Spaces:
Running
Running
better progress bars
Browse files
app.py
CHANGED
@@ -258,7 +258,6 @@ def rename_speakers(text: str, speaker_map: dict) -> str:
|
|
258 |
|
259 |
def process_audio(audio_file):
|
260 |
try:
|
261 |
-
# Save uploaded file with a temporary name
|
262 |
temp_path = Path("temp_audio")
|
263 |
temp_path.mkdir(exist_ok=True)
|
264 |
temp_file = temp_path / "temp_audio.mp3"
|
@@ -267,34 +266,48 @@ def process_audio(audio_file):
|
|
267 |
with open(temp_file, "wb") as f:
|
268 |
f.write(audio_file)
|
269 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
270 |
# Get transcript
|
271 |
transcriber = Transcriber(os.getenv("ASSEMBLYAI_API_KEY"))
|
272 |
utterances = transcriber.get_transcript(temp_file)
|
273 |
-
|
274 |
-
# Generate original transcript
|
275 |
dialogues = list(group_utterances_by_speaker(utterances))
|
276 |
original = format_chunk(dialogues, markdown=True)
|
277 |
|
278 |
-
# Show original
|
279 |
-
yield
|
|
|
|
|
|
|
|
|
280 |
|
281 |
try:
|
282 |
-
# Enhance transcript
|
283 |
enhancer = Enhancer(os.getenv("GOOGLE_API_KEY"))
|
284 |
chunks = prepare_audio_chunks(temp_file, utterances)
|
285 |
enhanced = asyncio.run(enhancer.enhance_chunks(chunks))
|
286 |
-
|
287 |
-
# Format final transcript
|
288 |
merged = "\n\n".join(chunk.strip() for chunk in enhanced)
|
289 |
merged = apply_markdown_formatting(merged)
|
290 |
|
291 |
-
|
|
|
|
|
|
|
|
|
|
|
292 |
|
293 |
except Exception as e:
|
294 |
-
yield
|
|
|
|
|
|
|
|
|
295 |
|
296 |
finally:
|
297 |
-
# Cleanup temp file
|
298 |
if os.path.exists(temp_file):
|
299 |
os.remove(temp_file)
|
300 |
|
@@ -333,12 +346,33 @@ with gr.Blocks(title="Transcript Enhancer") as demo:
|
|
333 |
|
334 |
with gr.Column():
|
335 |
gr.Markdown("### Enhanced Transcript")
|
|
|
|
|
|
|
|
|
336 |
enhanced_output = gr.Markdown()
|
337 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
338 |
transcribe_btn.click(
|
339 |
fn=process_audio,
|
340 |
inputs=[audio_input],
|
341 |
-
outputs=[
|
|
|
|
|
|
|
|
|
342 |
)
|
343 |
|
344 |
# Launch the app
|
|
|
258 |
|
259 |
def process_audio(audio_file):
|
260 |
try:
|
|
|
261 |
temp_path = Path("temp_audio")
|
262 |
temp_path.mkdir(exist_ok=True)
|
263 |
temp_file = temp_path / "temp_audio.mp3"
|
|
|
266 |
with open(temp_file, "wb") as f:
|
267 |
f.write(audio_file)
|
268 |
|
269 |
+
# Initial state - show generating message
|
270 |
+
yield (
|
271 |
+
gr.update(value="", visible=True), # original transcript
|
272 |
+
gr.update(value="", visible=True), # enhanced transcript
|
273 |
+
gr.update(value="🎯 Generating transcript...", visible=True) # status
|
274 |
+
)
|
275 |
+
|
276 |
# Get transcript
|
277 |
transcriber = Transcriber(os.getenv("ASSEMBLYAI_API_KEY"))
|
278 |
utterances = transcriber.get_transcript(temp_file)
|
|
|
|
|
279 |
dialogues = list(group_utterances_by_speaker(utterances))
|
280 |
original = format_chunk(dialogues, markdown=True)
|
281 |
|
282 |
+
# Show original and enhancing message
|
283 |
+
yield (
|
284 |
+
gr.update(value=original, visible=True),
|
285 |
+
gr.update(value="", visible=True),
|
286 |
+
gr.update(value="🔄 Enhancing transcript...", visible=True)
|
287 |
+
)
|
288 |
|
289 |
try:
|
|
|
290 |
enhancer = Enhancer(os.getenv("GOOGLE_API_KEY"))
|
291 |
chunks = prepare_audio_chunks(temp_file, utterances)
|
292 |
enhanced = asyncio.run(enhancer.enhance_chunks(chunks))
|
|
|
|
|
293 |
merged = "\n\n".join(chunk.strip() for chunk in enhanced)
|
294 |
merged = apply_markdown_formatting(merged)
|
295 |
|
296 |
+
# Show final result
|
297 |
+
yield (
|
298 |
+
gr.update(value=original, visible=True),
|
299 |
+
gr.update(value=merged, visible=True),
|
300 |
+
gr.update(visible=False) # hide status
|
301 |
+
)
|
302 |
|
303 |
except Exception as e:
|
304 |
+
yield (
|
305 |
+
gr.update(value=original, visible=True),
|
306 |
+
gr.update(value=f"Error: {str(e)}", visible=True),
|
307 |
+
gr.update(visible=False)
|
308 |
+
)
|
309 |
|
310 |
finally:
|
|
|
311 |
if os.path.exists(temp_file):
|
312 |
os.remove(temp_file)
|
313 |
|
|
|
346 |
|
347 |
with gr.Column():
|
348 |
gr.Markdown("### Enhanced Transcript")
|
349 |
+
status = gr.Markdown(
|
350 |
+
visible=False,
|
351 |
+
elem_classes="status-message"
|
352 |
+
)
|
353 |
enhanced_output = gr.Markdown()
|
354 |
|
355 |
+
# Add some CSS
|
356 |
+
gr.Markdown("""
|
357 |
+
<style>
|
358 |
+
.status-message {
|
359 |
+
padding: 8px 15px;
|
360 |
+
border-radius: 4px;
|
361 |
+
background-color: #f0f0f0;
|
362 |
+
margin-bottom: 10px;
|
363 |
+
display: inline-block;
|
364 |
+
}
|
365 |
+
</style>
|
366 |
+
""")
|
367 |
+
|
368 |
transcribe_btn.click(
|
369 |
fn=process_audio,
|
370 |
inputs=[audio_input],
|
371 |
+
outputs=[
|
372 |
+
original_output,
|
373 |
+
enhanced_output,
|
374 |
+
status
|
375 |
+
]
|
376 |
)
|
377 |
|
378 |
# Launch the app
|