File size: 4,287 Bytes
2bbabc3
 
 
 
 
 
 
 
 
 
 
 
 
9f383ea
7481e8e
2bbabc3
 
 
f88a8c0
7481e8e
 
 
51e442c
7481e8e
2bbabc3
 
 
 
 
 
 
 
51e442c
 
 
2bbabc3
 
51e442c
 
 
2bbabc3
 
 
7481e8e
 
 
 
 
51e442c
 
 
7481e8e
51e442c
2bbabc3
51e442c
2bbabc3
 
 
 
 
 
 
 
51e442c
2bbabc3
51e442c
 
 
 
 
 
2bbabc3
51e442c
 
 
2bbabc3
 
51e442c
 
 
 
 
 
2bbabc3
 
 
 
 
9f383ea
2bbabc3
 
 
 
51e442c
2bbabc3
 
 
 
51e442c
 
 
 
 
 
2bbabc3
 
 
 
 
51e442c
 
2bbabc3
 
51e442c
 
da83510
51e442c
 
 
 
2bbabc3
 
 
 
 
 
 
 
 
 
 
 
 
51e442c
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
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
import tempfile
import edge_tts
import gradio as gr
import asyncio

language_dict = {
    "Amharic": {
        "Ameha": "am-ET-AmehaNeural",
        "Mekdes": "am-ET-MekdesNeural"
    },
    "English": {
        "Ryan": "en-GB-RyanNeural",
        "Clara": "en-CA-ClaraNeural"
    },
    "Tigrinya": {}  # Empty because we'll use Amharic voices with labels
}

async def text_to_speech_edge(text, language, speaker):
    if language == "Tigrinya":
        clean_speaker = speaker.replace(" (Amharic Voice)", "")
        voice = language_dict["Amharic"][clean_speaker]
    else:
        voice = language_dict[language][speaker]
    
    try:
        communicate = edge_tts.Communicate(text, voice)
        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
            tmp_path = tmp_file.name
            await asyncio.wait_for(communicate.save(tmp_path), timeout=30)
        return tmp_path
        
    except asyncio.TimeoutError:
        error_msg = ("αˆ΅αˆ…α‰°α‰΅: αŒŠα‹œ αŠ αˆα‰‹αˆα’ αŠ₯α‰£αŠ­α‹Ž αŠ₯αŠ•α‹°αŒˆαŠ“ α‹­αˆžαŠ­αˆ©α’"
                    if language in ["Amharic", "Tigrinya"]
                    else "Error: Timeout. Please try again.")
        raise gr.Error(error_msg)
    except Exception as e:
        error_msg = (f"αˆ΅αˆ…α‰°α‰΅: {str(e)}"
                    if language in ["Amharic", "Tigrinya"]
                    else f"Error: {str(e)}")
        raise gr.Error(error_msg)

def update_speakers(language):
    if language == "Tigrinya":
        speakers = [f"{name} (Amharic Voice)" for name in language_dict["Amharic"].keys()]
    else:
        speakers = list(language_dict[language].keys())
    
    return gr.Dropdown(
        choices=speakers,
        value=speakers[0],
        label=f"Select Speaker {'(Amharic Voices)' if language == 'Tigrinya' else ''}"
    )

with gr.Blocks(title="Amharic, English & Tigrinya TTS", theme=gr.themes.Soft()) as demo:
    gr.HTML("""
    <style>
        h1 { 
            color: #2E86C1; 
            text-align: center;
            background: linear-gradient(45deg, #FF007F, #2E86C1);
            -webkit-background-clip: text;
            -webkit-text-fill-color: transparent;
            margin-bottom: 20px;
        }
        .notice {
            font-size: 0.9em;
            color: #666;
            text-align: center;
            margin: 10px 0;
            font-style: italic;
        }
        .gradio-button {
            background: linear-gradient(45deg, #FF007F, #2E86C1) !important;
            color: white !important;
        }
    </style>
    <center>
        <h1>Amharic, English & Tigrinya Text-to-Speech</h1>
        <div class="notice">
            Note: Tigrinya uses Amharic-accented voices until dedicated models become available
        </div>
    </center>
    """)

    with gr.Row():
        with gr.Column():
            language = gr.Dropdown(
                choices=["Amharic", "English", "Tigrinya"],
                value="Amharic",
                label="Select Language / α‰‹αŠ•α‰‹ α‹­αˆαˆ¨αŒ‘"
            )
            input_text = gr.Textbox(
                lines=5,
                label="Enter Text / αŒ½αˆ‘α α‹«αˆ΅αŒˆα‰‘",
                placeholder="Type your text here... / αŒ½αˆ‘αα‹ŽαŠ• α‹­αŒ»α‰..."
            )
            speaker = gr.Dropdown(
                label="Select Speaker / αŠ αˆ­α‰²αˆ΅α‰΅ α‹­αˆαˆ¨αŒ‘",
                interactive=True
            )
            run_btn = gr.Button(
                value="Generate Audio / α‹΅αˆα… ፍጠር",
                variant="primary"
            )

        with gr.Column():
            output_audio = gr.Audio(
                type="filepath",
                label="Generated Audio / α‹¨α‰°αˆαŒ αˆ¨ α‹΅αˆα…",
                interactive=False
            )

    # Initialize speakers dropdown
    demo.load(
        fn=lambda: gr.update(choices=list(language_dict["Amharic"].keys())),
        outputs=speaker
    )

    # Update speakers when language changes
    language.change(
        update_speakers,
        inputs=language,
        outputs=speaker
    )

    run_btn.click(
        text_to_speech_edge,
        inputs=[input_text, language, speaker],
        outputs=output_audio
    )

if __name__ == "__main__":
    demo.launch(server_port=7860, share=False)