python/FunASR-XL.git

			@@ -1,10 +1,12 @@
			from scipy.fftpack import shift
			from itertools import zip_longest

			import torch
			import copy
			import codecs
			import logging
			import edit_distance
			import argparse
			import pdb
			import numpy as np
			from typing import Any, List, Tuple, Union

			@@ -13,10 +15,11 @@
			us_peaks,
			char_list,
			vad_offset=0.0,
			force_time_shift=-1.5
			force_time_shift=-1.5,
			sil_in_str=True
			):
			if not len(char_list):
			return []
			return "", []
			START_END_THRESHOLD = 5
			MAX_TOKEN_DURATION = 12
			TIME_RATE = 10.0 * 6 / 1000 / 3 # 3 times upsampled
			@@ -66,6 +69,8 @@
			timestamp_list[i][1] = timestamp_list[i][1] + vad_offset / 1000.0
			res_txt = ""
			for char, timestamp in zip(new_char_list, timestamp_list):
			#if char != '<sil>':
			if not sil_in_str and char == '<sil>': continue
			res_txt += "{} {} {};".format(char, str(timestamp[0]+0.0005)[:5], str(timestamp[1]+0.0005)[:5])
			res = []
			for char, timestamp in zip(new_char_list, timestamp_list):
			@@ -84,6 +89,7 @@
			return res
			if len(text_postprocessed) == 0:
			return res

			if punc_id_list is None or len(punc_id_list) == 0:
			res.append({
			'text': text_postprocessed.split(),
			@@ -92,36 +98,45 @@
			})
			return res
			if len(punc_id_list) != len(time_stamp_postprocessed):
			res.append({
			'text': text_postprocessed.split(),
			"start": time_stamp_postprocessed[0][0],
			"end": time_stamp_postprocessed[-1][1]
			})
			return res

			print(" warning length mistach!!!!!!")
			sentence_text = ''
			sentence_start = time_stamp_postprocessed[0][0]
			sentence_end = time_stamp_postprocessed[0][1]
			texts = text_postprocessed.split()
			for i in range(len(punc_id_list)):
			sentence_text += texts[i]
			if punc_id_list[i] == 2:
			punc_stamp_text_list = list(zip_longest(punc_id_list, time_stamp_postprocessed, texts, fillvalue=None))
			for punc_stamp_text in punc_stamp_text_list:
			punc_id, time_stamp, text = punc_stamp_text
			sentence_text += text if text is not None else ''
			punc_id = int(punc_id) if punc_id is not None else 1
			sentence_end = time_stamp[1] if time_stamp is not None else sentence_end

			if punc_id == 2:
			sentence_text += ','
			res.append({
			'text': sentence_text,
			"start": sentence_start,
			"end": time_stamp_postprocessed[i][1]
			"end": sentence_end
			})
			sentence_text = ''
			sentence_start = time_stamp_postprocessed[i][1]
			elif punc_id_list[i] == 3:
			sentence_start = sentence_end
			elif punc_id == 3:
			sentence_text += '.'
			res.append({
			'text': sentence_text,
			"start": sentence_start,
			"end": time_stamp_postprocessed[i][1]
			"end": sentence_end
			})
			sentence_text = ''
			sentence_start = time_stamp_postprocessed[i][1]
			sentence_start = sentence_end
			elif punc_id == 4:
			sentence_text += '?'
			res.append({
			'text': sentence_text,
			"start": sentence_start,
			"end": sentence_end
			})
			sentence_text = ''
			sentence_start = sentence_end
			return res


			@@ -134,9 +149,9 @@
			uttid_intersection = self._intersection(uttid_list1, uttid_list2)
			res = self.as_cal(uttid_intersection, ts_dict1, ts_dict2)
			logging.warning("Average shift of {} and {}: {}.".format(file1, file2, str(res)[:8]))
			logging.warning("Following timestamp pair differs most: {}, detail:{}".format(self.max_shift_uttid))
			logging.warning("Following timestamp pair differs most: {}, detail:{}".format(self.max_shift, self.max_shift_uttid))

			def _intersection(list1, list2):
			def _intersection(self, list1, list2):
			set1 = set(list1)
			set2 = set(list2)
			if set1 == set2:
			@@ -170,8 +185,9 @@
			return uttid_list, ts_dict

			def _shift(self, filtered_timestamp_list1, filtered_timestamp_list2):
			shift_time = 0
			for fts1, fts2 in zip(filtered_timestamp_list1, filtered_timestamp_list2):
			shift_time = abs(fts1[0] - fts2[0]) + abs(fts1[1] - fts2[1])
			shift_time += abs(fts1[0] - fts2[0]) + abs(fts1[1] - fts2[1])
			num_tokens = len(filtered_timestamp_list1)
			return shift_time, num_tokens

			@@ -232,13 +248,54 @@
			return self._accumlated_shift / self._accumlated_tokens


			SUPPORTED_MODES = ['cal_aas']
			def convert_external_alphas(alphas_file, text_file, output_file):
			from funasr.models.predictor.cif import cif_wo_hidden
			with open(alphas_file, 'r') as f1, open(text_file, 'r') as f2, open(output_file, 'w') as f3:
			for line1, line2 in zip(f1.readlines(), f2.readlines()):
			line1 = line1.rstrip()
			line2 = line2.rstrip()
			assert line1.split()[0] == line2.split()[0]
			uttid = line1.split()[0]
			alphas = [float(i) for i in line1.split()[1:]]
			new_alphas = np.array(remove_chunk_padding(alphas))
			new_alphas[-1] += 1e-4
			text = line2.split()[1:]
			if len(text) + 1 != int(new_alphas.sum()):
			# force resize
			new_alphas *= (len(text) + 1) / int(new_alphas.sum())
			peaks = cif_wo_hidden(torch.Tensor(new_alphas).unsqueeze(0), 1.0-1e-4)
			if " " in text:
			text = text.split()
			else:
			text = [i for i in text]
			res_str, _ = ts_prediction_lfr6_standard(new_alphas, peaks[0], text,
			force_time_shift=-7.0,
			sil_in_str=False)
			f3.write("{} {}\n".format(uttid, res_str))


			def remove_chunk_padding(alphas):
			# remove the padding part in alphas if using chunk paraformer for GPU
			START_ZERO = 45
			MID_ZERO = 75
			REAL_FRAMES = 360 # for chunk based encoder 10-120-10 and fsmn padding 5
			alphas = alphas[START_ZERO:] # remove the padding at beginning
			new_alphas = []
			while True:
			new_alphas = new_alphas + alphas[:REAL_FRAMES]
			alphas = alphas[REAL_FRAMES+MID_ZERO:]
			if len(alphas) < REAL_FRAMES: break
			return new_alphas

			SUPPORTED_MODES = ['cal_aas', 'read_ext_alphas']


			def main(args):
			if args.mode == 'cal_aas':
			asc = AverageShiftCalculator()
			asc(args.input, args.input2)
			elif args.mode == 'read_ext_alphas':
			convert_external_alphas(args.input, args.input2, args.output)
			else:
			logging.error("Mode {} not in SUPPORTED_MODES: {}.".format(args.mode, SUPPORTED_MODES))