developer28 commited on
Commit
168cef5
Β·
verified Β·
1 Parent(s): 3e5d7ea

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +31 -30
app.py CHANGED
@@ -13,7 +13,7 @@ try:
13
  except:
14
  WHISPER_AVAILABLE = False
15
 
16
- # βœ… Download audio from YouTube
17
  def download_audio(url, cookies_path=None):
18
  try:
19
  temp_dir = tempfile.mkdtemp()
@@ -25,15 +25,9 @@ def download_audio(url, cookies_path=None):
25
  'quiet': True,
26
  'noplaylist': True,
27
  'cookiefile': cookies_path if cookies_path else None,
28
- 'user_agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64)',
29
  'referer': 'https://www.youtube.com/',
30
  'force_ipv4': True,
31
- 'http_headers': {
32
- 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64)',
33
- 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
34
- 'Accept-Language': 'en-US,en;q=0.5',
35
- 'Referer': 'https://www.youtube.com/'
36
- },
37
  }
38
 
39
  with YoutubeDL(ydl_opts) as ydl:
@@ -50,7 +44,7 @@ def download_audio(url, cookies_path=None):
50
  traceback.print_exc()
51
  return None, f"❌ Download error: {str(e)}"
52
 
53
- # βœ… Transcribe audio using Whisper
54
  def transcribe_audio(path):
55
  if not WHISPER_AVAILABLE:
56
  return "❌ Whisper not available. Please install openai-whisper."
@@ -62,13 +56,13 @@ def transcribe_audio(path):
62
  traceback.print_exc()
63
  return f"❌ Transcription failed: {str(e)}"
64
 
65
- # βœ… Extract stock-related information
66
  def extract_stock_info(text):
67
  try:
68
  companies = re.findall(r'\b[A-Z][a-z]+(?: [A-Z][a-z]+)*\b', text)
69
  symbols = re.findall(r'\b[A-Z]{2,5}\b', text)
70
  prices = re.findall(r'\$\d+(?:\.\d{1,2})?', text)
71
- actions = re.findall(r'\b(buy|sell|hold|target|bullish|bearish|stop loss)\b', text, re.IGNORECASE)
72
 
73
  result = "=== STOCK RECOMMENDATION ANALYSIS ===\n\n"
74
  if companies:
@@ -82,7 +76,7 @@ def extract_stock_info(text):
82
 
83
  recommendations = []
84
  for line in text.split("."):
85
- if any(word in line.lower() for word in ['buy', 'sell', 'target', 'hold']):
86
  recommendations.append(line.strip())
87
 
88
  if recommendations:
@@ -98,7 +92,7 @@ def extract_stock_info(text):
98
  except Exception as e:
99
  return f"❌ Stock info extraction failed: {str(e)}"
100
 
101
- # βœ… Save uploaded cookies.txt
102
  def save_cookies(file):
103
  if file is None:
104
  return None
@@ -115,11 +109,11 @@ def save_cookies(file):
115
  print(f"❌ Failed to handle cookies.txt: {e}")
116
  return None
117
 
118
- # βœ… YouTube pipeline
119
- def run_pipeline(url, cookies_file):
120
  try:
121
  if not WHISPER_AVAILABLE:
122
- return "❌ Whisper is not installed. Run: pip install openai-whisper", ""
123
  if not url:
124
  return "❌ YouTube URL required", ""
125
 
@@ -133,22 +127,24 @@ def run_pipeline(url, cookies_file):
133
  return transcript, ""
134
 
135
  stock_info = extract_stock_info(transcript)
136
- return "βœ… Complete", stock_info
 
 
 
137
 
138
  except Exception as e:
139
  tb = traceback.format_exc()
140
  print(tb)
141
  return f"❌ Unhandled Error:\n{tb}", ""
142
 
143
- # βœ… Audio file upload pipeline
144
- def run_pipeline_audio(audio_file):
145
  try:
146
  if not WHISPER_AVAILABLE:
147
- return "❌ Whisper is not installed. Run: pip install openai-whisper", ""
148
  if audio_file is None:
149
  return "❌ No audio file uploaded", ""
150
 
151
- # Handle both file-like and NamedString (path string)
152
  temp_audio_path = tempfile.mktemp(suffix=os.path.splitext(str(audio_file))[-1])
153
  if hasattr(audio_file, "read"):
154
  with open(temp_audio_path, "wb") as f:
@@ -161,35 +157,40 @@ def run_pipeline_audio(audio_file):
161
  return transcript, ""
162
 
163
  stock_info = extract_stock_info(transcript)
164
- return "βœ… Complete", stock_info
 
 
 
165
 
166
  except Exception as e:
167
  tb = traceback.format_exc()
168
  print(tb)
169
  return f"❌ Unhandled Error:\n{tb}", ""
170
 
171
- # βœ… Gradio UI
172
  with gr.Blocks(title="Stock Insights from YouTube or Audio") as demo:
173
  gr.Markdown("""
174
- # 🎧 Extract Stock Recommendations from YouTube or Uploaded Audio
175
- Upload a YouTube URL or an audio file. We'll transcribe it and extract stock-related insights!
176
  """)
177
 
178
  with gr.Tab("πŸ“Ί From YouTube Video"):
179
  with gr.Row():
180
- url_input = gr.Textbox(label="πŸŽ₯ YouTube Video URL")
181
  cookie_input = gr.File(label="cookies.txt (optional)", file_types=[".txt"])
 
182
  yt_run_btn = gr.Button("πŸš€ Extract from YouTube")
183
  yt_status = gr.Textbox(label="Status")
184
- yt_result = gr.Textbox(label="Stock Info", lines=12)
185
- yt_run_btn.click(fn=run_pipeline, inputs=[url_input, cookie_input], outputs=[yt_status, yt_result])
186
 
187
  with gr.Tab("🎡 From Uploaded Audio"):
188
  audio_input = gr.File(label="Upload Audio File", file_types=[".mp3", ".wav", ".m4a", ".webm"])
 
189
  audio_run_btn = gr.Button("πŸš€ Extract from Audio")
190
  audio_status = gr.Textbox(label="Status")
191
- audio_result = gr.Textbox(label="Stock Info", lines=12)
192
- audio_run_btn.click(fn=run_pipeline_audio, inputs=[audio_input], outputs=[audio_status, audio_result])
193
 
194
  if __name__ == "__main__":
195
  demo.launch(debug=True)
 
13
  except:
14
  WHISPER_AVAILABLE = False
15
 
16
+ # Download audio from YouTube
17
  def download_audio(url, cookies_path=None):
18
  try:
19
  temp_dir = tempfile.mkdtemp()
 
25
  'quiet': True,
26
  'noplaylist': True,
27
  'cookiefile': cookies_path if cookies_path else None,
28
+ 'user_agent': 'Mozilla/5.0',
29
  'referer': 'https://www.youtube.com/',
30
  'force_ipv4': True,
 
 
 
 
 
 
31
  }
32
 
33
  with YoutubeDL(ydl_opts) as ydl:
 
44
  traceback.print_exc()
45
  return None, f"❌ Download error: {str(e)}"
46
 
47
+ # Transcribe using Whisper
48
  def transcribe_audio(path):
49
  if not WHISPER_AVAILABLE:
50
  return "❌ Whisper not available. Please install openai-whisper."
 
56
  traceback.print_exc()
57
  return f"❌ Transcription failed: {str(e)}"
58
 
59
+ # Extract stock insights
60
  def extract_stock_info(text):
61
  try:
62
  companies = re.findall(r'\b[A-Z][a-z]+(?: [A-Z][a-z]+)*\b', text)
63
  symbols = re.findall(r'\b[A-Z]{2,5}\b', text)
64
  prices = re.findall(r'\$\d+(?:\.\d{1,2})?', text)
65
+ actions = re.findall(r'\b(buy|sell|hold|target|bullish|bearish|stop loss|accumulate|short|take profit|entry|exit)\b', text, re.IGNORECASE)
66
 
67
  result = "=== STOCK RECOMMENDATION ANALYSIS ===\n\n"
68
  if companies:
 
76
 
77
  recommendations = []
78
  for line in text.split("."):
79
+ if any(word in line.lower() for word in ['buy', 'sell', 'target', 'hold', 'accumulate', 'short', 'entry', 'exit']):
80
  recommendations.append(line.strip())
81
 
82
  if recommendations:
 
92
  except Exception as e:
93
  return f"❌ Stock info extraction failed: {str(e)}"
94
 
95
+ # Save cookies
96
  def save_cookies(file):
97
  if file is None:
98
  return None
 
109
  print(f"❌ Failed to handle cookies.txt: {e}")
110
  return None
111
 
112
+ # YouTube flow
113
+ def run_pipeline(url, cookies_file, show_transcript):
114
  try:
115
  if not WHISPER_AVAILABLE:
116
+ return "❌ Whisper not installed", ""
117
  if not url:
118
  return "❌ YouTube URL required", ""
119
 
 
127
  return transcript, ""
128
 
129
  stock_info = extract_stock_info(transcript)
130
+ if show_transcript:
131
+ return "βœ… Complete", f"πŸ“œ Transcript:\n\n{transcript}\n\n\n{stock_info}"
132
+ else:
133
+ return "βœ… Complete", stock_info
134
 
135
  except Exception as e:
136
  tb = traceback.format_exc()
137
  print(tb)
138
  return f"❌ Unhandled Error:\n{tb}", ""
139
 
140
+ # Audio upload flow
141
+ def run_pipeline_audio(audio_file, show_transcript):
142
  try:
143
  if not WHISPER_AVAILABLE:
144
+ return "❌ Whisper not installed", ""
145
  if audio_file is None:
146
  return "❌ No audio file uploaded", ""
147
 
 
148
  temp_audio_path = tempfile.mktemp(suffix=os.path.splitext(str(audio_file))[-1])
149
  if hasattr(audio_file, "read"):
150
  with open(temp_audio_path, "wb") as f:
 
157
  return transcript, ""
158
 
159
  stock_info = extract_stock_info(transcript)
160
+ if show_transcript:
161
+ return "βœ… Complete", f"πŸ“œ Transcript:\n\n{transcript}\n\n\n{stock_info}"
162
+ else:
163
+ return "βœ… Complete", stock_info
164
 
165
  except Exception as e:
166
  tb = traceback.format_exc()
167
  print(tb)
168
  return f"❌ Unhandled Error:\n{tb}", ""
169
 
170
+ # Gradio UI
171
  with gr.Blocks(title="Stock Insights from YouTube or Audio") as demo:
172
  gr.Markdown("""
173
+ # πŸ“ˆ Extract Stock Recommendations from YouTube or Uploaded Audio
174
+ Upload a YouTube video or audio file. We'll transcribe it using Whisper and extract stock insights.
175
  """)
176
 
177
  with gr.Tab("πŸ“Ί From YouTube Video"):
178
  with gr.Row():
179
+ url_input = gr.Textbox(label="πŸŽ₯ YouTube URL")
180
  cookie_input = gr.File(label="cookies.txt (optional)", file_types=[".txt"])
181
+ show_transcript_yt = gr.Checkbox(label="Show Transcript", value=False)
182
  yt_run_btn = gr.Button("πŸš€ Extract from YouTube")
183
  yt_status = gr.Textbox(label="Status")
184
+ yt_result = gr.Textbox(label="Transcript & Stock Info", lines=15)
185
+ yt_run_btn.click(fn=run_pipeline, inputs=[url_input, cookie_input, show_transcript_yt], outputs=[yt_status, yt_result])
186
 
187
  with gr.Tab("🎡 From Uploaded Audio"):
188
  audio_input = gr.File(label="Upload Audio File", file_types=[".mp3", ".wav", ".m4a", ".webm"])
189
+ show_transcript_audio = gr.Checkbox(label="Show Transcript", value=False)
190
  audio_run_btn = gr.Button("πŸš€ Extract from Audio")
191
  audio_status = gr.Textbox(label="Status")
192
+ audio_result = gr.Textbox(label="Transcript & Stock Info", lines=15)
193
+ audio_run_btn.click(fn=run_pipeline_audio, inputs=[audio_input, show_transcript_audio], outputs=[audio_status, audio_result])
194
 
195
  if __name__ == "__main__":
196
  demo.launch(debug=True)