def praat_syllable_nuclei(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
praat = audio.PraatRunner("praat_syllable_nuclei", out_dir=out_folder)
p = ProgressPipeline(file_finder | praat, n_threads=num_threads, quiet=True)
return p
def lex(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".txt")
feats = lexicosyntactic.Lexicosyntactic("lexicosyntactic", out_dir=out_folder, cfg_file="default.conf")
p = ProgressPipeline(file_finder | feats, n_threads=num_threads, quiet=True)
return p
def kaldi_asr(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
resample = audio.ResampleWav("resampled_audio", out_dir=out_folder, new_sr=8000)
asr = audio.KaldiASR("kaldi_asr", out_dir=out_folder)
p = ProgressPipeline(file_finder | resample | asr, n_threads=num_threads, quiet=True)
return p
def split_speech_eaf(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
splitter = EafSegments(in_folder)
split = audio.SplitSegments("split_speech", out_dir=out_folder, segment_mapping_fn=splitter.get_segs_for_file)
p = ProgressPipeline(file_finder | split, n_threads=num_threads, quiet=True)
return p
def opensmile_is10(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
is10 = audio.OpenSmileRunner("is10", out_dir=out_folder, conf_file="IS10_paraling.conf", out_flag="-csvoutput")
p = ProgressPipeline(file_finder | is10, n_threads=num_threads, quiet=True)
return p
def matlab(in_folder, out_folder, num_threads):
from nodes import matlab as mtlb
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
is10 = mtlb.MatlabRunner("matlab_acoustics", out_dir=out_folder, function="extract_acoustics", out_ext=".txt")
p = ProgressPipeline(file_finder | is10, n_threads=num_threads, quiet=True)
return p
def lex_chinese(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".txt")
feats = lexicosyntactic_multi.ChineseLex("lex_chinese", out_dir=out_folder)
p = ProgressPipeline(file_finder | feats,
n_threads=num_threads,
quiet=True,
exec_name="ParallelExecutor2")
return p
def main(in_folder, out_folder, num_threads):
file_finder = helper.FindFiles("file_finder", dir=in_folder, ext=".wav")
resample = audio.ResampleWav("resampled_audio",
out_dir=out_folder,
new_sr=8000)
is10_lld = audio.OpenSmileRunner("is10_lld",
out_dir=out_folder,
conf_file="IS10_paraling.conf",
out_flag="-lldcsvoutput")
is10 = audio.OpenSmileRunner("is10",
out_dir=out_folder,
conf_file="IS10_paraling.conf",
out_flag="-csvoutput")
splitter = EafSegments(in_folder)
split = audio.SplitSegments("utterances",
out_dir=out_folder,
segment_mapping_fn=splitter.get_segs_for_file,
out_streams=["audio", "label"])
is10_lld_per_utterance = audio.OpenSmileRunner(
"is10_lld_per_utterance",
out_dir=out_folder,
in_streams="audio",
conf_file="IS10_paraling.conf",
out_flag="-lldcsvoutput")
is10_per_utterance = audio.OpenSmileRunner("is10_per_utterance",
out_dir=out_folder,
in_streams="audio",
conf_file="IS10_paraling.conf",
out_flag="-csvoutput")
asr = audio.KaldiASR("kaldi_asr", out_dir=out_folder, in_streams="audio")
p = ProgressPipeline(file_finder | resample | [
is10_lld, is10,
split | [asr, is10_lld_per_utterance, is10_per_utterance]
],
n_threads=num_threads,
quiet=True)
return p