Spaces:
Running
Running
Update main.py
Browse files
main.py
CHANGED
@@ -32,20 +32,6 @@ if not secret_api_endpoint or not secret_api_endpoint_2 or not secret_api_endpoi
|
|
32 |
alternate_models = {"gpt-4o-mini", "claude-3-haiku", "llama-3.1-70b", "mixtral-8x7b"}
|
33 |
|
34 |
available_model_ids = []
|
35 |
-
def check_server_status():
|
36 |
-
server_down = True
|
37 |
-
|
38 |
-
def decorator(func):
|
39 |
-
async def wrapper(*args, **kwargs):
|
40 |
-
if server_down:
|
41 |
-
raise HTTPException(
|
42 |
-
status_code=503,
|
43 |
-
detail="Server is currently unavailable. All services are temporarily down."
|
44 |
-
)
|
45 |
-
return await func(*args, **kwargs)
|
46 |
-
return wrapper
|
47 |
-
return decorator
|
48 |
-
|
49 |
class Payload(BaseModel):
|
50 |
model: str
|
51 |
messages: list
|
@@ -157,30 +143,41 @@ async def fetch_models():
|
|
157 |
return await get_models()
|
158 |
|
159 |
|
|
|
|
|
|
|
|
|
|
|
160 |
@app.post("/chat/completions")
|
161 |
@app.post("/v1/chat/completions")
|
162 |
-
|
163 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
164 |
model_to_use = payload.model if payload.model else "gpt-4o-mini"
|
|
|
165 |
# Validate model availability
|
166 |
if model_to_use not in available_model_ids:
|
167 |
raise HTTPException(
|
168 |
-
status_code=400,
|
169 |
detail=f"Model '{model_to_use}' is not available. Check /models for the available model list."
|
170 |
)
|
171 |
-
|
172 |
# Proceed with the request handling
|
173 |
payload_dict = payload.dict()
|
174 |
payload_dict["model"] = model_to_use
|
175 |
-
|
176 |
# Select the appropriate endpoint
|
177 |
endpoint = secret_api_endpoint_2 if model_to_use in alternate_models else secret_api_endpoint
|
178 |
current_time = (datetime.datetime.utcnow() + datetime.timedelta(hours=5, minutes=30)).strftime("%Y-%m-%d %I:%M:%S %p")
|
179 |
aaip = request.client.host
|
180 |
print(f"Time: {current_time}, {aaip}")
|
181 |
print(payload_dict)
|
182 |
-
|
183 |
-
|
184 |
async def stream_generator(payload_dict):
|
185 |
scraper = cloudscraper.create_scraper() # Create a CloudScraper session
|
186 |
try:
|
@@ -198,12 +195,11 @@ async def get_completion(payload: Payload,request: Request):
|
|
198 |
raise HTTPException(status_code=404, detail="The requested resource was not found.")
|
199 |
elif response.status_code >= 500:
|
200 |
raise HTTPException(status_code=500, detail="Server error. Try again later.")
|
201 |
-
|
202 |
# Stream response lines to the client
|
203 |
for line in response.iter_lines():
|
204 |
if line:
|
205 |
yield line.decode('utf-8') + "\n"
|
206 |
-
|
207 |
except requests.exceptions.RequestException as req_err:
|
208 |
# Handle request-specific errors
|
209 |
print(response.text)
|
@@ -212,7 +208,7 @@ async def get_completion(payload: Payload,request: Request):
|
|
212 |
# Handle unexpected errors
|
213 |
print(response.text)
|
214 |
raise HTTPException(status_code=500, detail=f"An unexpected error occurred: {e}")
|
215 |
-
|
216 |
return StreamingResponse(stream_generator(payload_dict), media_type="application/json")
|
217 |
# Remove the duplicated endpoint and combine the functionality
|
218 |
@app.get("/images/generations") #pollinations.ai thanks to them :)
|
|
|
32 |
alternate_models = {"gpt-4o-mini", "claude-3-haiku", "llama-3.1-70b", "mixtral-8x7b"}
|
33 |
|
34 |
available_model_ids = []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
class Payload(BaseModel):
|
36 |
model: str
|
37 |
messages: list
|
|
|
143 |
return await get_models()
|
144 |
|
145 |
|
146 |
+
from fastapi import HTTPException
|
147 |
+
from fastapi.responses import JSONResponse
|
148 |
+
|
149 |
+
server_status = False # Set to False to simulate the server being down
|
150 |
+
|
151 |
@app.post("/chat/completions")
|
152 |
@app.post("/v1/chat/completions")
|
153 |
+
async def get_completion(payload: Payload, request: Request):
|
154 |
+
# Check server status
|
155 |
+
if not server_status:
|
156 |
+
return JSONResponse(
|
157 |
+
status_code=503,
|
158 |
+
content={"message": "Server is down. Please try again later."}
|
159 |
+
)
|
160 |
+
|
161 |
model_to_use = payload.model if payload.model else "gpt-4o-mini"
|
162 |
+
|
163 |
# Validate model availability
|
164 |
if model_to_use not in available_model_ids:
|
165 |
raise HTTPException(
|
166 |
+
status_code=400,
|
167 |
detail=f"Model '{model_to_use}' is not available. Check /models for the available model list."
|
168 |
)
|
169 |
+
|
170 |
# Proceed with the request handling
|
171 |
payload_dict = payload.dict()
|
172 |
payload_dict["model"] = model_to_use
|
173 |
+
|
174 |
# Select the appropriate endpoint
|
175 |
endpoint = secret_api_endpoint_2 if model_to_use in alternate_models else secret_api_endpoint
|
176 |
current_time = (datetime.datetime.utcnow() + datetime.timedelta(hours=5, minutes=30)).strftime("%Y-%m-%d %I:%M:%S %p")
|
177 |
aaip = request.client.host
|
178 |
print(f"Time: {current_time}, {aaip}")
|
179 |
print(payload_dict)
|
180 |
+
|
|
|
181 |
async def stream_generator(payload_dict):
|
182 |
scraper = cloudscraper.create_scraper() # Create a CloudScraper session
|
183 |
try:
|
|
|
195 |
raise HTTPException(status_code=404, detail="The requested resource was not found.")
|
196 |
elif response.status_code >= 500:
|
197 |
raise HTTPException(status_code=500, detail="Server error. Try again later.")
|
198 |
+
|
199 |
# Stream response lines to the client
|
200 |
for line in response.iter_lines():
|
201 |
if line:
|
202 |
yield line.decode('utf-8') + "\n"
|
|
|
203 |
except requests.exceptions.RequestException as req_err:
|
204 |
# Handle request-specific errors
|
205 |
print(response.text)
|
|
|
208 |
# Handle unexpected errors
|
209 |
print(response.text)
|
210 |
raise HTTPException(status_code=500, detail=f"An unexpected error occurred: {e}")
|
211 |
+
|
212 |
return StreamingResponse(stream_generator(payload_dict), media_type="application/json")
|
213 |
# Remove the duplicated endpoint and combine the functionality
|
214 |
@app.get("/images/generations") #pollinations.ai thanks to them :)
|