MCPcopy Index your code
hub / github.com/modelscope/FunASR / sentence_postprocess

Function sentence_postprocess

funasr/utils/postprocess_utils.py:165–278  ·  view source on GitHub ↗

Sentence postprocess. Args: words: TODO. time_stamp: TODO.

(words: List[Any], time_stamp: List[List] = None)

Source from the content-addressed store, hash-verified

163
164
165def sentence_postprocess(words: List[Any], time_stamp: List[List] = None):
166 """Sentence postprocess.
167
168 Args:
169 words: TODO.
170 time_stamp: TODO.
171 """
172 middle_lists = []
173 word_lists = []
174 word_item = ""
175 ts_lists = []
176
177 # wash words lists
178 for i in words:
179 word = ""
180 if isinstance(i, str):
181 word = i
182 else:
183 word = i.decode("utf-8")
184
185 if word in ["<s>", "</s>", "<unk>", "<OOV>"]:
186 continue
187 else:
188 middle_lists.append(word)
189
190 # all chinese characters
191 if isAllChinese(middle_lists):
192 for i, ch in enumerate(middle_lists):
193 word_lists.append(ch.replace(" ", ""))
194 if time_stamp is not None:
195 ts_lists = time_stamp
196
197 # all alpha characters
198 elif isAllAlpha(middle_lists):
199 ts_flag = True
200 for i, ch in enumerate(middle_lists):
201 if ts_flag and time_stamp is not None:
202 begin = time_stamp[i][0]
203 end = time_stamp[i][1]
204 word = ""
205 if "@@" in ch:
206 word = ch.replace("@@", "")
207 word_item += word
208 if time_stamp is not None:
209 ts_flag = False
210 end = time_stamp[i][1]
211 else:
212 word_item += ch
213 word_lists.append(word_item)
214 word_lists.append(" ")
215 word_item = ""
216 if time_stamp is not None:
217 ts_flag = True
218 end = time_stamp[i][1]
219 ts_lists.append([begin, end])
220 begin = end
221
222 # mix characters

Callers

nothing calls this directly

Calls 4

isAllChineseFunction · 0.70
isAllAlphaFunction · 0.70
abbr_disposeFunction · 0.70
decodeMethod · 0.45

Tested by

no test coverage detected

Used in the wild real call sites across dependent graphs

searching dependent graphs…