Some stuff

interfaces/__init__.py

@@ -0,0 +1 @@
+from .gradio_ui import GenerateUI

interfaces/gradio_ui.py

@@ -0,0 +1,51 @@
+import importlib
+import json
+import os
+import gradio as gr
+from src import engines
+
+class GenerateUI:
+    def __init__(self):
+        self.engines = self.get_engines()
+
+    def get_engines(self):
+        engines_d = {}
+        with open(os.path.join(os.getcwd(), "src", "engines", "engines.json"), "r") as f:
+            engine_types = json.load(f)
+        for engine_type, engine_list in engine_types.items():
+            engines_d[engine_type] = {}
+            for engine_name in engine_list:
+                module = importlib.import_module(f"src.engines.{engine_type}.{engine_name}")
+                engine_class = getattr(module, engine_name)
+                engines_d[engine_type][engine_name] = engine_class
+        return engines_d
+
+    def launch_ui(self):
+        with gr.Blocks() as main_block:
+            for engine_type, engines in self.engines.items():
+                switch_dropdown = gr.Dropdown(list(engines.keys()), label=engine_type)
+                engine_blocks = []
+
+                for engine_name, engine_class in engines.items():
+                    with gr.Blocks(elem_id=f"{engine_type}_{engine_name}_block", visible=False) as engine_block:
+                        options = engine_class().get_options()
+                        for option in options:
+                            engine_block.add(option)
+                        engine_blocks.append(engine_block)
+
+                def switch_engine(engine_name, engine_blocks=engine_blocks, switch_dropdown=switch_dropdown):
+                    for block in engine_blocks:
+                        block.visible = block.elem_id.startswith(f"{engine_type}_{engine_name}")
+
+                switch_dropdown.change(switch_engine, inputs=[switch_dropdown], outputs=engine_blocks)
+
+                # Initially show the first engine's options
+                if engines:
+                    first_engine_name = list(engines.keys())[0]
+                    switch_engine(first_engine_name)
+
+        main_block.launch()
+
+if __name__ == "__main__":
+    ui_generator = GenerateUI()
+    ui_generator.launch_ui()

main.py

@@ -0,0 +1,5 @@
+from interfaces import GenerateUI
+
+if __name__ == "__main__":
+    ui_generator = GenerateUI()
+    ui_generator.launch_ui()

src/__init__.py

@@ -0,0 +1 @@
+from . import engines

src/engines/BaseEngine.py

@@ -0,0 +1,10 @@
+from abc import ABC, abstractmethod
+import gradio as gr
+
+class BaseEngine(ABC):
+    options: list
+    name: str
+    description: str
+
+    def __init__(self):
+        pass

src/engines/TTSEngine/BaseTTSEngine.py

@@ -0,0 +1,10 @@
+from abc import ABC, abstractmethod
+from ..BaseEngine import BaseEngine
+
+
+class BaseTTSEngine(BaseEngine):
+    pass
+
+    @abstractmethod
+    def synthesize(self, text: str, path: str) -> str:
+        pass

src/engines/TTSEngine/CoquiTTSEngine.py

@@ -0,0 +1,111 @@
+import gradio as gr
+import TTS
+import os
+import torch
+
+from .BaseTTSEngine import BaseTTSEngine
+
+
+class CoquiTTSEngine(BaseTTSEngine):
+    voices = [
+        "Claribel Dervla",
+        "Daisy Studious",
+        "Gracie Wise",
+        "Tammie Ema",
+        "Alison Dietlinde",
+        "Ana Florence",
+        "Annmarie Nele",
+        "Asya Anara",
+        "Brenda Stern",
+        "Gitta Nikolina",
+        "Henriette Usha",
+        "Sofia Hellen",
+        "Tammy Grit",
+        "Tanja Adelina",
+        "Vjollca Johnnie",
+        "Andrew Chipper",
+        "Badr Odhiambo",
+        "Dionisio Schuyler",
+        "Royston Min",
+        "Viktor Eka",
+        "Abrahan Mack",
+        "Adde Michal",
+        "Baldur Sanjin",
+        "Craig Gutsy",
+        "Damien Black",
+        "Gilberto Mathias",
+        "Ilkin Urbano",
+        "Kazuhiko Atallah",
+        "Ludvig Milivoj",
+        "Suad Qasim",
+        "Torcull Diarmuid",
+        "Viktor Menelaos",
+        "Zacharie Aimilios",
+        "Nova Hogarth",
+        "Maja Ruoho",
+        "Uta Obando",
+        "Lidiya Szekeres",
+        "Chandra MacFarland",
+        "Szofi Granger",
+        "Camilla Holmström",
+        "Lilya Stainthorpe",
+        "Zofija Kendrick",
+        "Narelle Moon",
+        "Barbora MacLean",
+        "Alexandra Hisakawa",
+        "Alma María",
+        "Rosemary Okafor",
+        "Ige Behringer",
+        "Filip Traverse",
+        "Damjan Chapman",
+        "Wulf Carlevaro",
+        "Aaron Dreschner",
+        "Kumar Dahl",
+        "Eugenio Mataracı",
+        "Ferran Simen",
+        "Xavier Hayasaka",
+        "Luis Moray",
+        "Marcos Rudaski",
+    ]
+    name = "Coqui TTS"
+    description = "Coqui TTS engine."
+    languages = [
+        "en",  # English
+        "es",  # Spanish
+        "fr",  # French
+        "de",  # German
+        "it",  # Italian
+        "pt",  # Portuguese
+        "pl",  # Polish
+        "tr",  # Turkish
+        "ru",  # Russian
+        "nl",  # Dutch
+        "cs",  # Czech
+        "ar",  # Arabic
+        "zh-cn",  # Chinese (Simplified)
+        "ja",  # Japanese
+        "hu",  # Hungarian
+        "ko",  # Korean
+        "hi",  # Hindi
+    ]
+    options = [
+        gr.Dropdown(
+            voices, value=voices[0], label="voice", max_choices=1
+        ),
+        gr.Dropwdown(
+            languages, value=languages[0], label="language", max_choices=1
+        ),
+    ]
+    def __init__(self):
+        super().__init__()
+
+        os.environ["COQUI_TOS_AGREED"] = "1"
+        self.tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2")
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        self.tts.to(device)
+
+    def synthesize(self, text: str, path: str) -> str:
+        voice = self.options[0].value
+        language = self.options[1].value
+        self.tts.tts_to_file(text=text, file_path=path, lang=language, speaker=voice)
+        return path

src/engines/TTSEngine/ElevenLabsTTSEngine.py

@@ -0,0 +1,13 @@
+from .BaseTTSEngine import AbstractTTSEngine
+import gradio as gr
+
+class ElevenLabsTTSEngine(AbstractTTSEngine):
+    options = [gr.Radio(["Neutral", "Happy", "Sad"], label="emotion")]
+    name = "ElevenLabs"
+    description = "ElevenLabs TTS engine."
+
+    def __init__(self):
+        super().__init__()
+
+    def synthesize(self, text: str, path: str) -> str:
+        pass

src/engines/TTSEngine/__init__.py

@@ -0,0 +1,3 @@
+from .BaseTTSEngine import BaseTTSEngine
+from .CoquiTTSEngine import CoquiTTSEngine
+from .ElevenLabsTTSEngine import ElevenLabsTTSEngine

src/engines/__init__.py

@@ -0,0 +1,2 @@
+from . import TTSEngine
+from .BaseEngine import BaseEngine

src/engines/engines.json

@@ -0,0 +1,6 @@
+{
+    "TTSEngine": [
+        "CoquiTTSEngine",
+        "ElevenLabsTTSEngine"
+    ]
+}

src/engines/script/AbstractScriptEngine.py

@@ -0,0 +1,4 @@
+from src.engines.BaseEngine import AbstractEngine
+
+class AbstractScriptEngine(AbstractEngine):
+    pass

src/engines/script/ComicalScriptEngine.py

@@ -0,0 +1,6 @@
+from AbstractScriptEngine import AbstractScriptEngine
+
+class ComicalScriptEngine(AbstractScriptEngine):
+    def __init__(self):
+        super().__init__()
+        self.options = {"comicality": ["Low", "Medium", "High"]}

src/engines/script/DrammaticScriptEngine.py

@@ -0,0 +1,6 @@
+from AbstractScriptEngine import AbstractScriptEngine
+
+class DramaticScriptEngine(AbstractScriptEngine):
+    def __init__(self):
+        super().__init__()
+        self.options = {"tone": ["Serious", "Light-hearted"]}