Update main.py
Browse files
main.py
CHANGED
@@ -1,6 +1,10 @@
|
|
1 |
import os
|
2 |
import uuid
|
3 |
import logging
|
|
|
|
|
|
|
|
|
4 |
from datetime import datetime
|
5 |
from typing import Any, Dict, List, Optional
|
6 |
|
@@ -8,7 +12,23 @@ import httpx
|
|
8 |
from fastapi import FastAPI, HTTPException, Depends
|
9 |
from pydantic import BaseModel
|
10 |
from starlette.middleware.cors import CORSMiddleware
|
11 |
-
from starlette.responses import StreamingResponse
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
# Mock implementations for ImageResponse and to_data_uri
|
14 |
class ImageResponse:
|
@@ -20,6 +40,7 @@ def to_data_uri(image: Any) -> str:
|
|
20 |
# Placeholder for actual image encoding
|
21 |
return "data:image/png;base64,..." # Replace with actual base64 data
|
22 |
|
|
|
23 |
class AsyncGeneratorProvider:
|
24 |
pass
|
25 |
|
@@ -33,7 +54,7 @@ class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
|
|
33 |
supports_stream = True
|
34 |
supports_system_message = True
|
35 |
supports_message_history = True
|
36 |
-
|
37 |
default_model = 'blackbox'
|
38 |
models = [
|
39 |
'blackbox',
|
@@ -58,13 +79,13 @@ class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
|
|
58 |
'llama-3.1-70b': {'mode': True, 'id': "llama-3.1-70b"},
|
59 |
'llama-3.1-405b': {'mode': True, 'id': "llama-3.1-405b"},
|
60 |
}
|
61 |
-
|
62 |
userSelectedModel = {
|
63 |
"gpt-4o": "gpt-4o",
|
64 |
"gemini-pro": "gemini-pro",
|
65 |
'claude-sonnet-3.5': "claude-sonnet-3.5",
|
66 |
}
|
67 |
-
|
68 |
model_aliases = {
|
69 |
"gemini-flash": "gemini-1.5-flash",
|
70 |
"flux": "ImageGenerationLV45LJp",
|
@@ -115,7 +136,7 @@ class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
|
|
115 |
if not messages[0]['content'].startswith(prefix):
|
116 |
messages[0]['content'] = f"{prefix} {messages[0]['content']}"
|
117 |
|
118 |
-
async with
|
119 |
if image is not None:
|
120 |
messages[-1]["data"] = {
|
121 |
"fileText": image_name,
|
@@ -166,16 +187,14 @@ class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
|
|
166 |
else:
|
167 |
raise Exception("Image URL not found in the response")
|
168 |
else:
|
169 |
-
async for chunk in response.
|
170 |
if chunk:
|
171 |
decoded_chunk = chunk.decode()
|
172 |
decoded_chunk = re.sub(r'\$@\$v=[^$]+\$@\$', '', decoded_chunk)
|
173 |
if decoded_chunk.strip():
|
174 |
yield decoded_chunk
|
175 |
|
176 |
-
#
|
177 |
-
app = FastAPI()
|
178 |
-
|
179 |
class Message(BaseModel):
|
180 |
role: str
|
181 |
content: str
|
@@ -183,19 +202,22 @@ class Message(BaseModel):
|
|
183 |
class ChatRequest(BaseModel):
|
184 |
model: str
|
185 |
messages: List[Message]
|
|
|
186 |
|
187 |
-
|
|
|
|
|
|
|
188 |
|
189 |
@app.post("/v1/chat/completions")
|
190 |
-
async def chat_completions(
|
191 |
-
request: ChatRequest, app_secret: str = Depends(verify_app_secret)
|
192 |
-
):
|
193 |
logger.info(f"Received chat completion request for model: {request.model}")
|
194 |
|
195 |
-
|
|
|
196 |
raise HTTPException(
|
197 |
status_code=400,
|
198 |
-
detail=f"Model {request.model} is not allowed. Allowed models are: {', '.join(
|
199 |
)
|
200 |
|
201 |
# Generate a UUID for the conversation
|
@@ -205,10 +227,7 @@ async def chat_completions(
|
|
205 |
"attachments": [],
|
206 |
"conversationId": conversation_id,
|
207 |
"prompt": "\n".join(
|
208 |
-
[
|
209 |
-
f"{'User' if msg.role == 'user' else 'Assistant'}: {msg.content}"
|
210 |
-
for msg in request.messages
|
211 |
-
]
|
212 |
),
|
213 |
}
|
214 |
|
@@ -217,13 +236,12 @@ async def chat_completions(
|
|
217 |
async def generate():
|
218 |
async with httpx.AsyncClient() as client:
|
219 |
try:
|
220 |
-
async with client.stream('POST', f'{
|
221 |
response.raise_for_status()
|
222 |
async for line in response.aiter_lines():
|
223 |
if line and line != "[DONE]":
|
224 |
content = json.loads(line)["data"]
|
225 |
-
yield f"data: {json.dumps(
|
226 |
-
yield f"data: {json.dumps(create_chat_completion_data('', request.model, 'stop'))}\n\n"
|
227 |
yield "data: [DONE]\n\n"
|
228 |
except httpx.HTTPStatusError as e:
|
229 |
logger.error(f"HTTP error occurred: {e}")
|
@@ -233,16 +251,13 @@ async def chat_completions(
|
|
233 |
raise HTTPException(status_code=500, detail=str(e))
|
234 |
|
235 |
if request.stream:
|
236 |
-
logger.info("Streaming response")
|
237 |
return StreamingResponse(generate(), media_type="text/event-stream")
|
238 |
else:
|
239 |
-
logger.info("Non-streaming response")
|
240 |
full_response = ""
|
241 |
async for chunk in generate():
|
242 |
if chunk.startswith("data: ") and not chunk[6:].startswith("[DONE]"):
|
243 |
data = json.loads(chunk[6:])
|
244 |
-
|
245 |
-
full_response += data["choices"][0]["delta"]["content"]
|
246 |
|
247 |
return {
|
248 |
"id": f"chatcmpl-{uuid.uuid4()}",
|
|
|
1 |
import os
|
2 |
import uuid
|
3 |
import logging
|
4 |
+
import json
|
5 |
+
import re
|
6 |
+
import random
|
7 |
+
import string
|
8 |
from datetime import datetime
|
9 |
from typing import Any, Dict, List, Optional
|
10 |
|
|
|
12 |
from fastapi import FastAPI, HTTPException, Depends
|
13 |
from pydantic import BaseModel
|
14 |
from starlette.middleware.cors import CORSMiddleware
|
15 |
+
from starlette.responses import StreamingResponse
|
16 |
+
|
17 |
+
# Setup logging
|
18 |
+
logging.basicConfig(level=logging.INFO)
|
19 |
+
logger = logging.getLogger(__name__)
|
20 |
+
|
21 |
+
# FastAPI app setup
|
22 |
+
app = FastAPI()
|
23 |
+
|
24 |
+
# CORS middleware setup
|
25 |
+
app.add_middleware(
|
26 |
+
CORSMiddleware,
|
27 |
+
allow_origins=["*"],
|
28 |
+
allow_credentials=True,
|
29 |
+
allow_methods=["*"],
|
30 |
+
allow_headers=["*"],
|
31 |
+
)
|
32 |
|
33 |
# Mock implementations for ImageResponse and to_data_uri
|
34 |
class ImageResponse:
|
|
|
40 |
# Placeholder for actual image encoding
|
41 |
return "data:image/png;base64,..." # Replace with actual base64 data
|
42 |
|
43 |
+
# Define models and providers
|
44 |
class AsyncGeneratorProvider:
|
45 |
pass
|
46 |
|
|
|
54 |
supports_stream = True
|
55 |
supports_system_message = True
|
56 |
supports_message_history = True
|
57 |
+
|
58 |
default_model = 'blackbox'
|
59 |
models = [
|
60 |
'blackbox',
|
|
|
79 |
'llama-3.1-70b': {'mode': True, 'id': "llama-3.1-70b"},
|
80 |
'llama-3.1-405b': {'mode': True, 'id': "llama-3.1-405b"},
|
81 |
}
|
82 |
+
|
83 |
userSelectedModel = {
|
84 |
"gpt-4o": "gpt-4o",
|
85 |
"gemini-pro": "gemini-pro",
|
86 |
'claude-sonnet-3.5': "claude-sonnet-3.5",
|
87 |
}
|
88 |
+
|
89 |
model_aliases = {
|
90 |
"gemini-flash": "gemini-1.5-flash",
|
91 |
"flux": "ImageGenerationLV45LJp",
|
|
|
136 |
if not messages[0]['content'].startswith(prefix):
|
137 |
messages[0]['content'] = f"{prefix} {messages[0]['content']}"
|
138 |
|
139 |
+
async with httpx.AsyncClient(headers=headers) as session:
|
140 |
if image is not None:
|
141 |
messages[-1]["data"] = {
|
142 |
"fileText": image_name,
|
|
|
187 |
else:
|
188 |
raise Exception("Image URL not found in the response")
|
189 |
else:
|
190 |
+
async for chunk in response.aiter_bytes():
|
191 |
if chunk:
|
192 |
decoded_chunk = chunk.decode()
|
193 |
decoded_chunk = re.sub(r'\$@\$v=[^$]+\$@\$', '', decoded_chunk)
|
194 |
if decoded_chunk.strip():
|
195 |
yield decoded_chunk
|
196 |
|
197 |
+
# Message and chat request models
|
|
|
|
|
198 |
class Message(BaseModel):
|
199 |
role: str
|
200 |
content: str
|
|
|
202 |
class ChatRequest(BaseModel):
|
203 |
model: str
|
204 |
messages: List[Message]
|
205 |
+
stream: Optional[bool] = False
|
206 |
|
207 |
+
# Verify app secret (placeholder)
|
208 |
+
async def verify_app_secret(app_secret: str):
|
209 |
+
if app_secret != os.getenv("APP_SECRET"):
|
210 |
+
raise HTTPException(status_code=403, detail="Forbidden")
|
211 |
|
212 |
@app.post("/v1/chat/completions")
|
213 |
+
async def chat_completions(request: ChatRequest, app_secret: str = Depends(verify_app_secret)):
|
|
|
|
|
214 |
logger.info(f"Received chat completion request for model: {request.model}")
|
215 |
|
216 |
+
# Validate model
|
217 |
+
if request.model not in Blackbox.models:
|
218 |
raise HTTPException(
|
219 |
status_code=400,
|
220 |
+
detail=f"Model {request.model} is not allowed. Allowed models are: {', '.join(Blackbox.models)}",
|
221 |
)
|
222 |
|
223 |
# Generate a UUID for the conversation
|
|
|
227 |
"attachments": [],
|
228 |
"conversationId": conversation_id,
|
229 |
"prompt": "\n".join(
|
230 |
+
[f"{msg.role}: {msg.content}" for msg in request.messages]
|
|
|
|
|
|
|
231 |
),
|
232 |
}
|
233 |
|
|
|
236 |
async def generate():
|
237 |
async with httpx.AsyncClient() as client:
|
238 |
try:
|
239 |
+
async with client.stream('POST', f'{Blackbox.api_endpoint}', headers=headers, json=json_data, timeout=120.0) as response:
|
240 |
response.raise_for_status()
|
241 |
async for line in response.aiter_lines():
|
242 |
if line and line != "[DONE]":
|
243 |
content = json.loads(line)["data"]
|
244 |
+
yield f"data: {json.dumps(content)}\n\n"
|
|
|
245 |
yield "data: [DONE]\n\n"
|
246 |
except httpx.HTTPStatusError as e:
|
247 |
logger.error(f"HTTP error occurred: {e}")
|
|
|
251 |
raise HTTPException(status_code=500, detail=str(e))
|
252 |
|
253 |
if request.stream:
|
|
|
254 |
return StreamingResponse(generate(), media_type="text/event-stream")
|
255 |
else:
|
|
|
256 |
full_response = ""
|
257 |
async for chunk in generate():
|
258 |
if chunk.startswith("data: ") and not chunk[6:].startswith("[DONE]"):
|
259 |
data = json.loads(chunk[6:])
|
260 |
+
full_response += data.get("choices", [{}])[0].get("delta", {}).get("content", "")
|
|
|
261 |
|
262 |
return {
|
263 |
"id": f"chatcmpl-{uuid.uuid4()}",
|