Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -383,6 +383,7 @@ def process_youtube_video(url="", keywords=""):
|
|
| 383 |
summary = ""
|
| 384 |
sentiment_label = "N/A"
|
| 385 |
recommendations = ""
|
|
|
|
| 386 |
|
| 387 |
if not url.strip():
|
| 388 |
return thumbnail, "Please enter a YouTube URL", sentiment_label, recommendations
|
|
@@ -393,36 +394,49 @@ def process_youtube_video(url="", keywords=""):
|
|
| 393 |
|
| 394 |
thumbnail = f"https://img.youtube.com/vi/{video_id}/maxresdefault.jpg"
|
| 395 |
|
|
|
|
| 396 |
try:
|
| 397 |
-
|
|
|
|
|
|
|
|
|
|
| 398 |
try:
|
| 399 |
-
|
| 400 |
-
|
| 401 |
-
|
| 402 |
-
#
|
| 403 |
-
|
| 404 |
-
|
| 405 |
-
|
| 406 |
-
|
| 407 |
-
for lang_code in ['en', 'en-US', 'en-GB', 'a.en']:
|
| 408 |
-
try:
|
| 409 |
-
transcript = transcript_list.find_transcript([lang_code])
|
| 410 |
text = " ".join([t['text'] for t in transcript.fetch()])
|
| 411 |
break
|
| 412 |
-
|
| 413 |
-
|
| 414 |
-
|
| 415 |
-
|
| 416 |
-
|
|
|
|
| 417 |
transcript = transcript_list.find_generated_transcript(['en'])
|
| 418 |
text = " ".join([t['text'] for t in transcript.fetch()])
|
| 419 |
-
|
| 420 |
-
|
| 421 |
-
|
| 422 |
-
|
| 423 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 424 |
|
| 425 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 426 |
cleaned_text = re.sub(r'[^\w\s.]', '', text)
|
| 427 |
cleaned_text = ' '.join(cleaned_text.split())
|
| 428 |
|
|
@@ -442,8 +456,7 @@ def process_youtube_video(url="", keywords=""):
|
|
| 442 |
summary = model.generate_content(f"Summarize this content: {cleaned_text[:4000]}").text
|
| 443 |
|
| 444 |
except Exception as e:
|
| 445 |
-
|
| 446 |
-
return thumbnail, f"⚠️ Unable to process video: {str(e)}", "N/A", recommendations
|
| 447 |
|
| 448 |
# Get recommendations
|
| 449 |
if keywords.strip():
|
|
@@ -452,7 +465,7 @@ def process_youtube_video(url="", keywords=""):
|
|
| 452 |
return thumbnail, summary, sentiment_label, recommendations
|
| 453 |
|
| 454 |
except Exception as e:
|
| 455 |
-
print(f"Debug - Main Error: {str(e)}")
|
| 456 |
return None, f"Error: {str(e)}", "N/A", ""
|
| 457 |
|
| 458 |
|
|
|
|
| 383 |
summary = ""
|
| 384 |
sentiment_label = "N/A"
|
| 385 |
recommendations = ""
|
| 386 |
+
text = ""
|
| 387 |
|
| 388 |
if not url.strip():
|
| 389 |
return thumbnail, "Please enter a YouTube URL", sentiment_label, recommendations
|
|
|
|
| 394 |
|
| 395 |
thumbnail = f"https://img.youtube.com/vi/{video_id}/maxresdefault.jpg"
|
| 396 |
|
| 397 |
+
# Method 1: Direct transcript fetch
|
| 398 |
try:
|
| 399 |
+
transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=['en'])
|
| 400 |
+
text = " ".join([t['text'] for t in transcript])
|
| 401 |
+
except:
|
| 402 |
+
# Method 2: List available transcripts
|
| 403 |
try:
|
| 404 |
+
transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
|
| 405 |
+
transcript = None
|
| 406 |
+
|
| 407 |
+
# Try English variants
|
| 408 |
+
for lang_code in ['en', 'en-US', 'en-GB', 'a.en']:
|
| 409 |
+
try:
|
| 410 |
+
transcript = transcript_list.find_transcript([lang_code])
|
| 411 |
+
if transcript:
|
|
|
|
|
|
|
|
|
|
| 412 |
text = " ".join([t['text'] for t in transcript.fetch()])
|
| 413 |
break
|
| 414 |
+
except:
|
| 415 |
+
continue
|
| 416 |
+
|
| 417 |
+
# Try auto-generated if no manual transcript
|
| 418 |
+
if not text:
|
| 419 |
+
try:
|
| 420 |
transcript = transcript_list.find_generated_transcript(['en'])
|
| 421 |
text = " ".join([t['text'] for t in transcript.fetch()])
|
| 422 |
+
except:
|
| 423 |
+
# Try translation as last resort
|
| 424 |
+
try:
|
| 425 |
+
manual_transcript = transcript_list.find_manually_created_transcript()
|
| 426 |
+
translated = manual_transcript.translate('en')
|
| 427 |
+
text = " ".join([t['text'] for t in translated.fetch()])
|
| 428 |
+
except:
|
| 429 |
+
raise Exception("No available transcripts found")
|
| 430 |
+
|
| 431 |
+
except Exception as e:
|
| 432 |
+
return thumbnail, f"⚠️ No transcripts available: {str(e)}", sentiment_label, recommendations
|
| 433 |
|
| 434 |
+
if not text:
|
| 435 |
+
return thumbnail, "⚠️ Could not extract transcript text", sentiment_label, recommendations
|
| 436 |
+
|
| 437 |
+
# Process valid transcript
|
| 438 |
+
try:
|
| 439 |
+
# Clean text
|
| 440 |
cleaned_text = re.sub(r'[^\w\s.]', '', text)
|
| 441 |
cleaned_text = ' '.join(cleaned_text.split())
|
| 442 |
|
|
|
|
| 456 |
summary = model.generate_content(f"Summarize this content: {cleaned_text[:4000]}").text
|
| 457 |
|
| 458 |
except Exception as e:
|
| 459 |
+
return thumbnail, f"⚠️ Error processing content: {str(e)}", sentiment_label, recommendations
|
|
|
|
| 460 |
|
| 461 |
# Get recommendations
|
| 462 |
if keywords.strip():
|
|
|
|
| 465 |
return thumbnail, summary, sentiment_label, recommendations
|
| 466 |
|
| 467 |
except Exception as e:
|
| 468 |
+
print(f"Debug - Main Error: {str(e)}")
|
| 469 |
return None, f"Error: {str(e)}", "N/A", ""
|
| 470 |
|
| 471 |
|