def get_segs_for_file(self, audio_file, sample_rate):
base_name = strip_ext(os.path.basename(audio_file))
possible_files = [x for x in self.seg_files if base_name in x]
res = []
if len(possible_files) > 0:
seg_file = os.path.join(self.root_dir, possible_files[0])
if len(possible_files) > 1:
get_logger().log(
logging.WARNING,
"Found multiple matches for %s (%s). Using %s" %
(audio_file, " ".join(possible_files), seg_file))
res = get_eaf_segs(seg_file, sample_rate, base_name,
self.add_labels)
else:
get_logger().log(
logging.WARNING,
"No seg file found for % in %s" % (audio_file, self.root_dir))
return res
def run(self, filepath):
self.log(logging.INFO, "Starting %s" % (filepath))
out_file = self.derive_new_file_path(filepath, ".csv")
if file_utils.should_run(filepath, out_file):
t = transcript.PlaintextTranscript(
filepath=filepath,
label=None,
pos_tagger_path=self.pos_tagger_path)
transcript_utterances_fillers = None
if self.filler_dir:
file_id = os.path.basename(file_utils.strip_ext(filepath))
if file_id in self.filler_files:
filler_file = os.path.join(self.filler_dir,
self.filler_files[file_id])
filler_transcript = transcript.PlaintextTranscript(
filepath=filler_file,
label=None,
pos_tagger_path=self.pos_tagger_path)
transcript_utterances_fillers = filler_transcript.tokens
self.feature_extractor.extract(
t,
out_csv=out_file,
transcript_utterances_fillers=transcript_utterances_fillers)
self.log(logging.INFO, "Done %s -> %s" % (filepath, out_file))
self.emit(out_file)
def derive_new_file_path(self, old_file, new_ext=None):
old_fname = os.path.basename(old_file)
if new_ext is not None:
ext = new_ext
if not ext.startswith("."):
ext = "." + ext
new_fname = file_utils.strip_ext(old_fname) + ext
else:
new_fname = old_fname
return os.path.join(self.out_dir, new_fname)
def setup(self, cfg_file, utterance_sep=" . ", filler_dir=None):
path_output_parses = os.path.join(self.out_dir, "stanford_parses")
path_output_lu_parses = os.path.join(self.out_dir, "lu_parses")
path_output_rst = os.path.join(self.out_dir, "rst_output")
self.output_csv = os.path.join(
self.out_dir, 'textfeatures%s.csv' %
(datetime.datetime.strftime(datetime.datetime.now(), "%Y%m%d")))
do_wnic = True
lexical_list, do_lexical, pragmatic_list, do_pragmatic, semantic_list, do_semantic, syntactic_list, do_syntactic = load_conf(
cfg_file)
parser_path = config.stanford_parser_path
pos_tagger_path = config.stanford_pos_path
lu_analyzer_path = config.lu_analyzer_path
path_to_stanford_cp = config.path_to_stanford_cp
cfg_rules_path = config.cfg_rules_path
path_to_dictionary = config.path_to_dictionary
path_to_freq_norms = config.path_to_freq_norms
path_to_image_norms = config.path_to_image_norms
path_to_anew = config.path_to_anew
path_to_warringer = config.path_to_warringer
path_to_mpqa_lexicon = config.path_to_mpqa_lexicon
path_to_rst_python = config.path_to_rst_python
path_to_rst = config.path_to_rst
path_to_lda_model = config.path_to_lda_model
path_to_lda_wordids = config.path_to_lda_wordids
self.filler_dir = filler_dir
self.pos_tagger_path = pos_tagger_path
self.feature_extractor = feature.FeatureExtractor(
utterance_sep=utterance_sep,
path_output_lu_parses=path_output_lu_parses,
path_output_parses=path_output_parses,
parser_path=parser_path,
cfg_rules_path=cfg_rules_path,
pos_tagger_path=pos_tagger_path,
path_to_dictionary=path_to_dictionary,
lu_analyzer_path=lu_analyzer_path,
path_to_freq_norms=path_to_freq_norms,
path_to_image_norms=path_to_image_norms,
path_to_anew=path_to_anew,
path_to_warringer=path_to_warringer,
do_wnic=do_wnic,
path_to_mpqa_lexicon=path_to_mpqa_lexicon,
path_to_rst_python=path_to_rst_python,
path_to_rst=path_to_rst,
path_output_rst=path_output_rst,
path_to_stanford_cp=path_to_stanford_cp,
path_to_lda_model=path_to_lda_model,
path_to_lda_wordids=path_to_lda_wordids,
do_lexical=do_lexical,
do_syntactic=do_syntactic,
do_semantic=do_semantic,
do_pragmatic=do_pragmatic,
lexical_list=lexical_list,
syntactic_list=syntactic_list,
semantic_list=semantic_list,
pragmatic_list=pragmatic_list)
self.filler_files = {
os.path.basename(file_utils.strip_ext(x)): x
for x in os.listdir(self.filler_dir)
}