# turn the string into a list
warpa_ph = arpa_phon_str.split()
except Exception, e:
warpa_ph = ["NO MATCH"]
return warpa_ph
# ----------------------------------------------------------------------------
context = {}
# ---- Long Text Box --------------------------------------------
form1 = EnglishInputForm()
context['form_1'] = form1
# if request.method == 'GET':
# input_string = request.GET.get('input', '')
if request.method == 'POST':
form1 = EnglishInputForm(request.POST)
context['form_1'] = form1
if form1.is_valid():
# read data from form
input_string = form1.cleaned_data['english_input_f']
context['english_input_v'] = input_string
# Make a map that shows capitals, miniscules, apostrophes, and non-letters
# char_to_remove = ['0', '1', '2']
# arpa_phon_clean = arpa_phon_str.translate(None, ''.join(char_to_remove))
# turn the string into a list
warpa_ph = arpa_phon_str.split()
except Exception, e:
warpa_ph = ["NO MATCH"]
return warpa_ph
# ----------------------------------------------------------------------------
context = {}
# ---- Long Text Box --------------------------------------------
form1 = EnglishInputForm()
context['form_1'] = form1
if request.method == 'POST':
form1 = EnglishInputForm(request.POST)
context['form_1'] = form1
if form1.is_valid():
# read data from form
input_string = form1.cleaned_data['english_input_f']
context['english_input_v'] = input_string
# Make a map that shows capitals, miniscules, apostrophes, and non-letters
input_length = len(input_string)
# ----------------------------------------------------------------------------
def input_output(request):
context = {}
# ---- Long Text Box --------------------------------------------
form1 = EnglishInputForm()
context['form_1'] = form1
if request.method == 'POST':
form1 = EnglishInputForm(request.POST)
context['form_1'] = form1
if form1.is_valid():
# read data from form
input_string = form1.cleaned_data['english_input_f']
context['english_input_v'] = input_string
# Make a map that shows capitals, miniscules, apostrophes, and non-letters
input_length = len(input_string)
input_map1 = []
for i in range(0, input_length):
# test if char is a capital letter and, thus, part of a word
if ord(input_string[i]) > 64 and ord(input_string[i]) < 91:
input_map1.append("C")
# test if char is a miniscule letter and, thus, part of a word
elif ord(input_string[i]) > 96 and ord(input_string[i]) < 123:
input_map1.append("m")
# test for apostrophes (which will be dealt with below)
elif input_string[i] == "'":
input_map1.append("a")
# for now we'll say everything else is not part of a word
# we'll probably have to deal with hyphens at some point
else:
input_map1.append(input_string[i])
# Test if char is a non-terminal apostrophe (and, thus, part
# of a word) or a single-quote mark or terminal apostrophe
# (and, thus, not part of a word)
for i in range(0, input_length):
# find the apostrophe marks
if input_map1[i] == "a":
# test if apostrophe starts or ends the string
if i == 0 or i == input_length - 1:
input_map1[i] = "'"
# test if apostrophe is part of a word (ie, has letters on both sides)
elif input_map1[i - 1] in (
"C", "m") and input_map1[i + 1] in ("C", "m"):
input_map1[i] = input_map1[i + 1]
# every other apostrophe is not part of a word
else:
input_map1[i] = "'"
print "input_map1 (joined): %s" % ''.join(input_map1)
# Finalize the map, categorizing words as all-caps (CC),
# initial-cap (Cm), and no-caps (mm)
# Also make a list of words to be transliterated.
input_map2 = []
word_index = []
for i in range(0, input_length):
# put non-words directly into the final map
if input_map1[i] not in ("C", "m"):
input_map2.append(input_map1[i])
# record index for the end of a word
# print "im1[i-1]: %s" % input_map1[i-1]
if i == 0:
pass
elif input_map1[i - 1] in ("C", "m"):
word_index.append(i)
# everything else is part of a word
else:
# find first-letters and put them in map
# record the index for the start of the word
if i == 0 or input_map1[i - 1] not in ("C", "m"):
input_map2.append(input_map1[i])
word_index.append(i)
# find second-letters (if present) and put them in map
elif i == 1 and input_map1[0] in ("C", "m"):
input_map2[0] = input_map1[0] + input_map1[1]
elif input_map1[i - 2] not in (
"C", "m") and input_map1[i - 1] in ("C", "m"):
cur_index = len(input_map2) - 1
let_1 = input_map2[cur_index]
let_2 = input_map1[i]
input_map2[cur_index] = let_1 + let_2
# Check if last char of string is a letter and make it end of word
if i == len(input_map1) - 1 and input_map1[i] in ("C", "m"):
word_index.append(i + 1)
# ignore all other letters [remember that at this stage
# we're ignoring the possibility of mixing upper and
# lower case letters except init caps]
print "input_map2 (joined): %s" % ''.join(input_map2)
print "word_index: %s" % word_index
# generate list of words to be transliterated
word_list = []
word_count = len(word_index)
for i in range(0, word_count, 2):
start = word_index[i]
stop = word_index[i + 1]
word_list.append(input_string[start:stop])
print "word_list: %s" % word_list
# Get Arpabet correlations for each word in word_list
arpa_phonemes = {}
word_num = 0
for word in word_list:
# Since the word might not be in the dictionary, we
# try to fail softly
try:
arpa_output = DictCMU.objects.get(entry__iexact=word)
# the db stores pronunc info as unicode; we force it to str
arpa_phon_str = str(arpa_output.phonemes)
# remove accent characters
char_to_remove = ['0', '1', '2']
arpa_phon_clean = arpa_phon_str.translate(
None, ''.join(char_to_remove))
# turn the string into a list
arpa_phonemes[word] = arpa_phon_clean.split()
except Exception, e:
arpa_phonemes[word] = "NO MATCH"
word_num += 1
print "arpa_phonemes: %s" % arpa_phonemes
# The correlation between the Arpabet and the Deseret script.
# Note: This is a provisional correlation for early proof-of-concept
# work; it has serious flaws in the correlations.
arpa_corr_U = {
'IY': u'\uD801\uDC00',
'EY': u'\uD801\uDC01',
'AA': u'\uD801\uDC02',
'AO': u'\uD801\uDC03',
'OW': u'\uD801\uDC04',
'UW': u'\uD801\uDC05',
'IH': u'\uD801\uDC06',
'EH': u'\uD801\uDC07',
'AE': u'\uD801\uDC08',
'AH': u'\uD801\uDC0A',
'UH': u'\uD801\uDC0B',
'AY': u'\uD801\uDC0C',
'AW': u'\uD801\uDC0D',
'W': u'\uD801\uDC0E',
'Y': u'\uD801\uDC0F',
'HH': u'\uD801\uDC10',
'P': u'\uD801\uDC11',
'B': u'\uD801\uDC12',
'T': u'\uD801\uDC13',
'D': u'\uD801\uDC14',
'CH': u'\uD801\uDC15',
'JH': u'\uD801\uDC16',
'K': u'\uD801\uDC17',
'G': u'\uD801\uDC18',
'F': u'\uD801\uDC19',
'V': u'\uD801\uDC1A',
'TH': u'\uD801\uDC1B',
'DH': u'\uD801\uDC1C',
'S': u'\uD801\uDC1D',
'Z': u'\uD801\uDC1E',
'SH': u'\uD801\uDC1F',
'ZH': u'\uD801\uDC20',
'R': u'\uD801\uDC21',
'L': u'\uD801\uDC22',
'M': u'\uD801\uDC23',
'N': u'\uD801\uDC24',
'NG': u'\uD801\uDC25',
'OY': u'\uD801\uDC26',
'ER': u'\uD801\uDC21'
}
arpa_corr_m = {
'IY': u'\uD801\uDC28',
'EY': u'\uD801\uDC29',
'AA': u'\uD801\uDC2A',
'AO': u'\uD801\uDC2B',
'OW': u'\uD801\uDC2C',
'UW': u'\uD801\uDC2D',
'IH': u'\uD801\uDC2E',
'EH': u'\uD801\uDC2F',
'AE': u'\uD801\uDC30',
'AH': u'\uD801\uDC32',
'UH': u'\uD801\uDC33',
'AY': u'\uD801\uDC34',
'AW': u'\uD801\uDC35',
'W': u'\uD801\uDC36',
'Y': u'\uD801\uDC37',
'HH': u'\uD801\uDC38',
'P': u'\uD801\uDC39',
'B': u'\uD801\uDC3A',
'T': u'\uD801\uDC3B',
'D': u'\uD801\uDC3C',
'CH': u'\uD801\uDC3D',
'JH': u'\uD801\uDC3E',
'K': u'\uD801\uDC3F',
'G': u'\uD801\uDC40',
'F': u'\uD801\uDC41',
'V': u'\uD801\uDC42',
'TH': u'\uD801\uDC43',
'DH': u'\uD801\uDC44',
'S': u'\uD801\uDC45',
'Z': u'\uD801\uDC46',
'SH': u'\uD801\uDC47',
'ZH': u'\uD801\uDC48',
'R': u'\uD801\uDC49',
'L': u'\uD801\uDC4A',
'M': u'\uD801\uDC4B',
'N': u'\uD801\uDC4C',
'NG': u'\uD801\uDC4D',
'OY': u'\uD801\uDC4E',
'ER': u'\uD801\uDC49'
}
# get unicode codes for Deseret characters that correlate to the Arpabet
# in arpa_phonemes, taking care to match capital and lower-case letters
unic_out = []
w_num = 0
for inp in input_map2:
# Check if inp is a word or not. If inp is a word...
if inp in ("CC", "C", "Cm", "mm", "m"):
word = word_list[w_num]
arpa_ph = arpa_phonemes[word]
# Deal with words not in dictionary
if arpa_ph == "NO MATCH":
a = "." * len(word)
unic_out.append(a)
w_num += 1
# All-Caps words
elif inp == "CC" or inp == "C":
for phone in arpa_ph:
unic_out.append(arpa_corr_U[phone])
w_num += 1
# Initial-Caps words
elif inp == "Cm":
unic_out.append(arpa_corr_U[arpa_ph[0]])
for i in range(1, len(arpa_ph)):
unic_out.append(arpa_corr_m[arpa_ph[i]])
w_num += 1
# No-Caps words
elif inp == "mm" or inp == "m":
for phone in arpa_ph:
unic_out.append(arpa_corr_m[phone])
w_num += 1
# If inp is not a word, insert the symbols directly.
else:
unic_out.append(inp)
# print "unic_out: %s" % unic_out
context['unic_out'] = unic_out
unic_str = ''.join(unic_out)
print "unic_str: %s" % unic_str
context['unic_str'] = unic_str
def input_output(request):
context = {}
# ---- Long Text Box --------------------------------------------
form1 = EnglishInputForm()
context['form_1'] = form1
if request.method == 'POST':
form1 = EnglishInputForm(request.POST)
context['form_1'] = form1
if form1.is_valid():
# read data from form
input_string = form1.cleaned_data['english_input_f']
context['english_input_v'] = input_string
# Make a map that shows capitals, miniscules, apostrophes, and non-letters
input_length = len(input_string)
input_map1 = []
for i in range(0, input_length):
# test if char is a capital letter and, thus, part of a word
if ord(input_string[i]) > 64 and ord(input_string[i]) < 91:
input_map1.append("C")
# test if char is a miniscule letter and, thus, part of a word
elif ord(input_string[i]) > 96 and ord(input_string[i]) < 123:
input_map1.append("m")
# test for apostrophes (which will be dealt with below)
elif input_string[i] == "'":
input_map1.append("a")
# for now we'll say everything else is not part of a word
# we'll probably have to deal with hyphens at some point
else:
input_map1.append(input_string[i])
# Test if char is a non-terminal apostrophe (and, thus, part
# of a word) or a single-quote mark or terminal apostrophe
# (and, thus, not part of a word)
for i in range(0, input_length):
# find the apostrophe marks
if input_map1[i] == "a":
# test if apostrophe starts or ends the string
if i == 0 or i == input_length - 1:
input_map1[i] = "'"
# test if apostrophe is part of a word (ie, has letters on both sides)
elif input_map1[i-1] in ("C", "m") and input_map1[i+1] in ("C", "m"):
input_map1[i] = input_map1[i+1]
# every other apostrophe is not part of a word
else:
input_map1[i] = "'"
print "input_map1 (joined): %s" % ''.join(input_map1)
# Finalize the map, categorizing words as all-caps (CC),
# initial-cap (Cm), and no-caps (mm)
# Also make a list of words to be transliterated.
input_map2 = []
word_index = []
for i in range(0, input_length):
# put non-words directly into the final map
if input_map1[i] not in ("C", "m"):
input_map2.append(input_map1[i])
# record index for the end of a word
# print "im1[i-1]: %s" % input_map1[i-1]
if i == 0:
pass
elif input_map1[i-1] in ("C", "m"):
word_index.append(i)
# everything else is part of a word
else:
# find first-letters and put them in map
# record the index for the start of the word
if i == 0 or input_map1[i-1] not in ("C", "m"):
input_map2.append(input_map1[i])
word_index.append(i)
# find second-letters (if present) and put them in map
elif i == 1 and input_map1[0] in ("C", "m"):
input_map2[0] = input_map1[0] + input_map1[1]
elif input_map1[i-2] not in ("C", "m") and input_map1[i-1] in ("C", "m"):
cur_index = len(input_map2) - 1
let_1 = input_map2[cur_index]
let_2 = input_map1[i]
input_map2[cur_index] = let_1 + let_2
# Check if last char of string is a letter and make it end of word
if i == len(input_map1) - 1 and input_map1[i] in ("C", "m"):
word_index.append(i+1)
# ignore all other letters [remember that at this stage
# we're ignoring the possibility of mixing upper and
# lower case letters except init caps]
print "input_map2 (joined): %s" % ''.join(input_map2)
print "word_index: %s" % word_index
# generate list of words to be transliterated
word_list = []
word_count = len(word_index)
for i in range(0, word_count, 2):
start = word_index[i]
stop = word_index[i+1]
word_list.append(input_string[start:stop])
print "word_list: %s" % word_list
# Get Arpabet correlations for each word in word_list
arpa_phonemes = {}
word_num = 0
for word in word_list:
# Since the word might not be in the dictionary, we
# try to fail softly
try:
arpa_output = DictCMU.objects.get(entry__iexact=word)
# the db stores pronunc info as unicode; we force it to str
arpa_phon_str = str(arpa_output.phonemes)
# remove accent characters
char_to_remove = ['0', '1', '2']
arpa_phon_clean = arpa_phon_str.translate(None, ''.join(char_to_remove))
# turn the string into a list
arpa_phonemes[word] = arpa_phon_clean.split()
except Exception, e:
arpa_phonemes[word] = "NO MATCH"
word_num += 1
print "arpa_phonemes: %s" % arpa_phonemes
# The correlation between the Arpabet and the Deseret script.
# Note: This is a provisional correlation for early proof-of-concept
# work; it has serious flaws in the correlations.
arpa_corr_U = {'IY': u'\uD801\uDC00', 'EY': u'\uD801\uDC01', 'AA': u'\uD801\uDC02', 'AO': u'\uD801\uDC03', 'OW': u'\uD801\uDC04', 'UW': u'\uD801\uDC05', 'IH': u'\uD801\uDC06', 'EH': u'\uD801\uDC07', 'AE': u'\uD801\uDC08', 'AH': u'\uD801\uDC0A', 'UH': u'\uD801\uDC0B', 'AY': u'\uD801\uDC0C', 'AW': u'\uD801\uDC0D', 'W': u'\uD801\uDC0E', 'Y': u'\uD801\uDC0F', 'HH': u'\uD801\uDC10', 'P': u'\uD801\uDC11', 'B': u'\uD801\uDC12', 'T': u'\uD801\uDC13', 'D': u'\uD801\uDC14', 'CH': u'\uD801\uDC15', 'JH': u'\uD801\uDC16', 'K': u'\uD801\uDC17', 'G': u'\uD801\uDC18', 'F': u'\uD801\uDC19', 'V': u'\uD801\uDC1A', 'TH': u'\uD801\uDC1B', 'DH': u'\uD801\uDC1C', 'S': u'\uD801\uDC1D', 'Z': u'\uD801\uDC1E', 'SH': u'\uD801\uDC1F', 'ZH': u'\uD801\uDC20', 'R': u'\uD801\uDC21', 'L': u'\uD801\uDC22', 'M': u'\uD801\uDC23', 'N': u'\uD801\uDC24', 'NG': u'\uD801\uDC25', 'OY': u'\uD801\uDC26', 'ER': u'\uD801\uDC21'}
arpa_corr_m = {'IY': u'\uD801\uDC28', 'EY': u'\uD801\uDC29', 'AA': u'\uD801\uDC2A', 'AO': u'\uD801\uDC2B', 'OW': u'\uD801\uDC2C', 'UW': u'\uD801\uDC2D', 'IH': u'\uD801\uDC2E', 'EH': u'\uD801\uDC2F', 'AE': u'\uD801\uDC30', 'AH': u'\uD801\uDC32', 'UH': u'\uD801\uDC33', 'AY': u'\uD801\uDC34', 'AW': u'\uD801\uDC35', 'W': u'\uD801\uDC36', 'Y': u'\uD801\uDC37', 'HH': u'\uD801\uDC38', 'P': u'\uD801\uDC39', 'B': u'\uD801\uDC3A', 'T': u'\uD801\uDC3B', 'D': u'\uD801\uDC3C', 'CH': u'\uD801\uDC3D', 'JH': u'\uD801\uDC3E', 'K': u'\uD801\uDC3F', 'G': u'\uD801\uDC40', 'F': u'\uD801\uDC41', 'V': u'\uD801\uDC42', 'TH': u'\uD801\uDC43', 'DH': u'\uD801\uDC44', 'S': u'\uD801\uDC45', 'Z': u'\uD801\uDC46', 'SH': u'\uD801\uDC47', 'ZH': u'\uD801\uDC48', 'R': u'\uD801\uDC49', 'L': u'\uD801\uDC4A', 'M': u'\uD801\uDC4B', 'N': u'\uD801\uDC4C', 'NG': u'\uD801\uDC4D', 'OY': u'\uD801\uDC4E', 'ER': u'\uD801\uDC49'}
# get unicode codes for Deseret characters that correlate to the Arpabet
# in arpa_phonemes, taking care to match capital and lower-case letters
unic_out = []
w_num = 0
for inp in input_map2:
# Check if inp is a word or not. If inp is a word...
if inp in ("CC", "C", "Cm", "mm", "m"):
word = word_list[w_num]
arpa_ph = arpa_phonemes[word]
# Deal with words not in dictionary
if arpa_ph == "NO MATCH":
a = "."*len(word)
unic_out.append(a)
w_num += 1
# All-Caps words
elif inp == "CC" or inp == "C":
for phone in arpa_ph:
unic_out.append(arpa_corr_U[phone])
w_num += 1
# Initial-Caps words
elif inp == "Cm":
unic_out.append(arpa_corr_U[arpa_ph[0]])
for i in range(1, len(arpa_ph)):
unic_out.append(arpa_corr_m[arpa_ph[i]])
w_num += 1
# No-Caps words
elif inp == "mm" or inp == "m":
for phone in arpa_ph:
unic_out.append(arpa_corr_m[phone])
w_num += 1
# If inp is not a word, insert the symbols directly.
else:
unic_out.append(inp)
# print "unic_out: %s" % unic_out
context['unic_out'] = unic_out
unic_str = ''.join(unic_out)
print "unic_str: %s" % unic_str
context['unic_str'] = unic_str
