acoustic_model/acoustic_model/acoustic_model.py

import os
import sys
import tempfile
import configparser
import subprocess

import pandas as pd


## ======================= user define =======================
repo_dir = 'C:\\Users\\Aki\\source\\repos\\acoustic_model'
curr_dir = repo_dir + '\\acoustic_model'
config_ini = curr_dir + '\\config.ini'
output_dir = 'd:\\OneDrive\\Research\\rug\\experiments\\friesian\\acoustic_model'
forced_alignment_module = 'C:\\Users\\Aki\\source\\repos\\forced-alignment'

sys.path.append(os.path.join(os.path.dirname(sys.path[0]), curr_dir))
sys.path.append(forced_alignment_module)
from forced_alignment import convert_phone_set


import acoustic_model_functions as am_func


## ======================= load variables =======================

config = configparser.ConfigParser()
config.sections()
config.read(config_ini)

config_hcopy = config['Settings']['config_hcopy']
config_train = config['Settings']['config_train']
FAME_dir	 = config['Settings']['FAME_dir']

lexicon_file = FAME_dir + '\\lexicon\\lex.asr'
dataset_list = ['devel', 'test', 'train']


## ======================= extract features =======================
##dataset = dataset_list[0]
#for dataset in dataset_list:
#	print(dataset)
	
	## make a script file for HCopy 
	#hcopy_scp = tempfile.NamedTemporaryFile(mode='w', delete=False)
	#hcopy_scp.close()

	## using the filelist in FAME! corpus
	#feature_dir = output_dir + '\\mfc\\' + dataset
	#am_func.make_hcopy_scp_from_filelist_in_fame(FAME_dir, dataset, feature_dir, hcopy_scp.name)

	## extract features
	#subprocessStr = 'HCopy -C ' + config_hcopy + ' -S ' + hcopy_scp.name
	#subprocess.call(subprocessStr, shell=True)

	#os.remove(hcopy_scp.name)


## ======================= make a list of features =======================
##dataset = dataset_list[2]
#for dataset in dataset_list:
#	print(dataset)

#	feature_dir = output_dir + '\\mfc\\' + dataset
#	hcompv_scp  = output_dir + '\\scp\\' + dataset + '.scp'

#	am_func.make_filelist(feature_dir, hcompv_scp)


## ======================= check the phonemes used in the lexicon =======================
phonelist = am_func.get_phonelist(lexicon_file) # 49
phonelist_list = list(phonelist)

#lines_g1 = am_func.find_phone(lexicon_file, 'g')
#lines_g2 = am_func.find_phone(lexicon_file, 'ɡ')


## ======================= make label file =======================
dataset = 'train'
hcompv_scp  = output_dir + '\\scp\\' + dataset + '.scp'
script_list = FAME_dir + '\\data\\' + dataset + '\\text'

lexicon = pd.read_table(lexicon_file, names=['word', 'pronunciation'])

with open(hcompv_scp) as fin:
	features = fin.read()
	features = features.split('\n')

with open(script_list, "rt", encoding="utf-8") as fin:
	scripts = fin.read()
	scripts = pd.Series(scripts.split('\n'))


feature = features[0]
file_basename = os.path.basename(feature).replace('.mfc', '')

# get words from scripts.
script = scripts[scripts.str.contains(file_basename)]
script_id  = script.index[0]
script_txt = script.get(script_id)
script_words = script_txt.split(' ')
del script_words[0]

# make the label file.
SCRIPT_WORDS = []
script_prons = []
all_prons_found = 1
for word in script_words:
	SCRIPT_WORDS.append(word.upper())
	extracted = lexicon[lexicon['word']==word]
	script_prons.append(extracted)
	all_prons_found *= len(extracted)
# make the dict file.

convert_phone_set.ipa2fame(phonelist_list)
phonelist_list