File size: 5,025 Bytes
7e9eac8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
#!/usr/bin/env python3
#
# Copyright      2022-2023  Xiaomi Corp.        (authors: Fangjun Kuang)
#
# See LICENSE for clarification regarding multiple authors
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# References:
# https://gradio.app/docs/#dropdown

import logging
import os
import uuid
from pathlib import Path

import gradio as gr

from decode import decode
from model import get_pretrained_model, get_vad, language_to_models

title = "# Next-gen Kaldi: Text-to-speech (TTS)"

description = """
This space shows how to convert text to speech with Next-gen Kaldi.

It is running on CPU within a docker container provided by Hugging Face.

See more information by visiting the following links:

- <https://github.com/k2-fsa/sherpa-onnx>

If you want to deploy it locally, please see
<https://k2-fsa.github.io/sherpa/>
"""

# css style is copied from
# https://huggingface.co/spaces/alphacep/asr/blob/main/app.py#L113
css = """
.result {display:flex;flex-direction:column}
.result_item {padding:15px;margin-bottom:8px;border-radius:15px;width:100%}
.result_item_success {background-color:mediumaquamarine;color:white;align-self:start}
.result_item_error {background-color:#ff7070;color:white;align-self:start}
"""


def update_model_dropdown(language: str):
    if language in language_to_models:
        choices = language_to_models[language]
        return gr.Dropdown.update(choices=choices, value=choices[0])

    raise ValueError(f"Unsupported language: {language}")


def build_html_output(s: str, style: str = "result_item_success"):
    return f"""
    <div class='result'>
        <div class='result_item {style}'>
          {s}
        </div>
    </div>
    """


def process(language: str, repo_id: str, text: str, sid: str):
    logging.info(f"Input text: {text}. sid: {sid}")
    sid = int(sid)
    tts = get_pretrained_model(repo_id)

    start = time.time()
    audio = tts.generate(text, sid=sid)
    end = time.time()

    if len(audio.samples) == 0:
        raise ValueError(
            "Error in generating audios. Please read previous error messages."
        )

    duration = len(audio.samples) / audio.sample_rate

    elapsed_seconds = end - start
    rtf = elapsed_seconds / duration

    info = f"""
    Wave duration  : {duration:.3f} s <br/>
    Processing time: {elapsed_seconds:.3f} s <br/>
    RTF: {elapsed_seconds:.3f}/{duration:.3f} = {rtf:.3f} <br/>
    """

    logging.info(info)
    logging.info(f"\nrepo_id: {repo_id}\ntext: {text}")

    filename = str(uuid.uuid4())
    filename = f"{filename}.wav"
    sf.write(
        filename,
        audio.samples,
        samplerate=audio.sample_rate,
        subtype="PCM_16",
    )

    return filename, build_html_output(info)


demo = gr.Blocks(css=css)


with demo:
    gr.Markdown(title)
    language_choices = list(language_to_models.keys())

    language_radio = gr.Radio(
        label="Language",
        choices=language_choices,
        value=language_choices[0],
    )

    model_dropdown = gr.Dropdown(
        choices=language_to_models[language_choices[0]],
        label="Select a model",
        value=language_to_models[language_choices[0]][0],
    )

    language_radio.change(
        update_model_dropdown,
        inputs=language_radio,
        outputs=model_dropdown,
    )

    with gr.Tabs():
        with gr.TabItem("Please input your text"):
            input_text = gr.Textbox(
                label="Input text",
                info="Your text",
                lines=3,
                placeholder="Please input your text here",
            )

            input_sid = gr.Textbox(
                label="Speaker ID",
                info="Speaker ID",
                lines=1,
                max_lines=1,
                value="0",
                placeholder="Speaker ID. Valid only for mult-speaker model",
            )
            input_button = gr.Button("Submit")

            output_audio = gr.Audio(label="Output")

            output_info = gr.HTML(label="Info")

        input_button.click(
            process,
            inputs=[
                language_radio,
                model_dropdown,
                input_text,
                input_sid,
            ],
            outputs=[
                output_audio,
                output_info,
            ],
        )

    gr.Markdown(description)

if __name__ == "__main__":
    formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s"

    logging.basicConfig(format=formatter, level=logging.INFO)

    demo.launch()