Spaces:
Running
on
Zero
Running
on
Zero
add 3 step
Browse files
app.py
CHANGED
|
@@ -4,49 +4,37 @@ from app.logger_config import (
|
|
| 4 |
)
|
| 5 |
import numpy as np
|
| 6 |
import gradio as gr
|
| 7 |
-
import asyncio
|
| 8 |
from fastrtc.webrtc import WebRTC
|
| 9 |
from fastrtc.utils import AdditionalOutputs
|
| 10 |
from pydub import AudioSegment
|
| 11 |
-
import time
|
| 12 |
-
import os
|
| 13 |
from gradio.utils import get_space
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
from app.utils import (
|
| 16 |
-
raise_error,
|
| 17 |
READ_SIZE,
|
| 18 |
generate_coturn_config,
|
| 19 |
read_and_stream_audio,
|
| 20 |
stop_streaming,
|
| 21 |
-
task,
|
| 22 |
-
task_fake
|
| 23 |
)
|
| 24 |
from app.session_utils import (
|
| 25 |
on_load,
|
| 26 |
on_unload,
|
| 27 |
get_active_session_hashes,
|
| 28 |
-
register_session_hash_code,
|
| 29 |
reset_all_active_sessions,
|
| 30 |
-
get_active_task_flag_file,
|
| 31 |
-
remove_active_task_flag_file
|
| 32 |
-
|
| 33 |
)
|
| 34 |
|
| 35 |
-
from app.ui_utils import (
|
| 36 |
-
SUPPORTED_LANGS_MAP,
|
| 37 |
-
EXAMPLE_CONFIGS,
|
| 38 |
-
apply_preset_if_example,
|
| 39 |
-
reset_to_defaults,
|
| 40 |
-
summarize_config,
|
| 41 |
-
handle_additional_outputs,
|
| 42 |
-
get_custom_theme,
|
| 43 |
-
on_file_load
|
| 44 |
-
)
|
| 45 |
-
import nemo.collections.asr as nemo_asr
|
| 46 |
-
from app.session_utils import (
|
| 47 |
-
get_active_task_flag_file,
|
| 48 |
-
get_session_hashe_chunks_dir
|
| 49 |
-
)
|
| 50 |
import spaces
|
| 51 |
# --------------------------------------------------------
|
| 52 |
# Initialization
|
|
@@ -109,43 +97,14 @@ with gr.Blocks(theme=theme, css=css_style) as demo:
|
|
| 109 |
"""
|
| 110 |
)
|
| 111 |
|
| 112 |
-
|
| 113 |
-
ui_components_oload_audio = [active_filepath,
|
| 114 |
main_audio.change(fn=on_file_load, inputs=[main_audio], outputs=ui_components_oload_audio)
|
| 115 |
-
# main_audio.stop_recording(fn=on_file_load, inputs=[main_audio], outputs=ui_components_one)
|
| 116 |
-
# main_audio.clear(fn=on_file_load, inputs=[main_audio], outputs=ui_components_one)
|
| 117 |
|
| 118 |
-
|
| 119 |
|
| 120 |
# === STEP 2 ===
|
| 121 |
-
with gr.Step("
|
| 122 |
-
gr.Markdown("### Step 2: Start audio streaming")
|
| 123 |
-
with gr.Group():
|
| 124 |
-
with gr.Column():
|
| 125 |
-
webrtc_stream = WebRTC(
|
| 126 |
-
label="Live Stream",
|
| 127 |
-
mode="receive",
|
| 128 |
-
modality="audio",
|
| 129 |
-
rtc_configuration=generate_coturn_config(),
|
| 130 |
-
visible=True,
|
| 131 |
-
inputs=main_audio,
|
| 132 |
-
)
|
| 133 |
-
start_stream_button = gr.Button("▶️ Start Streaming", variant="primary")
|
| 134 |
-
|
| 135 |
-
webrtc_stream.stream(
|
| 136 |
-
fn=read_and_stream_audio,
|
| 137 |
-
inputs=[active_filepath, session_hash_code,gr.State(READ_SIZE)],
|
| 138 |
-
outputs=[webrtc_stream],
|
| 139 |
-
trigger=start_stream_button.click,
|
| 140 |
-
concurrency_id="audio_stream",
|
| 141 |
-
concurrency_limit=10,
|
| 142 |
-
)
|
| 143 |
-
status_message_stream = gr.Markdown("", elem_id="status-message-stream", visible=False)
|
| 144 |
-
go_to_config = gr.Button("Go to Configuration", visible=False, variant="secondary")
|
| 145 |
-
go_to_config.click(lambda: gr.Walkthrough(selected=2), outputs=walkthrough)
|
| 146 |
-
|
| 147 |
-
# === STEP 3 ===
|
| 148 |
-
with gr.Step("Configuration", id=2):
|
| 149 |
gr.Markdown("### Step 3: Configure the Task")
|
| 150 |
with gr.Group():
|
| 151 |
with gr.Row():
|
|
@@ -234,24 +193,49 @@ with gr.Blocks(theme=theme, css=css_style) as demo:
|
|
| 234 |
)
|
| 235 |
|
| 236 |
go_to_task = gr.Button("Go to Task")
|
| 237 |
-
go_to_task.click(lambda: gr.Walkthrough(selected=
|
| 238 |
|
| 239 |
-
# === STEP
|
| 240 |
-
with gr.Step("Task", id=
|
| 241 |
-
with gr.Row():
|
| 242 |
-
gr.Markdown("## Step 4: Start the Task")
|
| 243 |
with gr.Row():
|
|
|
|
|
|
|
| 244 |
with gr.Column():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 245 |
status_slider = gr.Slider(
|
| 246 |
0, 100,
|
| 247 |
value=0,
|
| 248 |
-
label="Streaming Progress",
|
| 249 |
show_label=True,
|
| 250 |
interactive=False,
|
| 251 |
visible=False,
|
| 252 |
show_reset_button=False
|
| 253 |
)
|
| 254 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 255 |
with gr.Row():
|
| 256 |
gr.Markdown("---")
|
| 257 |
with gr.Row():
|
|
@@ -270,7 +254,7 @@ with gr.Blocks(theme=theme, css=css_style) as demo:
|
|
| 270 |
with gr.Row():
|
| 271 |
status_message_task = gr.Markdown("", elem_id="status-message-task",elem_classes=["info"], visible=False)
|
| 272 |
with gr.Row():
|
| 273 |
-
start_task_button = gr.Button("▶️ Start Task", visible=
|
| 274 |
stop_task_button = gr.Button("⏹️ Stop Task", visible=False,variant="stop")
|
| 275 |
|
| 276 |
stop_stream_button.click(
|
|
@@ -278,46 +262,16 @@ with gr.Blocks(theme=theme, css=css_style) as demo:
|
|
| 278 |
inputs=[session_hash_code],
|
| 279 |
)
|
| 280 |
|
| 281 |
-
def stop_task_fn(session_hash_code):
|
| 282 |
-
remove_active_task_flag_file(session_hash_code)
|
| 283 |
-
yield "Task stopped by user."
|
| 284 |
-
|
| 285 |
-
|
| 286 |
stop_task_button.click(
|
| 287 |
fn=stop_task_fn,
|
| 288 |
inputs=session_hash_code,
|
| 289 |
-
|
| 290 |
)
|
| 291 |
-
# task(session_hash_code)
|
| 292 |
config_task_ui = [session_hash_code,task_type, lang_source, lang_target,
|
| 293 |
chunk_secs, left_context_secs, right_context_secs,
|
| 294 |
streaming_policy, alignatt_thr, waitk_lagging,
|
| 295 |
exclude_sink_frames, xatt_scores_layer, hallucinations_detector]
|
| 296 |
|
| 297 |
-
def start_task_asr_ast(
|
| 298 |
-
session_hash_code,
|
| 299 |
-
task_type, lang_source, lang_target,
|
| 300 |
-
chunk_secs, left_context_secs, right_context_secs,
|
| 301 |
-
streaming_policy, alignatt_thr, waitk_lagging,
|
| 302 |
-
exclude_sink_frames, xatt_scores_layer, hallucinations_detector
|
| 303 |
-
):
|
| 304 |
-
"""Stream transcription or translation results in real time."""
|
| 305 |
-
accumulated = ""
|
| 306 |
-
# Boucle sur le générateur de `task2()`
|
| 307 |
-
for result, status, current_chunk in task(
|
| 308 |
-
session_hash_code,
|
| 309 |
-
task_type, lang_source, lang_target,
|
| 310 |
-
chunk_secs, left_context_secs, right_context_secs,
|
| 311 |
-
streaming_policy, alignatt_thr, waitk_lagging,
|
| 312 |
-
exclude_sink_frames, xatt_scores_layer, hallucinations_detector
|
| 313 |
-
):
|
| 314 |
-
if status == "success":
|
| 315 |
-
yield accumulated + result, gr.update(visible=True,value=current_chunk , elem_classes=["info"]), gr.update(visible=False), gr.update(visible=True)
|
| 316 |
-
accumulated += result
|
| 317 |
-
elif status in ["warning","info" ]:
|
| 318 |
-
yield accumulated, gr.update(visible=True,value=result , elem_classes=[status]), gr.update(visible=False), gr.update(visible=True)
|
| 319 |
-
elif status in ["error", "done"]:
|
| 320 |
-
yield accumulated, gr.update(visible=True,value=result , elem_classes=[status]), gr.update(visible=True), gr.update(visible=False)
|
| 321 |
|
| 322 |
|
| 323 |
start_task_button.click(
|
|
@@ -330,11 +284,11 @@ with gr.Blocks(theme=theme, css=css_style) as demo:
|
|
| 330 |
exclude_sink_frames, xatt_scores_layer, hallucinations_detector
|
| 331 |
|
| 332 |
],
|
| 333 |
-
outputs=[task_output,status_message_task,start_task_button,stop_task_button]
|
| 334 |
)
|
| 335 |
|
| 336 |
ui_components = [
|
| 337 |
-
start_stream_button, stop_stream_button,
|
| 338 |
go_to_config, audio_source_step, status_slider,walkthrough,status_message_stream
|
| 339 |
]
|
| 340 |
|
|
|
|
| 4 |
)
|
| 5 |
import numpy as np
|
| 6 |
import gradio as gr
|
|
|
|
| 7 |
from fastrtc.webrtc import WebRTC
|
| 8 |
from fastrtc.utils import AdditionalOutputs
|
| 9 |
from pydub import AudioSegment
|
|
|
|
|
|
|
| 10 |
from gradio.utils import get_space
|
| 11 |
+
from app.supported_languages import (
|
| 12 |
+
SUPPORTED_LANGS_MAP,
|
| 13 |
+
)
|
| 14 |
+
from app.ui_utils import (
|
| 15 |
+
EXAMPLE_CONFIGS,
|
| 16 |
+
apply_preset_if_example,
|
| 17 |
+
reset_to_defaults,
|
| 18 |
+
summarize_config,
|
| 19 |
+
handle_additional_outputs,
|
| 20 |
+
get_custom_theme,
|
| 21 |
+
on_file_load,
|
| 22 |
+
start_task_asr_ast,
|
| 23 |
+
stop_task_fn
|
| 24 |
+
)
|
| 25 |
from app.utils import (
|
|
|
|
| 26 |
READ_SIZE,
|
| 27 |
generate_coturn_config,
|
| 28 |
read_and_stream_audio,
|
| 29 |
stop_streaming,
|
|
|
|
|
|
|
| 30 |
)
|
| 31 |
from app.session_utils import (
|
| 32 |
on_load,
|
| 33 |
on_unload,
|
| 34 |
get_active_session_hashes,
|
|
|
|
| 35 |
reset_all_active_sessions,
|
|
|
|
|
|
|
|
|
|
| 36 |
)
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
import spaces
|
| 39 |
# --------------------------------------------------------
|
| 40 |
# Initialization
|
|
|
|
| 97 |
"""
|
| 98 |
)
|
| 99 |
|
| 100 |
+
go_to_config = gr.Button("Go to Configuration", visible=False)
|
| 101 |
+
ui_components_oload_audio = [active_filepath, go_to_config]
|
| 102 |
main_audio.change(fn=on_file_load, inputs=[main_audio], outputs=ui_components_oload_audio)
|
|
|
|
|
|
|
| 103 |
|
| 104 |
+
go_to_config.click(lambda: gr.Walkthrough(selected=1), outputs=walkthrough)
|
| 105 |
|
| 106 |
# === STEP 2 ===
|
| 107 |
+
with gr.Step("Configuration", id=1)as config_step:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
gr.Markdown("### Step 3: Configure the Task")
|
| 109 |
with gr.Group():
|
| 110 |
with gr.Row():
|
|
|
|
| 193 |
)
|
| 194 |
|
| 195 |
go_to_task = gr.Button("Go to Task")
|
| 196 |
+
go_to_task.click(lambda: gr.Walkthrough(selected=2), outputs=walkthrough)
|
| 197 |
|
| 198 |
+
# === STEP 3 ===
|
| 199 |
+
with gr.Step("Task", id=2) as task_step:
|
|
|
|
|
|
|
| 200 |
with gr.Row():
|
| 201 |
+
gr.Markdown("## Step 4: Start Stream Task")
|
| 202 |
+
with gr.Group():
|
| 203 |
with gr.Column():
|
| 204 |
+
webrtc_stream = WebRTC(
|
| 205 |
+
label="Live Stream",
|
| 206 |
+
mode="receive",
|
| 207 |
+
modality="audio",
|
| 208 |
+
rtc_configuration=generate_coturn_config(),
|
| 209 |
+
visible=True,
|
| 210 |
+
inputs=main_audio,
|
| 211 |
+
icon= "https://cdn-icons-png.flaticon.com/128/18429/18429788.png",
|
| 212 |
+
pulse_color= "#df7a7a",
|
| 213 |
+
icon_radius= "10px",
|
| 214 |
+
icon_button_color= "rgb(255, 255, 255)",
|
| 215 |
+
height=150,
|
| 216 |
+
show_label=False
|
| 217 |
+
)
|
| 218 |
status_slider = gr.Slider(
|
| 219 |
0, 100,
|
| 220 |
value=0,
|
| 221 |
+
label="Streaming Progress %",
|
| 222 |
show_label=True,
|
| 223 |
interactive=False,
|
| 224 |
visible=False,
|
| 225 |
show_reset_button=False
|
| 226 |
)
|
| 227 |
+
start_stream_button = gr.Button("▶️ Start Streaming", variant="primary")
|
| 228 |
+
stop_stream_button = gr.Button("⏹️ Stop Streaming", visible=False,variant="stop")
|
| 229 |
+
webrtc_stream.stream(
|
| 230 |
+
fn=read_and_stream_audio,
|
| 231 |
+
inputs=[active_filepath, session_hash_code,gr.State(READ_SIZE)],
|
| 232 |
+
outputs=[webrtc_stream],
|
| 233 |
+
trigger=start_stream_button.click,
|
| 234 |
+
concurrency_id="audio_stream",
|
| 235 |
+
concurrency_limit=10,
|
| 236 |
+
)
|
| 237 |
+
status_message_stream = gr.Markdown("", elem_id="status-message-stream", visible=False)
|
| 238 |
+
|
| 239 |
with gr.Row():
|
| 240 |
gr.Markdown("---")
|
| 241 |
with gr.Row():
|
|
|
|
| 254 |
with gr.Row():
|
| 255 |
status_message_task = gr.Markdown("", elem_id="status-message-task",elem_classes=["info"], visible=False)
|
| 256 |
with gr.Row():
|
| 257 |
+
start_task_button = gr.Button("▶️ Start Task", visible=False, variant="primary")
|
| 258 |
stop_task_button = gr.Button("⏹️ Stop Task", visible=False,variant="stop")
|
| 259 |
|
| 260 |
stop_stream_button.click(
|
|
|
|
| 262 |
inputs=[session_hash_code],
|
| 263 |
)
|
| 264 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 265 |
stop_task_button.click(
|
| 266 |
fn=stop_task_fn,
|
| 267 |
inputs=session_hash_code,
|
| 268 |
+
outputs=task_output
|
| 269 |
)
|
|
|
|
| 270 |
config_task_ui = [session_hash_code,task_type, lang_source, lang_target,
|
| 271 |
chunk_secs, left_context_secs, right_context_secs,
|
| 272 |
streaming_policy, alignatt_thr, waitk_lagging,
|
| 273 |
exclude_sink_frames, xatt_scores_layer, hallucinations_detector]
|
| 274 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 275 |
|
| 276 |
|
| 277 |
start_task_button.click(
|
|
|
|
| 284 |
exclude_sink_frames, xatt_scores_layer, hallucinations_detector
|
| 285 |
|
| 286 |
],
|
| 287 |
+
outputs=[task_output,status_message_task,start_task_button,stop_task_button,config_step]
|
| 288 |
)
|
| 289 |
|
| 290 |
ui_components = [
|
| 291 |
+
start_stream_button, stop_stream_button,start_task_button,
|
| 292 |
go_to_config, audio_source_step, status_slider,walkthrough,status_message_stream
|
| 293 |
]
|
| 294 |
|