclass, voice, tense
if dbg1:
print "exiting after aMh"
exit(1)
fout.close()
def v_file_init_alt1_prf(intab, indir, outtab, outdir, n1, n2, tensecode,
voicecode):
""" A separate routine is needed to facilitate comparison to elisp
v-file-init-alt1-fut, since there is arrangement of 'non-duplicate'
cases prior to the actual conjugation generation.
Output is constructed by routine v-file-init-alt1-pre-helper, and the
this output is slightly altered before insertion into outtab (dummy class
removed). Also, since the passive depends only on the root (and not on
the class or pada), logic of this function discards duplicate records
(i.e., only one set of tables for a give root, even if that root
appears in multiple class-pada combinations in 'intab'.
The class-voice combinations to use for a given root are derived in a
complex way by function v-file-init3-prf. We mimic that function's
behavior here; the grammatical reason for this is unclear at
present (10-15-2015).
"""
nin = 0
nout = 0
filein = "%s/%s" % (indir, intab)
fileout = "%s/%s" % (outdir, outtab)
f = codecs.open(filein, "r", "utf-8")
fout = codecs.open(fileout, "w", "utf-8")
# phase 1
roots = {}
rootlist = []
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<', line)
if not m:
print "PROBLEM WITH INPUT LINE:", line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words = temp.split(' ')
#print "words=",words
(root, theclass, pada) = words
# pada is not used
if root not in roots:
roots[root] = [] # value is a list of classes
rootlist.append(root)
if theclass not in roots[root]:
roots[root].append(theclass)
f.close()
# phase 2
dbg = False
dbg1 = True
dbg1 = False
for root in rootlist:
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
multiclass = None
theclasses = roots[root]
if dbg1:
if root != 'akz':
continue
print root, theclasses
cvs = []
for theclass in theclasses:
if (theclass not in cvs) and (reduplicative_liw_P(root, theclass)):
if multiclass:
multiclass = multiclass + "/" + theclass
else:
multiclass = theclass
cvs.append(theclass)
if dbg1 and (root == 'akz'):
print "cvs=", cvs, "multiclass=", multiclass
if (len(cvs) == 0):
#print "%s has no prf. Classes=%s" %(root,theclasses)
continue
cvs = [cvs[0]] # just use first case
for theclass in cvs: # just one iteration, by previous statement
tenses = ['prf']
# 'p' (passive) is same as 'm' (middle) for 'fut pft con ben'
voices = ['a', 'm']
for voice in voices:
for tense in tenses:
dtype = None
try:
# note classid is not used, except for format of output in helper
outputs = v_file_init_alt1_pre_helper(root,
theclass,
voice,
tense,
dtype=dtype,
dbg=dbg)
except (NameError, MyException) as err:
print "\ncase=", nin, "line=", line
print err
continue
if isinstance(outputs, list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# change labeling from 'theclass' to 'multiclass'
old = " %s%s:" % (theclass, voice)
new = " %s%s:" % (multiclass, voice)
output1 = output.replace(old, new, 1)
if dbg1:
if root == 'akz':
print "CHK: output=", output, "\nold=", old, "new=", new, "\noutput1=", output1
exit(1)
fout.write("%s\n" % output1)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)", root, theclass, voice, tense
if dbg1:
print "exiting after aMh"
exit(1)
fout.close()
def v_file_init_alt1_aor(intab, indir, outtab, outdir, n1, n2, tensecode,
voicecode):
"""
Read from file like construct/dcpforms-MW.txt, and write
information for each line to output outdir/outtab
Determine aorist varieties, and for each variety, compute the
conjugation of that variety with the given root, class, voice as read from
input
Sample records are:
(Ap 5 P <MW=Ap,35737,1>)
(As 2 A <MW=As,39445,1>)
(BA 2 A <MW=vy-ati-BA,262287,1>)
NOTE: Uses aorist_varieties function. In this function, the 'pada' is
assumed to be a lower case 'p' (parasmaipada) or 'a' (atmanepada). I mention
this since we are, in particular, NOT using the Scharf convention of
'a' (for active voice, same as parasmaipada) or 'm' (for middle voice, same
as atmanepada).
"""
nin = 0
nout = 0
filein = "%s/%s" % (indir, intab)
fileout = "%s/%s" % (outdir, outtab)
f = codecs.open(filein, "r", "utf-8")
fout = codecs.open(fileout, "w", "utf-8")
# phase 1
roots = {}
rootlist = []
dbg = False
dbg1 = True
dbg1 = False
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<', line)
if not m:
print "PROBLEM WITH INPUT LINE:", line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words = temp.split(' ')
#print "words=",words
(root, theclass, pada) = words
# voice is used only in output, for comparision with Elisp output
if pada == 'P':
voice = 'a'
else:
voice = 'm'
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
upasargas = None
pada1 = pada.lower()
aorvars = aorist_varieties(root, theclass, pada1, upasargas, dbg)
if aorvars == None:
aorvars = []
for aorvar in aorvars:
tense = "aor%s" % aorvar
dtype = None
#print "v_file_init_alt1_pre_helper(%s,%s,%s,%s,%s)" %(root,theclass,voice,tense,dtype)
try:
outputs = v_file_init_alt1_pre_helper(root,
theclass,
voice,
tense,
dtype=dtype,
dbg=dbg)
except (NameError, MyException) as err:
print "\ncase=", nin, "line=", line, "tense=", tense
print err
continue
if isinstance(outputs, list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)", root, theclass, voice, tense
f.close()
fout.close()
def v_file_init_alt1_fut(intab,indir,outtab,outdir,n1,n2,tensecode,voicecode):
""" A separate routine is needed to facilitate comparison to elisp
v-file-init-alt1-fut, since there is arrangement of 'non-duplicate'
cases prior to the actual conjugation generation.
"""
nin = 0
nout = 0
filein = "%s/%s" %(indir,intab)
fileout = "%s/%s" %(outdir,outtab)
f = codecs.open(filein,"r","utf-8")
fout = codecs.open(fileout,"w","utf-8")
# phase 1
roots = {}
rootlist=[]
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<',line)
if not m:
print "PROBLEM WITH INPUT LINE:",line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words =temp.split(' ')
#print "words=",words
(root,theclass,pada) = words
# pada is not used
if root not in roots:
roots[root]=[] # value is a list of classes
rootlist.append(root)
if theclass not in roots[root]:
roots[root].append(theclass)
f.close()
# phase 2
dbg=False
dbg1=True
dbg1=False
for root in rootlist:
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
multiclass = None
class10 = None
class0 = None
theclasses = roots[root]
if dbg1:
if root != 'aMh':
continue
print root,theclasses
cvs=[]
for theclass in theclasses:
if theclass == '10':
class10 = theclass
elif class0 == None:
class0 = theclass
if (theclass not in cvs) and (theclass != "10"):
if multiclass:
multiclass = multiclass + "/" + theclass
else:
multiclass = theclass
cvs.append(theclass)
# reset cvs artificially
cvs = []
if class0 != None:
cvs.insert(0,[class0,multiclass])
if class10 != None:
cvs.insert(0,[class10,class10])
if dbg1:
print cvs
for (theclass,classid) in cvs:
tenses = ['fut','pft','con','ben']
# 'p' (passive) is same as 'm' (middle) for 'fut pft con ben'
voices = ['a','m']
for voice in voices:
for tense in tenses:
dtype=None
try:
# note classid is not used, except for format of output in helper
outputs = v_file_init_alt1_pre_helper(root,classid,voice,tense,dtype=dtype,dbg=dbg)
except (NameError,MyException) as err:
print "\ncase=",nin,"line=",line
print err
continue
if isinstance(outputs,list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# adjustment for fact that Elisp uses 'pop' instead of 'opt'
if tense == 'opt':
output = output.replace(' opt ',' pop ',1) # replace only 1st occurrence
if voice == 'p':
output = re.sub(r'[0-9]+p:','p:',output)
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)",root,theclass,voice,tense
if dbg1:
print "exiting after aMh"
exit(1)
fout.close()
def v_file_init_alt1_prf(intab,indir,outtab,outdir,n1,n2,tensecode,voicecode):
""" A separate routine is needed to facilitate comparison to elisp
v-file-init-alt1-fut, since there is arrangement of 'non-duplicate'
cases prior to the actual conjugation generation.
Output is constructed by routine v-file-init-alt1-pre-helper, and the
this output is slightly altered before insertion into outtab (dummy class
removed). Also, since the passive depends only on the root (and not on
the class or pada), logic of this function discards duplicate records
(i.e., only one set of tables for a give root, even if that root
appears in multiple class-pada combinations in 'intab'.
The class-voice combinations to use for a given root are derived in a
complex way by function v-file-init3-prf. We mimic that function's
behavior here; the grammatical reason for this is unclear at
present (10-15-2015).
"""
nin = 0
nout = 0
filein = "%s/%s" %(indir,intab)
fileout = "%s/%s" %(outdir,outtab)
f = codecs.open(filein,"r","utf-8")
fout = codecs.open(fileout,"w","utf-8")
# phase 1
roots = {}
rootlist=[]
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<',line)
if not m:
print "PROBLEM WITH INPUT LINE:",line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words =temp.split(' ')
#print "words=",words
(root,theclass,pada) = words
# pada is not used
if root not in roots:
roots[root]=[] # value is a list of classes
rootlist.append(root)
if theclass not in roots[root]:
roots[root].append(theclass)
f.close()
# phase 2
dbg=False
dbg1=True
dbg1=False
for root in rootlist:
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
multiclass = None
theclasses = roots[root]
if dbg1:
if root != 'akz':
continue
print root,theclasses
cvs=[]
for theclass in theclasses:
if (theclass not in cvs) and (reduplicative_liw_P(root,theclass)):
if multiclass:
multiclass = multiclass + "/" + theclass
else:
multiclass = theclass
cvs.append(theclass)
if dbg1 and (root == 'akz'):
print "cvs=",cvs,"multiclass=",multiclass
if (len(cvs) == 0):
#print "%s has no prf. Classes=%s" %(root,theclasses)
continue
cvs = [cvs[0]] # just use first case
for theclass in cvs: # just one iteration, by previous statement
tenses = ['prf']
# 'p' (passive) is same as 'm' (middle) for 'fut pft con ben'
voices = ['a','m']
for voice in voices:
for tense in tenses:
dtype=None
try:
# note classid is not used, except for format of output in helper
outputs = v_file_init_alt1_pre_helper(root,theclass,voice,tense,dtype=dtype,dbg=dbg)
except (NameError,MyException) as err:
print "\ncase=",nin,"line=",line
print err
continue
if isinstance(outputs,list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# change labeling from 'theclass' to 'multiclass'
old = " %s%s:" %(theclass,voice)
new = " %s%s:" %(multiclass,voice)
output1 = output.replace(old,new,1)
if dbg1:
if root == 'akz':
print "CHK: output=",output,"\nold=",old,"new=",new,"\noutput1=",output1
exit(1)
fout.write("%s\n" % output1)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)",root,theclass,voice,tense
if dbg1:
print "exiting after aMh"
exit(1)
fout.close()
def v_file_init_alt1_aor(intab,indir,outtab,outdir,n1,n2,tensecode,voicecode):
"""
Read from file like construct/dcpforms-MW.txt, and write
information for each line to output outdir/outtab
Determine aorist varieties, and for each variety, compute the
conjugation of that variety with the given root, class, voice as read from
input
Sample records are:
(Ap 5 P <MW=Ap,35737,1>)
(As 2 A <MW=As,39445,1>)
(BA 2 A <MW=vy-ati-BA,262287,1>)
NOTE: Uses aorist_varieties function. In this function, the 'pada' is
assumed to be a lower case 'p' (parasmaipada) or 'a' (atmanepada). I mention
this since we are, in particular, NOT using the Scharf convention of
'a' (for active voice, same as parasmaipada) or 'm' (for middle voice, same
as atmanepada).
"""
nin = 0
nout = 0
filein = "%s/%s" %(indir,intab)
fileout = "%s/%s" %(outdir,outtab)
f = codecs.open(filein,"r","utf-8")
fout = codecs.open(fileout,"w","utf-8")
# phase 1
roots = {}
rootlist=[]
dbg=False
dbg1=True
dbg1=False
for line in f:
line = line.rstrip('\r\n')
if line == '':
continue
m = re.search(r'\((.*?)<',line)
if not m:
print "PROBLEM WITH INPUT LINE:",line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words =temp.split(' ')
#print "words=",words
(root,theclass,pada) = words
# voice is used only in output, for comparision with Elisp output
if pada == 'P':
voice = 'a'
else:
voice = 'm'
nin = nin + 1
if not ((n1 <= nin) and (nin <= n2)):
continue
upasargas = None
pada1 = pada.lower()
aorvars=aorist_varieties(root,theclass,pada1,upasargas,dbg)
if aorvars == None:
aorvars = []
for aorvar in aorvars:
tense = "aor%s" % aorvar
dtype = None
#print "v_file_init_alt1_pre_helper(%s,%s,%s,%s,%s)" %(root,theclass,voice,tense,dtype)
try:
outputs = v_file_init_alt1_pre_helper(root,theclass,voice,tense,dtype=dtype,dbg=dbg)
except (NameError,MyException) as err:
print "\ncase=",nin,"line=",line,"tense=",tense
print err
continue
if isinstance(outputs,list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)",root,theclass,voice,tense
f.close()
fout.close()
def v_file_init_alt1_pre(intab,indir,outtab,outdir,n1,n2,tensecode,voicecode):
nin = 0
nout = 0
filein = "%s/%s" %(indir,intab)
fileout = "%s/%s" %(outdir,outtab)
f = codecs.open(filein,"r","utf-8")
fout = codecs.open(fileout,"w","utf-8")
roots = {}
for line in f:
line = line.rstrip('\r\n')
nin = nin + 1
if line == '':
continue
if not ((n1 <= nin) and (nin <= n2)):
continue
# Example. line = (aMS 10 P <MW=aMS,17,1>)
m = re.search(r'\((.*?)<',line)
if not m:
print "PROBLEM WITH INPUT LINE:",line
exit(1)
temp = m.group(1).strip()
#print "temp=",temp
words =temp.split(' ')
#print "words=",words
(root,theclass,pada) = words
dbg=False
dtype=None
if voicecode == 'p':
voice = 'p'
if root in roots:
continue # avoid duplicates
roots[root]=True
if preserve_elisp_errors:
#no good reason for this choice of Elisp code.
#In fact, 'theclass' IS used in constructing the passive base,
# for example 'aMS 10 p pre' 3s of '(aMSApyate aMSyate)'
# while aMS 0 p pre -> aMSyate
theclass = "0"
elif (pada == 'P'):
voice = "a"
else:
# pada == 'A'
voice = "m"
if tensecode == 'pre':
tenses = ['pre','ipf','ipv','pop'] # 4 special tenses
else:
print "UNKNOWN tensecode=",tensecode
exit(1)
for tense0 in tenses:
if tense0 == 'pop':
tense = 'opt'
else:
tense = tense0
try:
outputs = v_file_init_alt1_pre_helper(root,theclass,voice,tense,dtype=dtype,dbg=dbg)
except (NameError,MyException) as err:
print "\ncase=",nin,"line=",line
print err
continue
if isinstance(outputs,list):
#print "outputs is a list of length",len(outputs)
pass
elif outputs == None:
continue
else:
#print "outputs is not a list"
outputs = [outputs]
for output in outputs:
# adjustment for fact that Elisp uses 'pop' instead of 'opt'
if tense == 'opt':
output = output.replace(' opt ',' pop ',1) # replace only 1st occurrence
if voice == 'p':
output = re.sub(r'[0-9]+p:','p:',output)
fout.write("%s\n" % output)
nout = nout + 1
if len(outputs) == 0:
print "Warning(no result)",root,theclass,voice,tense
f.close()
fout.close()
