File size: 8,093 Bytes
3590c0c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f0a085b
03b48c7
f0a085b
3590c0c
f0a085b
 
21fcf42
3590c0c
 
 
 
 
 
 
 
8ebeeb8
 
 
3590c0c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ea07244
 
 
02084e1
 
 
 
 
ea07244
 
 
 
3590c0c
 
 
 
 
 
 
 
 
 
f0a085b
 
808a042
f0a085b
 
b922c32
 
 
21fcf42
b922c32
 
 
6393c35
 
 
 
 
 
 
 
 
b922c32
 
 
 
cfd7673
 
b922c32
 
 
3590c0c
ea07244
21fcf42
3590c0c
 
 
6393c35
 
 
 
 
 
 
 
 
3590c0c
 
 
 
1cd8b79
e4911f7
 
21fcf42
4384167
f0a085b
 
21fcf42
 
 
 
f0a085b
cfd7673
de273b3
cfd7673
f0a085b
03b48c7
 
 
 
ffd4c6b
de273b3
 
03b48c7
8ed6ca2
03b48c7
994c238
cfd7673
03b48c7
3590c0c
 
 
 
 
 
 
ea07244
3590c0c
 
 
 
 
 
 
ea07244
76241f8
ea07244
 
 
 
 
 
 
 
 
21fcf42
 
 
 
 
ea07244
3590c0c
 
e4911f7
02084e1
3590c0c
f0a085b
3590c0c
e4911f7
3590c0c
994c238
e4911f7
 
 
 
 
 
cfd7673
 
 
 
e4911f7
994c238
e4911f7
 
4360bb8
e4911f7
 
 
 
 
 
 
 
 
 
 
cfd7673
 
 
 
e4911f7
994c238
e4911f7
b922c32
3590c0c
 
ea07244
21fcf42
e4911f7
3590c0c
03b48c7
994c238
e4911f7
 
 
cfd7673
e4911f7
 
 
 
94e0fd2
e4911f7
 
 
21fcf42
e4911f7
 
 
 
 
 
cfd7673
03b48c7
3590c0c
 
 
 
 
 
 
 
 
 
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
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
#!/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
from pathlib import Path

import gradio as gr

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

title = "# Next-gen Kaldi: Generate subtitles for videos"

description = """
This space shows how to generate subtitles/captions with Next-gen Kaldi.

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

Please find test video files at
<https://huggingface.co/csukuangfj/vad/tree/main>

See more information by visiting the following links:

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

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(
            choices=choices,
            value=choices[0],
            interactive=True,
        )

    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 show_file_info(in_filename: str):
    logging.info(f"Input file: {in_filename}")
    _ = os.system(f"ffprobe -hide_banner -i '{in_filename}'")


def process_uploaded_video_file(
    language: str,
    repo_id: str,
    add_punctuation: str,
    in_filename: str,
):
    if in_filename is None or in_filename == "":
        return (
            "",
            build_html_output(
                "Please first upload a file and then click "
                'the button "submit for recognition"',
                "result_item_error",
            ),
            "",
            "",
        )

    logging.info(f"Processing uploaded file: {in_filename}")

    ans, all_text = process(language, repo_id, add_punctuation, in_filename)
    return (in_filename, ans[0]), ans[0], ans[1], ans[2], all_text


def process_uploaded_audio_file(
    language: str,
    repo_id: str,
    add_punctuation: str,
    in_filename: str,
):
    if in_filename is None or in_filename == "":
        return (
            "",
            build_html_output(
                "Please first upload a file and then click "
                'the button "submit for recognition"',
                "result_item_error",
            ),
            "",
            "",
        )

    logging.info(f"Processing uploaded file: {in_filename}")

    return process(language, repo_id, add_punctuation, in_filename)


def process(language: str, repo_id: str, add_punctuation: str, in_filename: str):
    logging.info(f"add_punctuation: {add_punctuation}")
    recognizer = get_pretrained_model(repo_id)
    vad = get_vad()
    if add_punctuation == "Yes":
        punct = get_punct_model()
    else:
        punct = None

    result, all_text = decode(recognizer, vad, punct, in_filename)
    logging.info(result)
    logging.info(all_text)

    srt_filename = Path(in_filename).with_suffix(".srt")
    with open(srt_filename, "w", encoding="utf-8") as f:
        f.write(result)

    show_file_info(in_filename)
    logging.info("Done")

    return (
        str(srt_filename),
        build_html_output("Done! Please download the SRT file", "result_item_success"),
        result,
        all_text,
    )


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,
    )
    punct_radio = gr.Radio(
        label="Whether to add punctuation",
        choices=["Yes", "No"],
        value="Yes",
    )

    with gr.Tabs():
        with gr.TabItem("Upload video from disk"):
            uploaded_video_file = gr.Video(
                sources=["upload"],
                label="Upload from disk",
                show_share_button=True,
            )
            upload_video_button = gr.Button("Submit for recognition")

            output_video = gr.Video(label="Output")
            output_srt_file_video = gr.File(
                label="Generated subtitles", show_label=True
            )

            output_info_video = gr.HTML(label="Info")
            output_textbox_video = gr.Textbox(
                label="Recognized speech from uploaded video file (srt format)"
            )
            all_output_textbox_video = gr.Textbox(
                label="Recognized speech from uploaded video file (all in one)"
            )

        with gr.TabItem("Upload audio from disk"):
            uploaded_audio_file = gr.Audio(
                sources=["upload"],  # Choose between "microphone", "upload"
                type="filepath",
                label="Upload audio from disk",
            )
            upload_audio_button = gr.Button("Submit for recognition")

            output_srt_file_audio = gr.File(
                label="Generated subtitles", show_label=True
            )

            output_info_audio = gr.HTML(label="Info")
            output_textbox_audio = gr.Textbox(
                label="Recognized speech from uploaded audio file (srt format)"
            )
            all_output_textbox_audio = gr.Textbox(
                label="Recognized speech from uploaded audio file (all in one)"
            )

        upload_video_button.click(
            process_uploaded_video_file,
            inputs=[
                language_radio,
                model_dropdown,
                punct_radio,
                uploaded_video_file,
            ],
            outputs=[
                output_video,
                output_srt_file_video,
                output_info_video,
                output_textbox_video,
                all_output_textbox_video,
            ],
        )

        upload_audio_button.click(
            process_uploaded_audio_file,
            inputs=[
                language_radio,
                model_dropdown,
                punct_radio,
                uploaded_audio_file,
            ],
            outputs=[
                output_srt_file_audio,
                output_info_audio,
                output_textbox_audio,
                all_output_textbox_audio,
            ],
        )

    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()