File size: 2,487 Bytes
fd43dfa
039f896
fd43dfa
 
 
039f896
fd43dfa
039f896
fd43dfa
 
 
039f896
fd43dfa
 
 
280807d
fd43dfa
 
 
 
 
 
 
 
 
 
 
996224c
fd43dfa
 
 
 
59b69bc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fd43dfa
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
import gradio as gr
import os
import asyncio
from conver import ConversationConfig, URLToAudioConverter
from dotenv import load_dotenv

load_dotenv()

async def synthesize(article_url, text_input, language="en"):
    if not article_url and not text_input:
        return "Error: Ingresa una URL o texto", None

    try:
        config = ConversationConfig()
        converter = URLToAudioConverter(config, llm_api_key=os.environ.get("TOGETHER_API_KEY"))
        
        # Voces humanizadas
        voices = {
            "en": ("en-US-AvaMultilingualNeural", "en-US-AndrewMultilingualNeural"),
            "es": ("es-ES-AlvaroNeural", "es-ES-ElviraNeural")
        }
        voice1, voice2 = voices.get(language, voices["en"])

        if text_input:
            output_file, conversation = await converter.text_to_audio(text_input, voice1, voice2)
        else:
            output_file, conversation = await converter.url_to_audio(article_url, voice1, voice2)
            
        return conversation, output_file
    except Exception as e:
        return f"Error: {str(e)}", None

async def synthesize(article_url, text_input, language="en", skip_llm=False):
    if not article_url and not text_input:
        return "Error: Ingresa una URL o texto", None

    try:
        config = ConversationConfig()
        converter = URLToAudioConverter(config, llm_api_key=os.environ.get("TOGETHER_API_KEY"))
        
        voices = {
            "en": ("en-US-AvaMultilingualNeural", "en-US-AndrewMultilingualNeural"),
            "es": ("es-ES-AlvaroNeural", "es-ES-ElviraNeural")
        }
        voice1, voice2 = voices.get(language, voices["en"])

        # Modo sin LLM (texto exacto)
        if skip_llm and text_input:
            return await converter.raw_text_to_audio(text_input, voice1, voice2)
            
        # Procesamiento normal (con LLM)
        if text_input:
            output_file, conversation = await converter.text_to_audio(text_input, voice1, voice2)
        else:
            output_file, conversation = await converter.url_to_audio(article_url, voice1, voice2)
            
        return conversation, output_file
    except Exception as e:
        return f"Error: {str(e)}", None

# Añade esto en tu interfaz (dentro de with gr.Blocks()):
skip_llm = gr.Checkbox(label="🔴 Modo libre (sin filtros LLM)", value=False)
btn.click(
    synthesize_sync,
    inputs=[text_url, text_input, language, skip_llm],
    outputs=[conv_display, aud]
)

demo.launch()