Spaces:
Sleeping
Sleeping
Nagesh Muralidhar
commited on
Commit
·
62fea8b
1
Parent(s):
72e705f
midterm-submission
Browse files- server/main.py +110 -110
server/main.py
CHANGED
|
@@ -39,10 +39,29 @@ chat_model = ChatOpenAI(
|
|
| 39 |
openai_api_key=openai_api_key
|
| 40 |
)
|
| 41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
# Initialize FastAPI app
|
| 43 |
app = FastAPI()
|
| 44 |
|
| 45 |
-
# Configure CORS
|
| 46 |
app.add_middleware(
|
| 47 |
CORSMiddleware,
|
| 48 |
allow_origins=["http://localhost:5173", "http://localhost:3000", "https://*.hf.space", "*"],
|
|
@@ -53,37 +72,49 @@ app.add_middleware(
|
|
| 53 |
max_age=600,
|
| 54 |
)
|
| 55 |
|
| 56 |
-
#
|
| 57 |
-
app.mount("/", StaticFiles(directory="static", html=True), name="frontend")
|
| 58 |
-
|
| 59 |
-
# Configure audio storage
|
| 60 |
audio_dir = os.path.join(os.path.dirname(__file__), "audio_storage")
|
| 61 |
os.makedirs(audio_dir, exist_ok=True)
|
| 62 |
|
| 63 |
-
# Mount the audio directory as a static file directory
|
| 64 |
-
app.mount("/audio-files", StaticFiles(directory=audio_dir), name="audio")
|
| 65 |
-
|
| 66 |
-
# Configure context storage
|
| 67 |
context_dir = os.path.join(os.path.dirname(__file__), "context_storage")
|
| 68 |
os.makedirs(context_dir, exist_ok=True)
|
| 69 |
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
final_podcast: Dict[str, Any]
|
| 81 |
-
|
| 82 |
-
class PodcastChatRequest(BaseModel):
|
| 83 |
-
message: str
|
| 84 |
|
| 85 |
-
|
| 86 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
|
| 88 |
@app.get("/audio-list")
|
| 89 |
async def list_audio_files():
|
|
@@ -103,55 +134,7 @@ async def list_audio_files():
|
|
| 103 |
except Exception as e:
|
| 104 |
raise HTTPException(status_code=500, detail=str(e))
|
| 105 |
|
| 106 |
-
|
| 107 |
-
async def delete_audio_file(filename: str):
|
| 108 |
-
"""Delete an audio file and its corresponding transcript."""
|
| 109 |
-
try:
|
| 110 |
-
# Delete audio file
|
| 111 |
-
file_path = os.path.join(audio_dir, filename)
|
| 112 |
-
if not os.path.exists(file_path):
|
| 113 |
-
raise HTTPException(status_code=404, detail="File not found")
|
| 114 |
-
|
| 115 |
-
# Get all audio files to determine the podcast ID
|
| 116 |
-
audio_files = [f for f in os.listdir(audio_dir) if f.endswith(('.mp3', '.wav'))]
|
| 117 |
-
try:
|
| 118 |
-
# Find the index (0-based) of the file being deleted
|
| 119 |
-
podcast_id = audio_files.index(filename) + 1 # Convert to 1-based ID
|
| 120 |
-
logger.info(f"Deleting podcast with ID: {podcast_id}")
|
| 121 |
-
|
| 122 |
-
# Path to transcripts file
|
| 123 |
-
transcripts_file = os.path.join(os.path.dirname(__file__), "transcripts", "podcasts.json")
|
| 124 |
-
|
| 125 |
-
# Update transcripts if file exists
|
| 126 |
-
if os.path.exists(transcripts_file):
|
| 127 |
-
with open(transcripts_file, 'r') as f:
|
| 128 |
-
transcripts = json.load(f)
|
| 129 |
-
|
| 130 |
-
# Remove the transcript at the corresponding index
|
| 131 |
-
if len(transcripts) >= podcast_id:
|
| 132 |
-
transcripts.pop(podcast_id - 1) # Convert back to 0-based index
|
| 133 |
-
|
| 134 |
-
# Save updated transcripts
|
| 135 |
-
with open(transcripts_file, 'w') as f:
|
| 136 |
-
json.dump(transcripts, f, indent=2)
|
| 137 |
-
logger.info(f"Removed transcript for podcast ID {podcast_id}")
|
| 138 |
-
|
| 139 |
-
# Delete the audio file
|
| 140 |
-
os.remove(file_path)
|
| 141 |
-
logger.info(f"Deleted audio file: {filename}")
|
| 142 |
-
|
| 143 |
-
return {"message": "File and transcript deleted successfully"}
|
| 144 |
-
|
| 145 |
-
except ValueError:
|
| 146 |
-
logger.error(f"Could not determine podcast ID for file: {filename}")
|
| 147 |
-
# Still delete the audio file even if transcript removal fails
|
| 148 |
-
os.remove(file_path)
|
| 149 |
-
return {"message": "Audio file deleted, but transcript could not be removed"}
|
| 150 |
-
|
| 151 |
-
except Exception as e:
|
| 152 |
-
logger.error(f"Error in delete_audio_file: {str(e)}")
|
| 153 |
-
raise HTTPException(status_code=500, detail=str(e))
|
| 154 |
-
|
| 155 |
@app.get("/audio/{filename}")
|
| 156 |
async def get_audio_file(filename: str):
|
| 157 |
"""Get an audio file by filename."""
|
|
@@ -288,42 +271,6 @@ async def get_podcast_context(podcast_id: str):
|
|
| 288 |
logger.error(f"Error in get_podcast_context: {str(e)}", exc_info=True)
|
| 289 |
raise HTTPException(status_code=500, detail=str(e))
|
| 290 |
|
| 291 |
-
@app.post("/chat")
|
| 292 |
-
async def chat(message: ChatMessage):
|
| 293 |
-
"""Process a chat message."""
|
| 294 |
-
try:
|
| 295 |
-
# Get API key
|
| 296 |
-
tavily_api_key = os.getenv("TAVILY_API_KEY")
|
| 297 |
-
if not tavily_api_key:
|
| 298 |
-
logger.error("Tavily API key not found")
|
| 299 |
-
raise HTTPException(status_code=500, detail="Tavily API key not configured")
|
| 300 |
-
|
| 301 |
-
# Initialize the workflow
|
| 302 |
-
try:
|
| 303 |
-
workflow = create_workflow(tavily_api_key)
|
| 304 |
-
logger.info("Workflow created successfully")
|
| 305 |
-
except Exception as e:
|
| 306 |
-
logger.error(f"Error creating workflow: {str(e)}")
|
| 307 |
-
raise HTTPException(status_code=500, detail=f"Error creating workflow: {str(e)}")
|
| 308 |
-
|
| 309 |
-
# Run the workflow with context
|
| 310 |
-
try:
|
| 311 |
-
result = await run_workflow(
|
| 312 |
-
workflow,
|
| 313 |
-
message.content,
|
| 314 |
-
agent_type=message.agent_type,
|
| 315 |
-
context=message.context
|
| 316 |
-
)
|
| 317 |
-
logger.info("Workflow completed successfully")
|
| 318 |
-
return result
|
| 319 |
-
except Exception as e:
|
| 320 |
-
logger.error(f"Error running workflow: {str(e)}")
|
| 321 |
-
raise HTTPException(status_code=500, detail=f"Error running workflow: {str(e)}")
|
| 322 |
-
|
| 323 |
-
except Exception as e:
|
| 324 |
-
logger.error(f"Error in chat endpoint: {str(e)}", exc_info=True)
|
| 325 |
-
raise HTTPException(status_code=500, detail=str(e))
|
| 326 |
-
|
| 327 |
@app.post("/podcast-chat/{podcast_id}", response_model=PodcastChatResponse)
|
| 328 |
async def podcast_chat(podcast_id: str, request: PodcastChatRequest):
|
| 329 |
"""Handle chat messages for a specific podcast."""
|
|
@@ -429,6 +376,59 @@ async def podcast_chat(podcast_id: str, request: PodcastChatRequest):
|
|
| 429 |
logger.error(f"Error in podcast chat: {str(e)}", exc_info=True)
|
| 430 |
raise HTTPException(status_code=500, detail=str(e))
|
| 431 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 432 |
if __name__ == "__main__":
|
| 433 |
import uvicorn
|
| 434 |
uvicorn.run(app, host="0.0.0.0", port=8000)
|
|
|
|
| 39 |
openai_api_key=openai_api_key
|
| 40 |
)
|
| 41 |
|
| 42 |
+
# Define Pydantic models
|
| 43 |
+
class ChatMessage(BaseModel):
|
| 44 |
+
content: str
|
| 45 |
+
context: Optional[Dict[str, Any]] = None
|
| 46 |
+
agent_type: Optional[str] = "believer"
|
| 47 |
+
|
| 48 |
+
class WorkflowResponse(BaseModel):
|
| 49 |
+
debate_history: List[Dict[str, str]]
|
| 50 |
+
supervisor_notes: List[str]
|
| 51 |
+
supervisor_chunks: List[Dict[str, List[str]]]
|
| 52 |
+
extractor_data: Dict[str, Any]
|
| 53 |
+
final_podcast: Dict[str, Any]
|
| 54 |
+
|
| 55 |
+
class PodcastChatRequest(BaseModel):
|
| 56 |
+
message: str
|
| 57 |
+
|
| 58 |
+
class PodcastChatResponse(BaseModel):
|
| 59 |
+
response: str
|
| 60 |
+
|
| 61 |
# Initialize FastAPI app
|
| 62 |
app = FastAPI()
|
| 63 |
|
| 64 |
+
# Configure CORS
|
| 65 |
app.add_middleware(
|
| 66 |
CORSMiddleware,
|
| 67 |
allow_origins=["http://localhost:5173", "http://localhost:3000", "https://*.hf.space", "*"],
|
|
|
|
| 72 |
max_age=600,
|
| 73 |
)
|
| 74 |
|
| 75 |
+
# Configure storage directories
|
|
|
|
|
|
|
|
|
|
| 76 |
audio_dir = os.path.join(os.path.dirname(__file__), "audio_storage")
|
| 77 |
os.makedirs(audio_dir, exist_ok=True)
|
| 78 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
context_dir = os.path.join(os.path.dirname(__file__), "context_storage")
|
| 80 |
os.makedirs(context_dir, exist_ok=True)
|
| 81 |
|
| 82 |
+
# API Routes
|
| 83 |
+
@app.post("/chat")
|
| 84 |
+
async def chat(message: ChatMessage):
|
| 85 |
+
"""Process a chat message."""
|
| 86 |
+
try:
|
| 87 |
+
# Get API key
|
| 88 |
+
tavily_api_key = os.getenv("TAVILY_API_KEY")
|
| 89 |
+
if not tavily_api_key:
|
| 90 |
+
logger.error("Tavily API key not found")
|
| 91 |
+
raise HTTPException(status_code=500, detail="Tavily API key not configured")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
|
| 93 |
+
# Initialize the workflow
|
| 94 |
+
try:
|
| 95 |
+
workflow = create_workflow(tavily_api_key)
|
| 96 |
+
logger.info("Workflow created successfully")
|
| 97 |
+
except Exception as e:
|
| 98 |
+
logger.error(f"Error creating workflow: {str(e)}")
|
| 99 |
+
raise HTTPException(status_code=500, detail=f"Error creating workflow: {str(e)}")
|
| 100 |
+
|
| 101 |
+
# Run the workflow with context
|
| 102 |
+
try:
|
| 103 |
+
result = await run_workflow(
|
| 104 |
+
workflow,
|
| 105 |
+
message.content,
|
| 106 |
+
agent_type=message.agent_type,
|
| 107 |
+
context=message.context
|
| 108 |
+
)
|
| 109 |
+
logger.info("Workflow completed successfully")
|
| 110 |
+
return result
|
| 111 |
+
except Exception as e:
|
| 112 |
+
logger.error(f"Error running workflow: {str(e)}")
|
| 113 |
+
raise HTTPException(status_code=500, detail=f"Error running workflow: {str(e)}")
|
| 114 |
+
|
| 115 |
+
except Exception as e:
|
| 116 |
+
logger.error(f"Error in chat endpoint: {str(e)}", exc_info=True)
|
| 117 |
+
raise HTTPException(status_code=500, detail=str(e))
|
| 118 |
|
| 119 |
@app.get("/audio-list")
|
| 120 |
async def list_audio_files():
|
|
|
|
| 134 |
except Exception as e:
|
| 135 |
raise HTTPException(status_code=500, detail=str(e))
|
| 136 |
|
| 137 |
+
# Add other API routes here...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
@app.get("/audio/{filename}")
|
| 139 |
async def get_audio_file(filename: str):
|
| 140 |
"""Get an audio file by filename."""
|
|
|
|
| 271 |
logger.error(f"Error in get_podcast_context: {str(e)}", exc_info=True)
|
| 272 |
raise HTTPException(status_code=500, detail=str(e))
|
| 273 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 274 |
@app.post("/podcast-chat/{podcast_id}", response_model=PodcastChatResponse)
|
| 275 |
async def podcast_chat(podcast_id: str, request: PodcastChatRequest):
|
| 276 |
"""Handle chat messages for a specific podcast."""
|
|
|
|
| 376 |
logger.error(f"Error in podcast chat: {str(e)}", exc_info=True)
|
| 377 |
raise HTTPException(status_code=500, detail=str(e))
|
| 378 |
|
| 379 |
+
@app.delete("/audio/{filename}")
|
| 380 |
+
async def delete_audio_file(filename: str):
|
| 381 |
+
"""Delete an audio file and its corresponding transcript."""
|
| 382 |
+
try:
|
| 383 |
+
# Delete audio file
|
| 384 |
+
file_path = os.path.join(audio_dir, filename)
|
| 385 |
+
if not os.path.exists(file_path):
|
| 386 |
+
raise HTTPException(status_code=404, detail="File not found")
|
| 387 |
+
|
| 388 |
+
# Get all audio files to determine the podcast ID
|
| 389 |
+
audio_files = [f for f in os.listdir(audio_dir) if f.endswith(('.mp3', '.wav'))]
|
| 390 |
+
try:
|
| 391 |
+
# Find the index (0-based) of the file being deleted
|
| 392 |
+
podcast_id = audio_files.index(filename) + 1 # Convert to 1-based ID
|
| 393 |
+
logger.info(f"Deleting podcast with ID: {podcast_id}")
|
| 394 |
+
|
| 395 |
+
# Path to transcripts file
|
| 396 |
+
transcripts_file = os.path.join(os.path.dirname(__file__), "transcripts", "podcasts.json")
|
| 397 |
+
|
| 398 |
+
# Update transcripts if file exists
|
| 399 |
+
if os.path.exists(transcripts_file):
|
| 400 |
+
with open(transcripts_file, 'r') as f:
|
| 401 |
+
transcripts = json.load(f)
|
| 402 |
+
|
| 403 |
+
# Remove the transcript at the corresponding index
|
| 404 |
+
if len(transcripts) >= podcast_id:
|
| 405 |
+
transcripts.pop(podcast_id - 1) # Convert back to 0-based index
|
| 406 |
+
|
| 407 |
+
# Save updated transcripts
|
| 408 |
+
with open(transcripts_file, 'w') as f:
|
| 409 |
+
json.dump(transcripts, f, indent=2)
|
| 410 |
+
logger.info(f"Removed transcript for podcast ID {podcast_id}")
|
| 411 |
+
|
| 412 |
+
# Delete the audio file
|
| 413 |
+
os.remove(file_path)
|
| 414 |
+
logger.info(f"Deleted audio file: {filename}")
|
| 415 |
+
|
| 416 |
+
return {"message": "File and transcript deleted successfully"}
|
| 417 |
+
|
| 418 |
+
except ValueError:
|
| 419 |
+
logger.error(f"Could not determine podcast ID for file: {filename}")
|
| 420 |
+
# Still delete the audio file even if transcript removal fails
|
| 421 |
+
os.remove(file_path)
|
| 422 |
+
return {"message": "Audio file deleted, but transcript could not be removed"}
|
| 423 |
+
|
| 424 |
+
except Exception as e:
|
| 425 |
+
logger.error(f"Error in delete_audio_file: {str(e)}")
|
| 426 |
+
raise HTTPException(status_code=500, detail=str(e))
|
| 427 |
+
|
| 428 |
+
# Static file mounts (must be last)
|
| 429 |
+
app.mount("/audio-files", StaticFiles(directory=audio_dir), name="audio")
|
| 430 |
+
app.mount("/", StaticFiles(directory="static", html=True), name="frontend")
|
| 431 |
+
|
| 432 |
if __name__ == "__main__":
|
| 433 |
import uvicorn
|
| 434 |
uvicorn.run(app, host="0.0.0.0", port=8000)
|