def __init__(self,ns=u'http://purl.org/linguistics/data/'):
#assign user defined namespace, or simply use default
self.ns=ns
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
self.converter=CharConverter('praat','uni')
#the universe to populate
#self.universe=LingUniverse()
#self.universe=linguniv
#will be used for interactive tier description
self.user_defined_types=[]
#for hashing time points, for speed
self.time_points={}
#init basic data members
self.segments=[]
self.morphemes=[]
self.translations=[]
self.notes=[]
def __init__(self):
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
#assume file is in XSAMPA, for now
self.converter = CharConverter('xsampa', 'uni')
#the universe to populate
#self.universe=LingUniverse()
#will be used for interactive tier description
self.user_defined_types = []
#for hashing time points
self.time_points = {}
def __init__(self,ns=u'http://purl.org/linguistics/data/'):
#assign user defined namespace, or simply use default
self.ns=ns
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
self.converter=CharConverter('praat','uni')
#the universe to populate
#self.universe=LingUniverse()
#self.universe=linguniv
#will be used for interactive tier description
self.user_defined_types=[]
#for hashing time points, for speed
self.time_points={}
#init basic data members
self.segments=[]
self.morphemes=[]
self.translations=[]
self.notes=[]
def toLatex(self,out_file):
#r=TableReader()
converter=CharConverter('ipa','tipa')
#map=ipa_tipa.getMap(0,4)
latex_header=u'%This document was autogenerated using from a Python script.\n\documentclass[letter,12pt]{article}\n\\usepackage{linguex}\n\n\\usepackage[tone,extra]{tipa}\n\n\\begin{document}\n\n'
outfile=codecs.open(out_file,'w','utf-8')
outfile.write(latex_header)
for i in self.igts:
outfile.write(u'\exg.')
outfile.write(u'\t')
p=printTipaList(i.phons)
#print p
#convert from ipa to tipa
#p=stringReplace(map,p)
p=converter.convert(p)
#print p
p=tipaClean(p)
#print p+'\n'
outfile.write(p+u'\\\\\n')
#take care of ctrl char's in latex
g=u'\t'+printList(i.glosses)+u'\\\\\n'
g=g.replace('_','\_')
outfile.write(g)
outfile.write(u'\t'+i.translation+u'\n\n')
outfile.write(u'\n\n\end{document}')
outfile.close()
def __init__(self):
# one converter per input file is read
# (maybe change this to global if multiple files are to be read in a batch?
# assume file is in XSAMPA, for now
self.converter = CharConverter("xsampa", "uni")
# the universe to populate
# self.universe=LingUniverse()
# will be used for interactive tier description
self.user_defined_types = []
# for hashing time points
self.time_points = {}
class PraatReader():
"""
A class for parsing and processing Praat TextGrid files
"""
#from Termset
# def __init__(self,ns=u'http://purl.org/linguistics/data/termset/'):
# """
# Init the termset and set its namespace
#
# @type ns: unicode
# @param ns: a unicode obj representing a namespace
# """
# self.namespace = Namespace(ns)
#
# self.ontology = Ontology()
def __init__(self,ns=u'http://purl.org/linguistics/data/'):
#assign user defined namespace, or simply use default
self.ns=ns
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
self.converter=CharConverter('praat','uni')
#the universe to populate
#self.universe=LingUniverse()
#self.universe=linguniv
#will be used for interactive tier description
self.user_defined_types=[]
#for hashing time points, for speed
self.time_points={}
#init basic data members
self.segments=[]
self.morphemes=[]
self.translations=[]
self.notes=[]
#def setNS(self,ns):
# """
# Set the namespace for the ontology.
# @type ns: str
# @param ns: a namespace string
# @rtype:
# @return:
# """
def setURI(self,uri):
"""
Ses the URI for the ontology.
@type uri: str
@param uri: a string representing a URI
"""
self.uri=uri
self.ontology=Ontology(uri)
def readPraat(self,filename):
"""
Reads and parses a Praat TextGrid file.
"""
try:
self.input_file=open(filename, 'r')
print('Trying to open '+filename+'...')
except IOError:
print error
print('Make sure the path is correct.')
print 'Reading Praat file...'
#test input file for proper type
try:
file_lines=self.input_file.readlines(8)
if file_lines[1]!='Object class = "TextGrid"\n':
raise IOError()
print 'This Praat file is good.'
#go to beginning of file
self.input_file.seek(0,0)
#temp variables
current_type=''
#get user defined ling types
for line in self.input_file:
#find the name of the tier
if line.find('name = "')!=-1:
self.user_defined_types.append(findQuoted(line))
self.input_file.seek(0,0)
#BEGIN MAIN ALGORITHM
print 'Processing contents...'
#process line by line
for line in self.input_file:
#reset tier type as different types are encountered
if line.find('IntervalTier')!=-1:
current_type='Interval'
elif line.find('PointTier')!=-1:
current_type='Point'
else: current_type=current_type
#find the name of the tier
if line.find('name = "')!=-1:
dstr_role = findQuoted(line)
#for handling individual intervals
if line.find('intervals [')!=-1:
try:
time_interval=[]
data=''
time_interval.append(findNum(self.input_file.next()))
time_interval.append(findNum(self.input_file.next()))
data=findQuoted(self.input_file.next())
#only build an interval if the text has content
if (data.isspace()!=1 and data!=''):
self.handleData(data,dstr_role,time_interval)
except StopIteration:
pass
except IOError:
print "Input file is not well formed or not of type TextGrid."
print file_lines[1]
def getUniverse(self):
"""
A method for returning the ling. universe
"""
return self.universe
def handleData(self,data,dstr_role,time_interval):
"""
Decide on the linguistic unit to instantiate
data string repr. of linguistic form or grammar unit label
dstr_role how the data is used in the Praat file (gloss, translation, etc)
time_interval time alignment
Instantiates units and add them to ling universe
"""
start=time_interval[0]
end=time_interval[1]
#build data obj's and add to universe
if dstr_role=='segment':
data=self.converter.convert(data)
#
#deal w. linguniv
#self.universe.addData(FormUnit('Munken',data,start,end))
#self.segments.append(FormUnit('Munken',data,start,end))
pass
#print data, 'is a segment'
elif dstr_role=='morpheme':
#self.universe.addData(Morpheme('Munken',data,start,end))
#elif dstr_role=='Checked tone':
#self.universe.addData(FormUnit(data)????)
#self.morphemes.append(Morpheme('Munken',data,start,end))
print data, ' is a Morpheme'
elif dstr_role=='translation':
#self.universe.addData(FormUnit('English',data,start,end))
#self.translations.append(FormUnit('English',data,start,end))
print data, ' is a translation'
#bug here: notes show up as forms of morphemes
#elif dstr_role=='notes':
# self.universe.addData(FormUnit('English',data,start,end))
def printIGT(self,out_file):
"""
build IGT objects based on number of translations
(since trans are common to the set of morphemes and segments );
assumes intervals are ordered (segment, syllable, morpheme, trans, notes)
"""
igts=[]
for t in self.translations:
temp_morphemes=[]
temp_segments=[]
#loop over morphemes to find those that are assoc'd with a translation
for g in self.morphemes:
if g.start>=t.start and g.end<=t.end:
temp_morphemes.append(g.label)
#loop over segment transcriptions to find those associated with each morpheme
for p in self.segments:
if p.start==g.start and p.end==g.end:
#print p.text
temp_segments.append(p.getString())
igts.append(IGT(temp_segments, temp_morphemes, t.segments))
c=IGTCollection(igts)
print 'there are', len(self.segments),'segment transcriptions'
#print 'there are', len(self.syllables),'syllables'
print 'there are', len(self.morphemes),'morphemes'
print 'there are ', len(self.translations),' translations'
print 'there are ', len(self.notes),' notes'
for t in self.translations:
print t.segments+' ',
#c.toStndOut()
c.toLatex(out_file)
def parseBibtex(self, bibfile):
"""
Parse a bibtext file and extract entries
@type bibfile: string
@param bibfile: the filename to be read
"""
f = open(bibfile, 'r')
file_lines = f.readlines()
#use the character converter from eltk.utils
converter = CharConverter('latex', 'uni')
#loop over input file
for line in file_lines:
#replace latex special character codes w. unicode
#
#only works for this latex style: \x{y}
#
#this won't work: \{xy}
#
line = converter.convert(line)
#if 'author' in line: print line
#get rid of leading and trailing spaces
line = line.strip()
#skip over blank lines
if line == '':
continue
#begin entry
if line[0] == '@':
if '@preamble' in line: pass
#get the name for the entry
entry_name = line[line.find("{") + 1:line.find(",")]
#create a new entry
self.entries[entry_name] = Entry()
#assign id using original bib files entry name
self.entries[entry_name].id = entry_name
#find the entry's type
type = line[1:line.find('{')].lower()
#normalize the type to camel-case, eg 'InCollection', not 'Incollection'
type = type.capitalize()
type = type.replace('Inproceedings', 'InProceedings')
type = type.replace('Incollection', 'InCollection')
type = type.replace('Inbook', 'InBook')
type = type.replace('Phdthesis', 'PhdThesis')
type = type.replace('Mastersthesis', 'MastersThesis')
type = type.replace('Techreport', 'TechReport')
self.entries[entry_name].type = type
#if field uses " and also contains internal {}'s
if line.find('\"') < line.find('{') and line.find('{') != -1:
#delete internal braces
line.replace('{', '')
line.replace('}', '')
#normalize in favor of {}'s, not quotes
#but leave internal quotes
line = line.replace('\"', '{', 1)
if line.count('\"') == 1:
line = line.replace('\"', '}')
else:
line = rreplace(line, '\"', '}')
#process fields, line by line
#non-integer fields
if 'author' in line:
#pick out string containing authors
authors_string = line[line.find('{') + 1:line.rfind('}')]
#reformat the string and add to object
self.entries[entry_name].authors = findAuthors(authors_string)
#set authors_string in Entry
self.entries[entry_name].authors_string = authors_string
if 'editor' in line:
#pick out string containing editors
editors_string = line[line.find('{') + 1:line.rfind('}')]
#reformat the string and add to object (use same as authors)
self.entries[entry_name].editors = findAuthors(editors_string)
#set editors_string in Entry
self.entries[entry_name].editors_string = editors_string
if 'title' in line:
#don't use title case due to bug in title() function (ie if there's a non-ascii char, then the next char get capitalized regardless of whether it's at the beginning of a word.)
self.entries[entry_name].title = line[line.find('{') +
1:line.rfind('}')]
#.title()
if 'booktitle' in line:
#don't use title case as per note above
self.entries[entry_name].booktitle = line[
line.find('{') + 1:line.rfind('}')] #.title()
if 'journal' in line:
self.entries[entry_name].journal = line[line.find('{') +
1:line.rfind('}')]
if 'pages' in line:
self.entries[entry_name].pages = line[line.find('{') +
1:line.rfind('}')]
if 'publisher' in line:
self.entries[entry_name].publisher = line[line.find('{') +
1:line.rfind('}')]
if 'address' in line:
self.entries[entry_name].address = line[line.find('{') +
1:line.rfind('}')]
if 'location' in line:
self.entries[entry_name].location = line[line.find('{') +
1:line.rfind('}')]
if 'school' in line:
self.entries[entry_name].school = line[line.find('{') +
1:line.rfind('}')]
if 'organization' in line:
self.entries[entry_name].organization = line[line.find('{') +
1:line.rfind('}')]
if 'institution' in line:
self.entries[entry_name].institution = line[line.find('{') +
1:line.rfind('}')]
if 'series' in line:
self.entries[entry_name].series = line[line.find('{') +
1:line.rfind('}')]
if 'edition' in line:
self.entries[entry_name].edition = line[line.find('{') +
1:line.rfind('}')]
if 'howpublished' in line:
self.entries[entry_name].howpublished = line[line.find('{') +
1:line.rfind('}')]
if 'month' in line:
self.entries[entry_name].month = line[line.find('{') +
1:line.rfind('}')]
if 'note' in line:
self.entries[entry_name].note = line[line.find('{') +
1:line.rfind('}')]
#various identifiers
if 'doi' in line:
self.entries[entry_name].doi = line[line.find('{') +
1:line.rfind('}')]
if 'isbn' in line:
self.entries[entry_name].isbn = line[line.find('{') +
1:line.rfind('}')]
if 'issn' in line:
self.entries[entry_name].issn = line[line.find('{') +
1:line.rfind('}')]
if 'lccn' in line:
self.entries[entry_name].lccn = line[line.find('{') +
1:line.rfind('}')]
#integer fields
if 'year' in line:
self.entries[entry_name].year = findInt(line)
if 'volume' in line:
self.entries[entry_name].volume = findInt(line)
if 'number' in line:
self.entries[entry_name].number = findInt(line)
if 'chapter' in line:
self.entries[entry_name].chapter = findInt(line)
print 'Found ' + str(len(self.entries)) + ' entries'
#post processing to fix entries w. editor and no author and
#to replace id with the convention form makeID()
#loop through entries
keys = self.entries.keys()
for k in keys:
new_id = ''
#make the new ID
if self.entries[k].authors_string == '':
new_id = makeID(self.entries[k].editors_string,
self.entries[k].year)
else:
new_id = makeID(self.entries[k].authors_string,
self.entries[k].year)
self.entries[new_id] = self.entries.pop(k)
self.entries[new_id].id = new_id
class ElanReader(object):
"""
A class for reading Elan eaf files
"""
def __init__(self):
# one converter per input file is read
# (maybe change this to global if multiple files are to be read in a batch?
# assume file is in XSAMPA, for now
self.converter = CharConverter("xsampa", "uni")
# the universe to populate
# self.universe=LingUniverse()
# will be used for interactive tier description
self.user_defined_types = []
# for hashing time points
self.time_points = {}
def setNS(self, ns):
"""
Set the namespace for the ontology.
@type ns: str
@param ns: a namespace string
@rtype:
@return:
"""
self.ns = ns
def setOntologyURI(self, uri):
"""
Ses the URI for the ontology.
@type uri: str
@param uri: a string representing a URI
"""
self.ontology = Ontology(uri)
def readElan(self, f):
"""Reads and parses an Elan file and populates the ling universe
@type f:string
@param f:an eaf file name
"""
print "Reading Elan file..."
self.file_in = open(f, "r")
# return IOError if root of eaf is not ANNOTATION_DOCUMENT
try:
file_lines = self.file_in.readlines(2)
if file_lines[1][1:20] != "ANNOTATION_DOCUMENT":
raise IOError
print "This Elan file is good."
except IOError:
print "Input file is not well formed or not of type 'eaf'."
# go to beginning of file
self.file_in.seek(0, 0)
# begin xml processing
self.dom = parse(self.file_in)
# store time points in a dict. for later use
for t in self.dom.getElementsByTagName("TIME_SLOT"):
self.time_points[t.getAttribute("TIME_SLOT_ID")] = t.getAttribute("TIME_VALUE")
# store alignable elements in a dict. for later use
self.alignable_elems = self.dom.getElementsByTagName("ALIGNABLE_ANNOTATION")
# BEGIN MAIN ALGORITHM:
print "Processing contents..."
# process tier by tier, based on whether they contain alignable or reference annotations
tier_elems = self.dom.getElementsByTagName("TIER")
# get user defined ling types
for tier in tier_elems:
self.user_defined_types.append(tier.getAttribute("LINGUISTIC_TYPE_REF"))
# print self.user_defined_types
for tier in tier_elems:
# dstr_role='data structure role'
# used to decide on what type of linguistic unit to instantiate
dstr_role = tier.getAttribute("LINGUISTIC_TYPE_REF")
alignable_elems = tier.getElementsByTagName("ALIGNABLE_ANNOTATION")
if len(alignable_elems) > 0:
self.handleAnnotation(alignable_elems, dstr_role)
ref_elems = tier.getElementsByTagName("REF_ANNOTATION")
if len(ref_elems) > 0:
self.handleAnnotation(ref_elems, dstr_role)
# BEGIN OTHER CLASS METHODS
def handleAnnotation(self, elems, dstr_role):
"""
Process ALIGNABLE_ANNOTATION and REF_ANNOTATION elements
@type elems:
@param elems: the elements to start from
@type dstr_role:
@param dstr_role: the element type
"""
# print 'There are ',len(elems),' annotation elements.'
for e in elems:
annot_val_elems = e.getElementsByTagName("ANNOTATION_VALUE")
if annot_val_elems > 0:
self.handleAnnotationValue(annot_val_elems, dstr_role, self.findTimeInterval(e))
def handleAnnotationValue(self, annot_val_elems, dstr_role, time_interval):
""" Process CDATA associated with ANNOTATION_VALUE elements.
annot_val_elems the elements to be processed
dstr_role element type to be passed on
time_interval time interval list to be passed on
Calls handleData(...)
"""
for v in annot_val_elems:
for c in v.childNodes:
if c.nodeType == 3:
self.handleData(c.data, dstr_role, time_interval)
def findTimeInterval(self, elem):
"""Build time interval list
elem the element to start from
Return a list containing the start and end times
"""
if elem.tagName == "ALIGNABLE_ANNOTATION":
start = elem.getAttribute("TIME_SLOT_REF1")
end = elem.getAttribute("TIME_SLOT_REF2")
# look up time refs in hash
return [self.time_points[start], self.time_points[end]]
elif elem.tagName == "REF_ANNOTATION":
ref = elem.getAttribute("ANNOTATION_REF")
for e in self.alignable_elems:
if e.getAttribute("ANNOTATION_ID") == ref:
# recursive call
return self.findTimeInterval(e)
def handleData(self, data, dstr_role, time_interval):
"""Decide on which linguistic units to instantiate.
data string repr. of linguistic form or grammar unit label
dstr_role how the data is used in the Elan file (gloss, translation, etc)
time_interval time alignment
Instantiates units and adds to ling. universe
"""
start = float(time_interval[0])
end = float(time_interval[1])
# build data obj's and add to universe
if dstr_role == "Sentence-level Transcription":
# convert unicode to string (nec. for character converter)
if type(data) == unicode:
data = str(data)
data = self.converter.convert(data)
print data, " is a Clause or Phrase"
# generalize this later
data = data.split()
for d in data:
print d, " is a SyntacticWord"
w = gold.SyntacticWord(self.ns + makeID(d), [])
# self.universe.addData(FormUnit('Koshin',d,start,end))
start = start + 0.00001
# self.universe.addData(FormUnit(data,start,end))
elif dstr_role == "Morpheme":
# generalize this later
###############################################
print data, "###data"
data = data.split()
for d in data:
print d, "####d"
# self.universe.addData(Morpheme('Koshin',d,start,end))
start = start + 0.00001
################################################
# self.universe.addData(Morpheme(data,start,end))
# elif dstr_role=='Checked tone':
# self.universe.addData(FormUnit(data,start,end))
elif dstr_role == "Sentence-level Translation":
# print 'trans'
pass
# self.universe.addData(FormUnit('English',data,start,end))
# elif dstr_role=='Notes':
# self.universe.addData(FormUnit(data,start,end)
def writeData(self, outfile):
self.ontology.save(outfile)
def toLatex(self, out_file):
"""
Outputs a tex file, converting from IPA to TIPA for each entry.
out_file The name of the tex output file
This requires the following tex ackages: tipa, linguex, supertabular
"""
print 'Printing to LeTex file...'
converter = CharConverter('ipa', 'tipa')
latex_header = u'%This document was autogenerated using from a Python script.\n\documentclass[letter,12pt]{article}\n\\usepackage{linguex}\n\n\\usepackage[tone,extra]{tipa}\n\\usepackage{supertabular}\n\n\\begin{document}\n\n'
outfile = codecs.open(out_file, 'w', 'utf-8')
outfile.write(latex_header)
############begin first section of bilingual dictionary###################3
outfile.write(u'\\section*{' + self.obj_lang + u'--' + self.meta_lang +
u' Dictionary}\n\n')
outfile.write(u'\\begin{supertabular}{lll}\n\n')
keys = self.entries_1.keys()
keys.sort()
for k in keys:
#headword
head_word = converter.convert(k)
head_word = tipaClean(head_word)
#alt forms
keys2 = self.entries_1[k].alt_forms.keys()
alts = '('
for k2 in keys2:
alts = alts + k2 + ': '
for i in self.entries_1[k].alt_forms[k2]:
if i is not None:
#################START HERE with tipa issue
#print i+' ',
i = converter.convert(i)
i = tipaClean(i)
#print i
alts = alts + '\\textipa{' + i + '}, '
alts = alts + ')'
alts = alts.replace(', )', ')')
if alts == '()': alts = ''
#translation
translation = ''
for t in self.entries_1[k].translation:
translation = translation + t + ', '
translation = translation[:len(translation) - 2]
translation = translation.replace('_', '\_')
outfile.write(u'\\textbf{\\textipa{' + head_word + u'}}&' + alts +
u'&' + translation + u'\\\\\\\\\n')
outfile.write(u'\end{supertabular}\n\n')
############begin next section of bilingual dictionary###################3
outfile.write(u'\\section*{' + self.meta_lang + u'--' + self.obj_lang +
u' Dictionary}\n\n')
outfile.write(u'\\begin{supertabular}{lll}\n\n')
keys_2 = self.entries_2.keys()
keys_2.sort()
for k in keys_2:
#head word
head_word = k
head_word = head_word.replace('_', '\_')
#unpack translations
trans = ''
for t in self.entries_2[k].translation:
if t is not None:
t = converter.convert(t)
t = tipaClean(t)
trans = trans + t + ', '
else:
trans = trans
trans = trans[:len(trans) - 2]
outfile.write(u'\\textbf{' + head_word + u'}&' + '' +
u'&\\textipa{' + trans + u'}\\\\\\\\\n')
outfile.write(u'\end{supertabular}\n\n\end{document}')
outfile.close()
def parseBibtex(self,bibfile):
"""
Parse a bibtext file and extract entries
@type bibfile: string
@param bibfile: the filename to be read
"""
f=open(bibfile,'r')
file_lines=f.readlines()
#use the character converter from eltk.utils
converter=CharConverter('latex','uni')
#loop over input file
for line in file_lines:
#replace latex special character codes w. unicode
#
#only works for this latex style: \x{y}
#
#this won't work: \{xy}
#
line=converter.convert(line)
#if 'author' in line: print line
#get rid of leading and trailing spaces
line=line.strip()
#skip over blank lines
if line=='':
continue
#begin entry
if line[0]=='@':
if '@preamble' in line: pass
#get the name for the entry
entry_name=line[line.find("{")+1:line.find(",")]
#create a new entry
self.entries[entry_name]=Entry()
#assign id using original bib files entry name
self.entries[entry_name].id=entry_name
#find the entry's type
type=line[1:line.find('{')].lower()
#normalize the type to camel-case, eg 'InCollection', not 'Incollection'
type=type.capitalize()
type=type.replace('Inproceedings','InProceedings')
type=type.replace('Incollection','InCollection')
type=type.replace('Inbook','InBook')
type=type.replace('Phdthesis','PhdThesis')
type=type.replace('Mastersthesis','MastersThesis')
type=type.replace('Techreport','TechReport')
self.entries[entry_name].type=type
#if field uses " and also contains internal {}'s
if line.find('\"')<line.find('{') and line.find('{')!=-1:
#delete internal braces
line.replace('{','')
line.replace('}','')
#normalize in favor of {}'s, not quotes
#but leave internal quotes
line=line.replace('\"','{',1)
if line.count('\"')==1:
line=line.replace('\"','}')
else:
line=rreplace(line,'\"','}')
#process fields, line by line
#non-integer fields
if 'author' in line:
#pick out string containing authors
authors_string=line[line.find('{')+1:line.rfind('}')]
#reformat the string and add to object
self.entries[entry_name].authors=findAuthors(authors_string)
#set authors_string in Entry
self.entries[entry_name].authors_string=authors_string
if 'editor' in line:
#pick out string containing editors
editors_string=line[line.find('{')+1:line.rfind('}')]
#reformat the string and add to object (use same as authors)
self.entries[entry_name].editors=findAuthors(editors_string)
#set editors_string in Entry
self.entries[entry_name].editors_string=editors_string
if 'title' in line:
#don't use title case due to bug in title() function (ie if there's a non-ascii char, then the next char get capitalized regardless of whether it's at the beginning of a word.)
self.entries[entry_name].title=line[line.find('{')+1:line.rfind('}')]
#.title()
if 'booktitle' in line:
#don't use title case as per note above
self.entries[entry_name].booktitle=line[line.find('{')+1:line.rfind('}')]#.title()
if 'journal' in line:
self.entries[entry_name].journal=line[line.find('{')+1:line.rfind('}')]
if 'pages' in line:
self.entries[entry_name].pages=line[line.find('{')+1:line.rfind('}')]
if 'publisher' in line:
self.entries[entry_name].publisher=line[line.find('{')+1:line.rfind('}')]
if 'address' in line:
self.entries[entry_name].address=line[line.find('{')+1:line.rfind('}')]
if 'location' in line:
self.entries[entry_name].location=line[line.find('{')+1:line.rfind('}')]
if 'school' in line:
self.entries[entry_name].school=line[line.find('{')+1:line.rfind('}')]
if 'organization' in line:
self.entries[entry_name].organization=line[line.find('{')+1:line.rfind('}')]
if 'institution' in line:
self.entries[entry_name].institution=line[line.find('{')+1:line.rfind('}')]
if 'series' in line:
self.entries[entry_name].series=line[line.find('{')+1:line.rfind('}')]
if 'edition' in line:
self.entries[entry_name].edition=line[line.find('{')+1:line.rfind('}')]
if 'howpublished' in line:
self.entries[entry_name].howpublished=line[line.find('{')+1:line.rfind('}')]
if 'month' in line:
self.entries[entry_name].month=line[line.find('{')+1:line.rfind('}')]
if 'note' in line:
self.entries[entry_name].note=line[line.find('{')+1:line.rfind('}')]
#various identifiers
if 'doi' in line:
self.entries[entry_name].doi=line[line.find('{')+1:line.rfind('}')]
if 'isbn' in line:
self.entries[entry_name].isbn=line[line.find('{')+1:line.rfind('}')]
if 'issn' in line:
self.entries[entry_name].issn=line[line.find('{')+1:line.rfind('}')]
if 'lccn' in line:
self.entries[entry_name].lccn=line[line.find('{')+1:line.rfind('}')]
#integer fields
if 'year' in line:
self.entries[entry_name].year=findInt(line)
if 'volume' in line:
self.entries[entry_name].volume=findInt(line)
if 'number' in line:
self.entries[entry_name].number=findInt(line)
if 'chapter' in line:
self.entries[entry_name].chapter=findInt(line)
print 'Found '+str(len(self.entries))+' entries'
#post processing to fix entries w. editor and no author and
#to replace id with the convention form makeID()
#loop through entries
keys=self.entries.keys()
for k in keys:
new_id=''
#make the new ID
if self.entries[k].authors_string=='':
new_id=makeID(self.entries[k].editors_string,self.entries[k].year)
else:
new_id=makeID(self.entries[k].authors_string,self.entries[k].year)
self.entries[new_id]=self.entries.pop(k)
self.entries[new_id].id=new_id
from itertools import combinations
from random import randint
derivWeight = .90
sampleSize = 100
minmean = 10
iters = 100
retention = {}
xmldoc = minidom.parse('./swadesh-retent.xml')
itemlist = xmldoc.getElementsByTagName('word')
for s in itemlist:
retention[s.attributes['sem'].value] = float(float(s.attributes['retention'].value)/float(100))
latex_converter=CharConverter('uni','tipa')
lgg = {}
xml_list = listdir('./LanguageXML')
widgets = ['Reading Language XML: ', Percentage(), ' ', Bar(marker=RotatingMarker()),
' ', ETA()]
pbar = ProgressBar(widgets=widgets,maxval=len(xml_list)).start()
for j in range(len(xml_list)):
name = xml_list[j]
lang = name.split('.')[0]
lgg[lang] = {}
xmlname = './LanguageXML/' + name
xmldoc = minidom.parse(xmlname)
itemlist = xmldoc.getElementsByTagName('word')
for s in itemlist :
mean = s.attributes['sem'].value
# -*- coding: UTF-8 -*-
from eltk.utils.CharConverter import *
"""
This script demostrates the use of CharConverter
"""
if __name__ == '__main__':
#latex to unicode test
c0 = CharConverter('latex', 'uni')
print c0.convert('author = {S\\\'{a}ndor Hervey},')
#unicode to latex test
c1 = CharConverter('uni', 'latex')
print c1.convert('ä')
#praat to uni test
c2 = CharConverter('praat', 'uni')
print c2.convert('\\t.o')
print 'If you do not see IPA characters, you may not have the required fonts installed.'
def toLatex(self,out_file):
"""
Outputs a tex file, converting from IPA to TIPA for each entry.
out_file The name of the tex output file
This requires the following tex ackages: tipa, linguex, supertabular
"""
print 'Printing to LeTex file...'
converter=CharConverter('ipa','tipa')
latex_header=u'%This document was autogenerated using from a Python script.\n\documentclass[letter,12pt]{article}\n\\usepackage{linguex}\n\n\\usepackage[tone,extra]{tipa}\n\\usepackage{supertabular}\n\n\\begin{document}\n\n'
outfile=codecs.open(out_file,'w','utf-8')
outfile.write(latex_header)
############begin first section of bilingual dictionary###################3
outfile.write(u'\\section*{'+self.obj_lang+u'--'+self.meta_lang+u' Dictionary}\n\n')
outfile.write(u'\\begin{supertabular}{lll}\n\n')
keys=self.entries_1.keys()
keys.sort()
for k in keys:
#headword
head_word=converter.convert(k)
head_word=tipaClean(head_word)
#alt forms
keys2=self.entries_1[k].alt_forms.keys()
alts='('
for k2 in keys2:
alts=alts+k2+': '
for i in self.entries_1[k].alt_forms[k2]:
if i is not None:
#################START HERE with tipa issue
#print i+' ',
i=converter.convert(i)
i=tipaClean(i)
#print i
alts=alts+'\\textipa{'+i+'}, '
alts=alts+')'
alts=alts.replace(', )',')')
if alts=='()': alts=''
#translation
translation=''
for t in self.entries_1[k].translation:
translation=translation+t+', '
translation=translation[:len(translation)-2]
translation=translation.replace('_','\_')
outfile.write(u'\\textbf{\\textipa{'+head_word+u'}}&'+alts+u'&'+translation+u'\\\\\\\\\n')
outfile.write(u'\end{supertabular}\n\n')
############begin next section of bilingual dictionary###################3
outfile.write(u'\\section*{'+self.meta_lang+u'--'+self.obj_lang+u' Dictionary}\n\n')
outfile.write(u'\\begin{supertabular}{lll}\n\n')
keys_2=self.entries_2.keys()
keys_2.sort()
for k in keys_2:
#head word
head_word=k
head_word=head_word.replace('_','\_')
#unpack translations
trans=''
for t in self.entries_2[k].translation:
if t is not None:
t=converter.convert(t)
t=tipaClean(t)
trans=trans+t+', '
else: trans=trans
trans=trans[:len(trans)-2]
outfile.write(u'\\textbf{'+head_word+u'}&'+''+u'&\\textipa{'+trans+u'}\\\\\\\\\n')
outfile.write(u'\end{supertabular}\n\n\end{document}')
outfile.close()
# -*- coding: UTF-8 -*-
from eltk.utils.CharConverter import *
"""
This script demostrates the use of CharConverter
"""
if __name__=='__main__':
#latex to unicode test
c0=CharConverter('latex','uni')
print c0.convert('author = {S\\\'{a}ndor Hervey},')
#unicode to latex test
c1=CharConverter('uni','latex')
print c1.convert('ä')
#praat to uni test
c2=CharConverter('praat','uni')
print c2.convert('\\t.o')
print 'If you do not see IPA characters, you may not have the required fonts installed.'
class PraatReader():
"""
A class for parsing and processing Praat TextGrid files
"""
#from Termset
# def __init__(self,ns=u'http://purl.org/linguistics/data/termset/'):
# """
# Init the termset and set its namespace
#
# @type ns: unicode
# @param ns: a unicode obj representing a namespace
# """
# self.namespace = Namespace(ns)
#
# self.ontology = Ontology()
def __init__(self,ns=u'http://purl.org/linguistics/data/'):
#assign user defined namespace, or simply use default
self.ns=ns
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
self.converter=CharConverter('praat','uni')
#the universe to populate
#self.universe=LingUniverse()
#self.universe=linguniv
#will be used for interactive tier description
self.user_defined_types=[]
#for hashing time points, for speed
self.time_points={}
#init basic data members
self.segments=[]
self.morphemes=[]
self.translations=[]
self.notes=[]
#def setNS(self,ns):
# """
# Set the namespace for the ontology.
# @type ns: str
# @param ns: a namespace string
# @rtype:
# @return:
# """
def setURI(self,uri):
"""
Ses the URI for the ontology.
@type uri: str
@param uri: a string representing a URI
"""
self.uri=uri
self.ontology=Ontology(uri)
def readPraat(self,filename):
"""
Reads and parses a Praat TextGrid file.
"""
try:
self.input_file=open(filename, 'r')
print('Trying to open '+filename+'...')
except IOError:
print error
print('Make sure the path is correct.')
print 'Reading Praat file...'
#test input file for proper type
try:
file_lines=self.input_file.readlines(8)
if file_lines[1]!='Object class = "TextGrid"\n':
raise IOError()
print 'This Praat file is good.'
#go to beginning of file
self.input_file.seek(0,0)
#temp variables
current_type=''
#get user defined ling types
for line in self.input_file:
#find the name of the tier
if line.find('name = "')!=-1:
self.user_defined_types.append(findQuoted(line))
self.input_file.seek(0,0)
#BEGIN MAIN ALGORITHM
print 'Processing contents...'
#process line by line
for line in self.input_file:
#reset tier type as different types are encountered
if line.find('IntervalTier')!=-1:
current_type='Interval'
elif line.find('PointTier')!=-1:
current_type='Point'
else: current_type=current_type
#find the name of the tier
if line.find('name = "')!=-1:
dstr_role = findQuoted(line)
#for handling individual intervals
if line.find('intervals [')!=-1:
try:
time_interval=[]
data=''
time_interval.append(findNum(self.input_file.next()))
time_interval.append(findNum(self.input_file.next()))
data=findQuoted(self.input_file.next())
#only build an interval if the text has content
if (data.isspace()!=1 and data!=''):
self.handleData(data,dstr_role,time_interval)
except StopIteration:
pass
except IOError:
print "Input file is not well formed or not of type TextGrid."
print file_lines[1]
def getUniverse(self):
"""
A method for returning the ling. universe
"""
return self.universe
def handleData(self,data,dstr_role,time_interval):
"""
Decide on the linguistic unit to instantiate
data string repr. of linguistic form or grammar unit label
dstr_role how the data is used in the Praat file (gloss, translation, etc)
time_interval time alignment
Instantiates units and add them to ling universe
"""
start=time_interval[0]
end=time_interval[1]
#build data obj's and add to universe
if dstr_role=='segment':
data=self.converter.convert(data)
#
#deal w. linguniv
#self.universe.addData(FormUnit('Munken',data,start,end))
#self.segments.append(FormUnit('Munken',data,start,end))
pass
#print data, 'is a segment'
elif dstr_role=='morpheme':
#self.universe.addData(Morpheme('Munken',data,start,end))
#elif dstr_role=='Checked tone':
#self.universe.addData(FormUnit(data)????)
#self.morphemes.append(Morpheme('Munken',data,start,end))
print data, ' is a Morpheme'
elif dstr_role=='translation':
#self.universe.addData(FormUnit('English',data,start,end))
#self.translations.append(FormUnit('English',data,start,end))
print data, ' is a translation'
#bug here: notes show up as forms of morphemes
#elif dstr_role=='notes':
# self.universe.addData(FormUnit('English',data,start,end))
def printIGT(self,out_file):
"""
build IGT objects based on number of translations
(since trans are common to the set of morphemes and segments );
assumes intervals are ordered (segment, syllable, morpheme, trans, notes)
"""
igts=[]
for t in self.translations:
temp_morphemes=[]
temp_segments=[]
#loop over morphemes to find those that are assoc'd with a translation
for g in self.morphemes:
if g.start>=t.start and g.end<=t.end:
temp_morphemes.append(g.label)
#loop over segment transcriptions to find those associated with each morpheme
for p in self.segments:
if p.start==g.start and p.end==g.end:
#print p.text
temp_segments.append(p.getString())
igts.append(IGT(temp_segments, temp_morphemes, t.segments))
c=IGTCollection(igts)
print 'there are', len(self.segments),'segment transcriptions'
#print 'there are', len(self.syllables),'syllables'
print 'there are', len(self.morphemes),'morphemes'
print 'there are ', len(self.translations),' translations'
print 'there are ', len(self.notes),' notes'
for t in self.translations:
print t.segments+' ',
#c.toStndOut()
c.toLatex(out_file)
class ElanReader(object):
"""
A class for reading Elan eaf files
"""
def __init__(self):
#one converter per input file is read
#(maybe change this to global if multiple files are to be read in a batch?
#assume file is in XSAMPA, for now
self.converter = CharConverter('xsampa', 'uni')
#the universe to populate
#self.universe=LingUniverse()
#will be used for interactive tier description
self.user_defined_types = []
#for hashing time points
self.time_points = {}
def setNS(self, ns):
"""
Set the namespace for the ontology.
@type ns: str
@param ns: a namespace string
@rtype:
@return:
"""
self.ns = ns
def setOntologyURI(self, uri):
"""
Ses the URI for the ontology.
@type uri: str
@param uri: a string representing a URI
"""
self.ontology = Ontology(uri)
def readElan(self, f):
"""Reads and parses an Elan file and populates the ling universe
@type f:string
@param f:an eaf file name
"""
print 'Reading Elan file...'
self.file_in = open(f, 'r')
#return IOError if root of eaf is not ANNOTATION_DOCUMENT
try:
file_lines = self.file_in.readlines(2)
if file_lines[1][1:20] != 'ANNOTATION_DOCUMENT':
raise IOError
print 'This Elan file is good.'
except IOError:
print "Input file is not well formed or not of type 'eaf'."
#go to beginning of file
self.file_in.seek(0, 0)
#begin xml processing
self.dom = parse(self.file_in)
#store time points in a dict. for later use
for t in self.dom.getElementsByTagName('TIME_SLOT'):
self.time_points[t.getAttribute('TIME_SLOT_ID')] = t.getAttribute(
'TIME_VALUE')
#store alignable elements in a dict. for later use
self.alignable_elems = self.dom.getElementsByTagName(
'ALIGNABLE_ANNOTATION')
#BEGIN MAIN ALGORITHM:
print 'Processing contents...'
#process tier by tier, based on whether they contain alignable or reference annotations
tier_elems = self.dom.getElementsByTagName('TIER')
#get user defined ling types
for tier in tier_elems:
self.user_defined_types.append(
tier.getAttribute('LINGUISTIC_TYPE_REF'))
#print self.user_defined_types
for tier in tier_elems:
#dstr_role='data structure role'
#used to decide on what type of linguistic unit to instantiate
dstr_role = tier.getAttribute('LINGUISTIC_TYPE_REF')
alignable_elems = tier.getElementsByTagName('ALIGNABLE_ANNOTATION')
if len(alignable_elems) > 0:
self.handleAnnotation(alignable_elems, dstr_role)
ref_elems = tier.getElementsByTagName('REF_ANNOTATION')
if len(ref_elems) > 0:
self.handleAnnotation(ref_elems, dstr_role)
#BEGIN OTHER CLASS METHODS
def handleAnnotation(self, elems, dstr_role):
"""
Process ALIGNABLE_ANNOTATION and REF_ANNOTATION elements
@type elems:
@param elems: the elements to start from
@type dstr_role:
@param dstr_role: the element type
"""
#print 'There are ',len(elems),' annotation elements.'
for e in elems:
annot_val_elems = e.getElementsByTagName('ANNOTATION_VALUE')
if annot_val_elems > 0:
self.handleAnnotationValue(annot_val_elems, dstr_role,
self.findTimeInterval(e))
def handleAnnotationValue(self, annot_val_elems, dstr_role, time_interval):
""" Process CDATA associated with ANNOTATION_VALUE elements.
annot_val_elems the elements to be processed
dstr_role element type to be passed on
time_interval time interval list to be passed on
Calls handleData(...)
"""
for v in annot_val_elems:
for c in v.childNodes:
if c.nodeType == 3:
self.handleData(c.data, dstr_role, time_interval)
def findTimeInterval(self, elem):
"""Build time interval list
elem the element to start from
Return a list containing the start and end times
"""
if elem.tagName == 'ALIGNABLE_ANNOTATION':
start = elem.getAttribute('TIME_SLOT_REF1')
end = elem.getAttribute('TIME_SLOT_REF2')
#look up time refs in hash
return [self.time_points[start], self.time_points[end]]
elif elem.tagName == 'REF_ANNOTATION':
ref = elem.getAttribute('ANNOTATION_REF')
for e in self.alignable_elems:
if e.getAttribute('ANNOTATION_ID') == ref:
#recursive call
return self.findTimeInterval(e)
def handleData(self, data, dstr_role, time_interval):
"""Decide on which linguistic units to instantiate.
data string repr. of linguistic form or grammar unit label
dstr_role how the data is used in the Elan file (gloss, translation, etc)
time_interval time alignment
Instantiates units and adds to ling. universe
"""
start = float(time_interval[0])
end = float(time_interval[1])
#build data obj's and add to universe
if dstr_role == 'Sentence-level Transcription':
#convert unicode to string (nec. for character converter)
if type(data) == unicode: data = str(data)
data = self.converter.convert(data)
print data, ' is a Clause or Phrase'
#generalize this later
data = data.split()
for d in data:
print d, ' is a SyntacticWord'
w = gold.SyntacticWord(self.ns + makeID(d), [])
#self.universe.addData(FormUnit('Koshin',d,start,end))
start = start + .00001
#self.universe.addData(FormUnit(data,start,end))
elif dstr_role == 'Morpheme':
#generalize this later
###############################################
print data, '###data'
data = data.split()
for d in data:
print d, '####d'
#self.universe.addData(Morpheme('Koshin',d,start,end))
start = start + .00001
################################################
#self.universe.addData(Morpheme(data,start,end))
#elif dstr_role=='Checked tone':
# self.universe.addData(FormUnit(data,start,end))
elif dstr_role == 'Sentence-level Translation':
#print 'trans'
pass
#self.universe.addData(FormUnit('English',data,start,end))
#elif dstr_role=='Notes':
# self.universe.addData(FormUnit(data,start,end)
def writeData(self, outfile):
self.ontology.save(outfile)
