def processFromPath(filename, filepath):
tg = ExtendedTextGrid(f=filepath)
tg.oprint = False
wtier = findWordsTier(tg, filename)
# saveTierAsLM(wtier,"words", filepath)
# lexiconFromTier(wtier, filepath)
tg.predictPhns(wtier.name, "gen_phones")
saveTierAsLM(generateLandmarkTier(tg.get_tier("gen_phones")), "landmarks", filepath)
def processFromPath(filename, filepath):
tg = ExtendedTextGrid(f=filepath)
tg.oprint = False
wtier = findWordsTier(tg, filename)
saveTierAsLM(wtier, "words", filepath)
lexiconFromTier(wtier, filepath)
tg.predictPhns(wtier.name, "gen_phones")
saveTierAsLM(generateLandmarkTier(tg.get_tier("gen_phones")), "landmarks",
filepath)
saveTierAsLM(tg.get_tier("gen_phones"), "phones", filepath)
BEFORE YOU RUN, PLEASE READ:
* FIRST, move your textgrid files to data/source folder. Work out of that folder.
REQUIRED TIER NAMES:
* "words": interval tier with full words
* either "landmarks" or "LM": point tier with t-cl style landmarks
* either "LMmod" or "LMmods" or "comments": point tier with t-cl-x style landmarks
NOTE: If a tier named "phones" exists, it will be renamed to "phones_prev" and a
new, generated "phones" tier will be used.
"""
TEXTGRID_FILES = ["Gabrieli/ASDH_SLI_3126_allison.TextGrid",
"Gabrieli/ASDL_BILDC3099_allison.TextGrid",
"Gabrieli/SLI_SLI3086_allison.TextGrid",
"Gabrieli/SLI3084_allison.TextGrid"]
from ExtendedTextGrid import *
for TEXTGRID_FILE in TEXTGRID_FILES:
print(TEXTGRID_FILE)
src = "data/source/" + TEXTGRID_FILE
print("Creating .tab from "+src+"...")
tg = ExtendedTextGrid(src)
tg.prepare()
tg.extractContext()
tg.saveTab()
print("Saved as "+tg.fname+".tab")
convProcess.py
Combines the functionality of generatePhonesTier.py, saveTierAsLM.py, and LexiconExtract.py, used for batch processing 16 textgrids at once.
"""
import sys, logging
# Require Python 3.x
if sys.version_info[0] < 3:
print("Error: The TextGrid processor requires Python 3.0 or above. Exiting.\n")
sys.exit(1)
#This will crash on python 2.x so it has to go below the version test
from ExtendedTextGrid import *
for i in range(1, 17):
filename = "conv{num:02d}g".format(num=i)
filepath = "../landmarks/matcher-data/"+filename+".TextGrid"
tg = ExtendedTextGrid(f=filepath)
tg.oprint = False
wtier = None
tier_names = [tier.name for tier in tg]
for pn in ["word", "words", "Word", "Words", filename, filename[:-1]]:
if pn in tier_names:
wtier = tg.get_tier(pn)
if wtier == None:
print(filename)
for i, tier in zip(range(len(tg)), tg):
print(str(i) + " " + tier.name)
wt = int(input("Words tier: "))
wtier = tg[wt]
from ExtendedTextGrid import *
# [Data Preparation]
# Define file path here
sourcetg = "C:\\Users\mz\Documents\\landmark\conv07.TextGrid"
sourcepd = "C:\\Users\mz\Documents\\landmark\data\conv07g_RS.TextGrid"
fname = "C:\\Users\mz\Documents\\landmark\data\conv07_ext.TextGrid"
# Load file
tg = ExtendedTextGrid(f=sourcetg)
tg.saveAs(fname)
### Apply lexicon, predict landmarks, and unify landmark label format
##tg.putPhns()
##tg.predictLM()
e = tg.convertLM(True)
### Convert words tier to context-rich objects
##tg.extendWords()
### Match landmarks with corresponding phones
##tg.linkLMtoWords("pred. LM")
##tg.linkLMtoWords("act. LM")
##tg.linkLMtoPhones()
### Align predicted and observed landmarks
##tg.alignLM()
### Adjust position of each predicted landmark to the aligned observed landmark
##lm_tier = tg.get_tier("pred. lm")
### Create a new tier to display landmark modifications
##tg.append(lm_tier.aligned())
##
### Convert textgrid into table
##tg.saveTab()
while not f:
path = input("Enter TextGrid file path: ")
if not path.lower().endswith(".textgrid"):
print("File must end with .textgrid")
continue
try:
f = open(path)
f.close()
except IOError:
print("File does not exist")
continue
return path
while 1: #Loop until putPhns() runs sucessfully, then break & save
tg = ExtendedTextGrid(pathFromInput())
try:
tg.putPhns()
break
except Exception:
print("There must be a tier named \"words\".")
lm = input("Do you want to generate a landmark tier (yes/no)? ").lower()
if lm == "yes" or lm == "y":
import phones_to_landmarks_dict
d = phones_to_landmarks_dict.d
phonetier = tg.get_tier("phones")
# Initiate new textgrid tiers for predicted landmarks, phonemes, voicing, and nosal info
lmtier = PointTier(name="LM", xmin=0, xmax=phonetier.xmax)
for interval in phonetier:
continue
return path
# while 1: #Loop until putPhns() runs sucessfully, then break & save
# tg = ExtendedTextGrid(pathFromInput())
# try:
# tg.putPhns()
# break
# except Exception:
# print("There must be a tier named \"words\".")
# lm = input("Do you want to generate a landmark tier (yes/no)? ").lower()
# if lm=="yes" or lm=="y":
from phones_to_landmarks_dict import d #Dictionary of phone -> landmark
tg = ExtendedTextGrid(pathFromInput())
phonetier = tg.get_tier("phones")
if not phonetier:
print("TextGrid must contain a tier named 'phones'")
sys.exit()
# Initiate new textgrid tiers for predicted landmarks, phonemes, voicing, and nosal info
lmtier = PointTier(name="LM", xmin = 0, xmax=phonetier.xmax)
for interval in phonetier:
try:
# non-words
phone = interval.text.lower().strip("\t \" +?.'[],012") # ignore uncertainty marks
if not ('<' in phone or '>' in phone or phone=='' or phone=="#"):
lms=d[phone]
duration = (interval.xmax-interval.xmin)/len(lms) # duration of each phoneme
from ExtendedTextGrid import *
# [Data Preparation]
# Define file path here
sourcetg = "C:\\Users\mz\Documents\\landmark\conv07.TextGrid"
sourcepd = "C:\\Users\mz\Documents\\landmark\data\conv07g_RS.TextGrid"
fname = "C:\\Users\mz\Documents\\landmark\data\conv07_ext.TextGrid"
# Load file
tg = ExtendedTextGrid(f=sourcetg)
tg.saveAs(fname)
### Apply lexicon, predict landmarks, and unify landmark label format
##tg.putPhns()
##tg.predictLM()
e=tg.convertLM(True)
### Convert words tier to context-rich objects
##tg.extendWords()
### Match landmarks with corresponding phones
##tg.linkLMtoWords("pred. LM")
##tg.linkLMtoWords("act. LM")
##tg.linkLMtoPhones()
### Align predicted and observed landmarks
##tg.alignLM()
### Adjust position of each predicted landmark to the aligned observed landmark
##lm_tier = tg.get_tier("pred. lm")
### Create a new tier to display landmark modifications
##tg.append(lm_tier.aligned())
##
### Convert textgrid into table
##tg.saveTab()
if not path.lower().endswith(".textgrid"):
print("File must end with .textgrid")
continue
try:
f = open(path)
f.close()
except IOError:
print("File does not exist")
continue
return path
if len(sys.argv) == 2:
try:
path = sys.argv[1]
f = open(path)
f.close()
except IOError:
path = pathFromInput()
else:
path = pathFromInput()
tg = ExtendedTextGrid(path)
for tier in tg:
if type(tier) == IntervalTier:
print("running on tier " + tier.name)
tier.fixIntervalSpaces()
tg.save()
print("Saved as " + tg.fname + ".TextGrid.")
while not f:
path = input("Enter TextGrid file path: ")
if not path.lower().endswith(".textgrid"):
print("File must end with .textgrid")
continue
try:
f = open(path)
f.close()
except IOError:
print("File does not exist")
continue
return path
if len(sys.argv) == 2:
try:
path = sys.argv[1]
f = open(path)
f.close()
except IOError:
path = pathFromInput()
else:
path = pathFromInput()
tg = ExtendedTextGrid(path)
for tier in tg:
if type(tier) == IntervalTier:
print("running on tier "+tier.name)
tier.fixIntervalSpaces()
tg.save()
print("Saved as "+tg.fname+".TextGrid.")
