Fix lang_seq binarizing error

yqzhishen · yqzhishen · commit 690bf32fb8fc · 2026-01-08T15:56:07.000+08:00
diff --git a/preprocessing/acoustic_binarizer.py b/preprocessing/acoustic_binarizer.py
@@ -92,7 +92,7 @@ def load_meta_data(self, raw_data_dir: pathlib.Path, ds_id, spk, lang):
                     'lang_seq': [
                         (
                             self.lang_map[lang if '/' not in p else p.split('/', maxsplit=1)[0]]
-                            if self.phoneme_dictionary.is_cross_lingual(p)
+                            if self.phoneme_dictionary.is_cross_lingual(p if '/' in p else f'{lang}/{p}')
                             else 0
                         )
                         for p in utterance_label['ph_seq'].split()
diff --git a/preprocessing/variance_binarizer.py b/preprocessing/variance_binarizer.py
@@ -153,7 +153,7 @@ def require(attr, optional=False):
                     'lang_seq': [
                         (
                             self.lang_map[lang if '/' not in p else p.split('/', maxsplit=1)[0]]
-                            if self.phoneme_dictionary.is_cross_lingual(p)
+                            if self.phoneme_dictionary.is_cross_lingual(p if '/' in p else f'{lang}/{p}')
                             else 0
                         )
                         for p in utterance_label['ph_seq'].split()

Original file line number	Diff line number	Diff line change
`@@ -92,7 +92,7 @@ def load_meta_data(self, raw_data_dir: pathlib.Path, ds_id, spk, lang):`
`92`	`92`	`'lang_seq': [`
`93`	`93`	`(`
`94`	`94`	`self.lang_map[lang if '/' not in p else p.split('/', maxsplit=1)[0]]`
`95`		`- if self.phoneme_dictionary.is_cross_lingual(p)`
	`95`	`+ if self.phoneme_dictionary.is_cross_lingual(p if '/' in p else f'{lang}/{p}')`
`96`	`96`	`else 0`
`97`	`97`	`)`
`98`	`98`	`for p in utterance_label['ph_seq'].split()`
Original file line number	Diff line number	Diff line change
`@@ -153,7 +153,7 @@ def require(attr, optional=False):`
`153`	`153`	`'lang_seq': [`
`154`	`154`	`(`
`155`	`155`	`self.lang_map[lang if '/' not in p else p.split('/', maxsplit=1)[0]]`
`156`		`- if self.phoneme_dictionary.is_cross_lingual(p)`
	`156`	`+ if self.phoneme_dictionary.is_cross_lingual(p if '/' in p else f'{lang}/{p}')`
`157`	`157`	`else 0`
`158`	`158`	`)`
`159`	`159`	`for p in utterance_label['ph_seq'].split()`