Spaces:
Sleeping
Sleeping
Update app/main.py
Browse files- app/main.py +3 -3
app/main.py
CHANGED
|
@@ -28,7 +28,7 @@ from app.utils.llm import ask_question, ask_arxiv_question
|
|
| 28 |
|
| 29 |
# ====== Config via ENV ======
|
| 30 |
FILE_MAX_MB: int = int(os.getenv("FILE_MAX_MB", "5")) # Max .txt upload size for endpoints
|
| 31 |
-
LATENCY_TIMEOUT_MS: int = int(os.getenv("LATENCY_TIMEOUT_MS", "
|
| 32 |
REQUEST_TIMEOUT_MS: int = int(os.getenv("REQUEST_TIMEOUT_MS", "15000")) # Optional timeout for /query
|
| 33 |
CORS_ORIGINS: str = os.getenv("CORS_ORIGINS", "*") # comma-separated list or "*"
|
| 34 |
|
|
@@ -270,7 +270,7 @@ async def query_llm(
|
|
| 270 |
input: LLMQueryInput,
|
| 271 |
mode: Mode = Query(Mode.short, description="Answer length"),
|
| 272 |
model_name: ModelName = Query(ModelName.mistral_v02, description="Choose the LLM"),
|
| 273 |
-
request_timeout_ms: int = Query(REQUEST_TIMEOUT_MS, ge=1000, le=
|
| 274 |
)
|
| 275 |
model_id = model_name.value
|
| 276 |
mode_val = mode.value
|
|
@@ -381,7 +381,7 @@ def build_ui():
|
|
| 381 |
with gr.Row():
|
| 382 |
mode = gr.Radio(["short", "detailed"], value="short", label="Answer length")
|
| 383 |
model_id = gr.Dropdown(MODEL_IDS, value=ModelName.mistral_v02.value, label="Choose the LLM")
|
| 384 |
-
request_timeout_ms = gr.Slider(1000,
|
| 385 |
help_btn = gr.Button("HELP", variant="secondary")
|
| 386 |
refresh_latency_btn = gr.Button("Refresh latency", variant="secondary")
|
| 387 |
|
|
|
|
| 28 |
|
| 29 |
# ====== Config via ENV ======
|
| 30 |
FILE_MAX_MB: int = int(os.getenv("FILE_MAX_MB", "5")) # Max .txt upload size for endpoints
|
| 31 |
+
LATENCY_TIMEOUT_MS: int = int(os.getenv("LATENCY_TIMEOUT_MS", "900000"))
|
| 32 |
REQUEST_TIMEOUT_MS: int = int(os.getenv("REQUEST_TIMEOUT_MS", "15000")) # Optional timeout for /query
|
| 33 |
CORS_ORIGINS: str = os.getenv("CORS_ORIGINS", "*") # comma-separated list or "*"
|
| 34 |
|
|
|
|
| 270 |
input: LLMQueryInput,
|
| 271 |
mode: Mode = Query(Mode.short, description="Answer length"),
|
| 272 |
model_name: ModelName = Query(ModelName.mistral_v02, description="Choose the LLM"),
|
| 273 |
+
request_timeout_ms: int = Query(REQUEST_TIMEOUT_MS, ge=1000, le=900000, description="Per-request timeout in ms")
|
| 274 |
)
|
| 275 |
model_id = model_name.value
|
| 276 |
mode_val = mode.value
|
|
|
|
| 381 |
with gr.Row():
|
| 382 |
mode = gr.Radio(["short", "detailed"], value="short", label="Answer length")
|
| 383 |
model_id = gr.Dropdown(MODEL_IDS, value=ModelName.mistral_v02.value, label="Choose the LLM")
|
| 384 |
+
request_timeout_ms = gr.Slider(1000, 900000, value=REQUEST_TIMEOUT_MS, step=500, label="Timeout (ms)")
|
| 385 |
help_btn = gr.Button("HELP", variant="secondary")
|
| 386 |
refresh_latency_btn = gr.Button("Refresh latency", variant="secondary")
|
| 387 |
|