add condition check if a ssml input and filter space line, test=tts

paddlespeech/t2s/exps/syn_utils.py

@@ -105,14 +105,15 @@ def get_sentences(text_file: Optional[os.PathLike], lang: str='zh'):
     sentences = []
     with open(text_file, 'rt') as f:
         for line in f:
-            items = re.split(r"\s+", line.strip(), 1)
-            utt_id = items[0]
-            if lang == 'zh':
-                sentence = "".join(items[1:])
-            elif lang == 'en':
-                sentence = " ".join(items[1:])
-            elif lang == 'mix':
-                sentence = " ".join(items[1:])
+            if line.strip() != "":
+                items = re.split(r"\s+", line.strip(), 1)
+                utt_id = items[0]
+                if lang == 'zh':
+                    sentence = "".join(items[1:])
+                elif lang == 'en':
+                    sentence = " ".join(items[1:])
+                elif lang == 'mix':
+                    sentence = " ".join(items[1:])
             sentences.append((utt_id, sentence))
     return sentences
 
@@ -182,11 +183,20 @@ def run_frontend(frontend: object,
                  to_tensor: bool=True):
     outs = dict()
     if lang == 'zh':
-        input_ids = frontend.get_input_ids_ssml(
-            text,
-            merge_sentences=merge_sentences,
-            get_tone_ids=get_tone_ids,
-            to_tensor=to_tensor)
+        input_ids = {}
+        if text.strip() != "" and re.match(r".*?<speak>.*?</speak>.*", text,
+                                           re.DOTALL):
+            input_ids = frontend.get_input_ids_ssml(
+                text,
+                merge_sentences=merge_sentences,
+                get_tone_ids=get_tone_ids,
+                to_tensor=to_tensor)
+        else:
+            input_ids = frontend.get_input_ids(
+                text,
+                merge_sentences=merge_sentences,
+                get_tone_ids=get_tone_ids,
+                to_tensor=to_tensor)
         phone_ids = input_ids["phone_ids"]
         if get_tone_ids:
             tone_ids = input_ids["tone_ids"]