Release v0.0.2

app.py

@@ -12,9 +12,10 @@ from torch.cuda import is_available
 
 
 class VoiceOption(Enum):
-    Nuri = "Нурі (жіночий) 👩"
+    Nuri = "Севіль (жіночий) 👩"
     Arslan = "Арслан (чоловічий) 👨"
-    Kemal = "Кемаль (чоловічий) 👨"
+    Kemal = "Ескандер (чоловічий) 👨"
+    Abibulla = "Абібулла (чоловічий) 👨"
 
 
 print(f"CUDA available? {is_available()}")
@@ -37,6 +38,7 @@ def tts(text: str, voice: str):
         VoiceOption.Nuri.value: Voices.Nuri.value,
         VoiceOption.Arslan.value: Voices.Arslan.value,
         VoiceOption.Kemal.value: Voices.Kemal.value,
+        VoiceOption.Abibulla.value: Voices.Abibulla.value,
     }
 
     speaker_name = voice_mapping[voice]
@@ -83,6 +85,7 @@ iface = gr.Interface(
             VoiceOption.Arslan.value,
         ],
         ["Селям! Ишлер насыл?", VoiceOption.Nuri.value],
+        ["Selâm! 123456789", VoiceOption.Abibulla.value],
     ],
 )
 iface.launch()

crh_preprocessor/preprocessor.py

@@ -1,15 +1,26 @@
 import re
 
-mappings = [
-    ["\u0069\u0307", "i"],  # ["i̇", "i"], handle i with dot, first occurence is two separate symbols
-]
+mapping = {
+    "n\u0303": "\xf1",
+    "g\u0306": "\u011f",
+    "i\u0307": "i",
+    "u\u0308": "\xfc",
+    "o\u0308": "\xf6",
+    "\xe7": "\u04ab",
+    "c\u0327": "\u04ab",
+    "s\u0327": "\u015f",
+    "a\u0302": "\xe2",
+    "w": "v",
+    "x": "ks"
+}
+ 
 
 def preprocess(text):
     text = text.lower()  # always treat lowercase
     text = " " + text + " "
 
-    for mapping in mappings:
-        text = re.sub(mapping[0], mapping[1], text)
+    for symbol in mapping.keys():
+        text = re.sub(symbol, mapping[symbol], text)
 
     numbers = {
         "0": "sıfır",

crh_tts/tts.py

@@ -13,6 +13,7 @@ class Voices(Enum):
     Arslan = "arslan"
     Nuri = "nuri"
     Kemal = "kemal"
+    Abibulla = "abibulla"
 
 
 class TTS:

requirements.txt

@@ -1,3 +1,3 @@
 gradio==3.6
 torch>=1.13
-TTS==0.8.0
+TTS==0.9.0