def copy_files(self):
run_cmds(f"cd {self.audio_path} && cp -R * {self.tuned_path}")
# Using bash instead of python to closely follow directions on
# https://cmusphinx.github.io/wiki/tutorialadapt/
for _dir in ["en-us", "cmudict-en-us.dict", "en-us.lm.bin"]:
path = os.path.join(self.model_path, _dir)
run_cmds(f"cp -a {path} {self.tuned_path}")
def write_test_files(self):
makedirs(self.test_dir)
wav_dir = os.path.join(self.test_dir, "wav/")
makedirs(wav_dir)
old_lm = os.path.join(self.tuned_path, "en-us.lm.bin")
new_lm = os.path.join(self.test_dir, "en-us.lm.bin")
old_dict = os.path.join(self.tuned_path, "cmudict-en-us.dict")
new_dict = os.path.join(self.test_dir, "cmudict-en-us.dict")
old_hmm = os.path.join(self.tuned_path, "en-us")
new_hmm = os.path.join(self.test_dir, "en-us")
old_mllr_matrix = os.path.join(self.tuned_path, "mllr_matrix")
new_mllr_matrix = os.path.join(self.test_dir, "mllr_matrix")
run_cmds(
[
f"cd {self.tuned_path}",
f"cp {self.file_ids_path} {self.test_file_ids_path}",
(f"cp {self.transcription_path} "
f"{self.test_transcription_path}"),
f"cp *wav {wav_dir}",
f"cp {old_lm} {new_lm}",
f"cp {old_dict} {new_dict}",
f"cp -R {old_hmm} {new_hmm}",
#f"cp {old_mllr_matrix} {new_mllr_matrix}",
f"cp {self.tuned_path}/sphinxtrain/scripts/decode/word_align.pl ./test/"
],
stdout=True)
def run_sphinx_fe(self):
run_cmds([
f"cd {self.tuned_path}",
(f"sphinx_fe -argfile en-us/feat.params "
f"-samprate 16000 -c {self.file_ids_path} "
"-di . -do . -ei wav -eo mfc -mswav yes")
],
stdout=True)
def run_mllr(self):
# NOTE: not nearly as effective as run_adapt.
# Now we just use that instead
run_cmds([
f"cd {self.tuned_path}",
("./mllr_solve\\\n"
" -meanfn en-us/means \\\n"
" -varfn en-us/variances \\\n"
" -outmllrfn mllr_matrix -accumdir .")
],
stdout=True)
def run_adapt(self):
cmds = [
f"cd {self.tuned_path}", "cp -R en-us en-us-adapt",
("./map_adapt -moddeffn en-us/mdef.txt -ts2cbfn .cont. "
"-meanfn en-us/means -varfn en-us/variances -mixwfn "
"en-us/mixture_weights -tmatfn en-us/transition_matrices "
"-accumdir . -mapmeanfn en-us-adapt/means -mapvarfn "
"en-us-adapt/variances -mapmixwfn "
"en-us-adapt/mixture_weights -maptmatfn "
"en-us-adapt/transition_matrices")
]
run_cmds(cmds, stdout=True)
def download_sphinxtrain(self):
# Must get installed from source for fixes
run_cmds([
f"cd {self.tuned_path}",
"git clone [email protected]:cmusphinx/sphinxtrain.git",
"cd sphinxtrain", "./autogen.sh", f"make -j {cpu_count()}",
"sudo make install"
],
stdout=True)
for fname in ["bw", "map_adapt", "mk_s2sendump", "mllr_solve"]:
old_path = os.path.join("/usr/local/libexec/sphinxtrain/", fname)
new_path = os.path.join(self.tuned_path, fname)
run_cmds(f"cp {old_path} {new_path}")
def convert_mdef(self):
#run_cmds("sudo apt -y install pocketsphinx")
run_cmds([
f"cd {self.tuned_path}",
"git clone [email protected]:cmusphinx/pocketsphinx.git",
"cd pocketsphinx", "./autogen.sh", f"make -j {cpu_count()}",
"sudo make install"
],
stdout=True)
tool_path = os.path.join(self.tuned_path, "pocketsphinx", "src",
"programs", "pocketsphinx_mdef_convert")
path = os.path.join(self.tuned_path, "en-us/mdef")
run_cmds(f"{tool_path} -text {path} {path}.txt", stdout=True)
def install_sphinx_base(self):
# https://bangladroid.wordpress.com/2017/02/16/installing-cmu-sphinx-on-ubuntu/
run_cmds("sudo apt-get install -y gcc automake autoconf libtool "
"bison swig python-dev libpulse-dev")
sphinx_path = os.path.join(self.tuned_path, "sphinx-src")
makedirs(sphinx_path, remake=True)
url = "https://github.com/cmusphinx/sphinxbase.git"
# sudo is used on the first command to ensure it's use
run_cmds([
f"sudo ls ", f"cd {sphinx_path}", f"git clone {url}",
"cd sphinxbase", "./autogen.sh", f"make -j {cpu_count()}",
"sudo make install",
f"cp src/sphinx_fe/sphinx_fe {self.tuned_path}"
],
stdout=True)
def _install_chromedriver(self):
path = '/'.join(self.driver_path.split('/')[:-1])
if os.path.exists(self.driver_path):
delete_paths(self.driver_path)
# Installs chromedriver
# https://gist.github.com/mikesmullin/2636776#gistcomment-2608206
cmd = ("LATEST_VERSION=$(curl -s "
"https://chromedriver.storage.googleapis.com/LATEST_RELEASE) &&"
" wget -O /tmp/chromedriver.zip "
"https://chromedriver.storage.googleapis.com/$LATEST_VERSION/"
"chromedriver_linux64.zip && "
"unzip /tmp/chromedriver.zip "
f"chromedriver -d {path};")
run_cmds([cmd])
def download_proper_en(self):
url = ("https://phoenixnap.dl.sourceforge.net/project/cmusphinx/"
"Acoustic%20and%20Language%20Models/US%20English/"
"cmusphinx-en-us-5.2.tar.gz")
path = os.path.join(self.tuned_path, "larger_sphinx.tar.gz")
logging.info("downloading file, this may take a while")
download_file(url, path)
logging.info("downloaded")
with tarfile.open(path) as f:
old_en_us_path = os.path.join(self.tuned_path, "en-us")
delete_paths(old_en_us_path)
f.extractall(old_en_us_path)
run_cmds([
f"cd {old_en_us_path}", f"mv * old_folder", "cd old_folder",
"mv * ..", "rm -rf old_folder"
],
stdout=True)
def run_bw(self):
run_cmds(
[
f"cd {self.tuned_path}",
(
"sudo ./bw \\\n"
" -hmmdir en-us \\\n"
" -moddeffn en-us/mdef.txt \\\n"
#"-ts2cbfn .ptm. \\\n"
" -ts2cbfn .cont. \\\n"
" -feat 1s_c_d_dd \\\n"
#"-svspec 0-12/13-25/26-38 \\\n"
" -lda en-us/feature_transform \\\n"
" -cmn current \\\n"
" -agc none \\\n"
" -dictfn cmudict-en-us.dict \\\n"
f" -ctlfn {self.file_ids_name} \\\n"
f" -lsnfn {self.transcription_name} \\\n"
" -accumdir .")
],
stdout=True)
def run_test_decoder(self):
for adapt in [False, True]:
if adapt:
run_cmds([
f"cd {self.tuned_path}", f"rm -rf test/en-us",
f"cp -R en-us-adapt test/en-us"
])
tool_path = os.path.join(self.tuned_path, "pocketsphinx", "src",
"programs", "pocketsphinx_batch")
run_cmds(
(
f"cd {self.test_dir} && \\\n"
f"{tool_path} \\\n"
f" -adcin yes \\\n"
f" -cepdir wav \\\n"
f" -cepext .wav \\\n"
f" -ctl test.fileids \\\n"
f" -lm en-us.lm.bin \\\n"
f" -dict cmudict-en-us.dict \\\n"
f" -hmm en-us \\\n" # for example en-us
f" -hyp test.hyp"),
stdout=True)
run_cmds([
f"cd {self.test_dir}",
"perl word_align.pl test.transcription test.hyp"
],
stdout=True)
input(f"test complete with adapt as {adapt}, hit enter")
def _install_google_chrome(self):
run_cmds(["sudo apt-get update -y"])
run_cmds(["sudo apt-get upgrade -y"])
chrome_install_base = "/tmp"
chrome_install_name = "google-chrome-stable_current_amd64.deb"
chrome_install_path = os.path.join(chrome_install_base,
chrome_install_name)
if os.path.exists(chrome_install_path):
delete_paths(chrome_install_path)
run_cmds([
f"cd {chrome_install_base}",
("wget https://dl.google.com/linux/direct/"
f"{chrome_install_name}"),
f"sudo apt install ./{chrome_install_name}"
])
def run(self):
run_cmds("sudo apt -y install pocketsphinx")
self.generate_new_model()
self.test_new_model()
input("backup audio in /etc/audio then press enter")