Spaces:
Paused
Paused
Update server.py
Browse files
server.py
CHANGED
@@ -1,39 +1,35 @@
|
|
1 |
import os
|
2 |
-
|
3 |
import requests
|
4 |
import logging
|
|
|
|
|
|
|
5 |
from dotenv import load_dotenv
|
6 |
-
import json
|
7 |
-
from typing import Dict, Any, Tuple, Union
|
8 |
|
9 |
# Load environment variables
|
10 |
load_dotenv()
|
11 |
|
12 |
-
|
|
|
|
|
13 |
|
|
|
14 |
ANTHROPIC_API_URL = os.getenv(
|
15 |
'ANTHROPIC_API_URL',
|
16 |
'https://relay.stagwellmarketingcloud.io/anthropic/v1/messages'
|
17 |
)
|
18 |
API_KEY = os.getenv('ANTHROPIC_API_KEY')
|
19 |
|
20 |
-
|
21 |
-
logging.basicConfig(level=logging.INFO)
|
22 |
-
logger = logging.getLogger(__name__)
|
23 |
-
|
24 |
-
def transform_to_anthropic_request(
|
25 |
-
request_data: Dict[str, Any]
|
26 |
-
) -> Dict[str, Any]:
|
27 |
"""Transform OpenAI-style request to Anthropic format."""
|
28 |
if 'messages' in request_data:
|
29 |
-
# Chat completion format
|
30 |
return {
|
31 |
"model": request_data.get('model', 'claude-3-5-sonnet-20240620'),
|
32 |
"max_tokens": request_data.get('max_tokens', 1024),
|
33 |
"messages": request_data['messages']
|
34 |
}
|
35 |
else:
|
36 |
-
# Regular completion format
|
37 |
return {
|
38 |
"model": request_data.get('model', 'claude-3-5-sonnet-20240620'),
|
39 |
"max_tokens": request_data.get('max_tokens', 1024),
|
@@ -79,59 +75,102 @@ def make_anthropic_request(
|
|
79 |
logger.error(error_msg)
|
80 |
return {"error": error_msg}, 500
|
81 |
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
if status_code != 200:
|
90 |
-
return jsonify(anthropic_response), status_code
|
91 |
-
|
92 |
-
return jsonify({
|
93 |
-
"id": "cmpl-" + anthropic_response.get('id', 'default'),
|
94 |
-
"object": "text_completion",
|
95 |
-
"created": anthropic_response.get('created', 0),
|
96 |
-
"choices": [{
|
97 |
-
"text": anthropic_response['content'][0]['text'],
|
98 |
-
"index": 0,
|
99 |
-
"finish_reason": "stop"
|
100 |
-
}],
|
101 |
-
"usage": {
|
102 |
-
"prompt_tokens": -1,
|
103 |
-
"completion_tokens": -1,
|
104 |
-
"total_tokens": -1
|
105 |
-
}
|
106 |
-
}), 200
|
107 |
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
116 |
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
-
|
134 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
|
136 |
-
if __name__ ==
|
137 |
-
|
|
|
1 |
import os
|
2 |
+
import json
|
3 |
import requests
|
4 |
import logging
|
5 |
+
from http.server import SimpleHTTPRequestHandler, ThreadingHTTPServer
|
6 |
+
from urllib.parse import parse_qs, urlparse
|
7 |
+
from typing import Dict, Any, Tuple
|
8 |
from dotenv import load_dotenv
|
|
|
|
|
9 |
|
10 |
# Load environment variables
|
11 |
load_dotenv()
|
12 |
|
13 |
+
# Configure logging
|
14 |
+
logging.basicConfig(level=logging.INFO)
|
15 |
+
logger = logging.getLogger(__name__)
|
16 |
|
17 |
+
# Constants
|
18 |
ANTHROPIC_API_URL = os.getenv(
|
19 |
'ANTHROPIC_API_URL',
|
20 |
'https://relay.stagwellmarketingcloud.io/anthropic/v1/messages'
|
21 |
)
|
22 |
API_KEY = os.getenv('ANTHROPIC_API_KEY')
|
23 |
|
24 |
+
def transform_to_anthropic_request(request_data: Dict[str, Any]) -> Dict[str, Any]:
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
"""Transform OpenAI-style request to Anthropic format."""
|
26 |
if 'messages' in request_data:
|
|
|
27 |
return {
|
28 |
"model": request_data.get('model', 'claude-3-5-sonnet-20240620'),
|
29 |
"max_tokens": request_data.get('max_tokens', 1024),
|
30 |
"messages": request_data['messages']
|
31 |
}
|
32 |
else:
|
|
|
33 |
return {
|
34 |
"model": request_data.get('model', 'claude-3-5-sonnet-20240620'),
|
35 |
"max_tokens": request_data.get('max_tokens', 1024),
|
|
|
75 |
logger.error(error_msg)
|
76 |
return {"error": error_msg}, 500
|
77 |
|
78 |
+
class AnthropicProxyHandler(SimpleHTTPRequestHandler):
|
79 |
+
def send_json_response(self, data: Dict[str, Any], status: int = 200):
|
80 |
+
"""Helper method to send JSON responses."""
|
81 |
+
self.send_response(status)
|
82 |
+
self.send_header('Content-Type', 'application/json')
|
83 |
+
self.end_headers()
|
84 |
+
self.wfile.write(json.dumps(data).encode('utf-8'))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
|
86 |
+
def do_GET(self):
|
87 |
+
"""Handle GET requests."""
|
88 |
+
if self.path == "/v1":
|
89 |
+
self.send_json_response({"status": True})
|
90 |
+
elif self.path == "/v1/models":
|
91 |
+
self.send_json_response({
|
92 |
+
"data": [{
|
93 |
+
"id": "claude-3-5-sonnet-20240620",
|
94 |
+
"object": "model",
|
95 |
+
"created": 1706745202,
|
96 |
+
"owned_by": "anthropic"
|
97 |
+
}]
|
98 |
+
})
|
99 |
+
else:
|
100 |
+
super().do_GET()
|
101 |
+
|
102 |
+
def do_POST(self):
|
103 |
+
"""Handle POST requests."""
|
104 |
+
content_length = int(self.headers.get('Content-Length', 0))
|
105 |
+
request_body = self.rfile.read(content_length)
|
106 |
+
request_data = json.loads(request_body)
|
107 |
|
108 |
+
if self.path == "/v1/completions":
|
109 |
+
logger.info(f"Received completion request: {json.dumps(request_data)}")
|
110 |
+
|
111 |
+
response, status_code = make_anthropic_request(request_data)
|
112 |
+
if status_code != 200:
|
113 |
+
self.send_json_response(response, status_code)
|
114 |
+
return
|
115 |
+
|
116 |
+
completion_response = {
|
117 |
+
"id": "cmpl-" + response.get('id', 'default'),
|
118 |
+
"object": "text_completion",
|
119 |
+
"created": response.get('created', 0),
|
120 |
+
"choices": [{
|
121 |
+
"text": response['content'][0]['text'],
|
122 |
+
"index": 0,
|
123 |
+
"finish_reason": "stop"
|
124 |
+
}],
|
125 |
+
"usage": {
|
126 |
+
"prompt_tokens": -1,
|
127 |
+
"completion_tokens": -1,
|
128 |
+
"total_tokens": -1
|
129 |
+
}
|
130 |
+
}
|
131 |
+
self.send_json_response(completion_response)
|
132 |
+
|
133 |
+
elif self.path == "/v1/chat/completions":
|
134 |
+
logger.info(
|
135 |
+
f"Received chat completion request: {json.dumps(request_data)}"
|
136 |
+
)
|
137 |
+
|
138 |
+
response, status_code = make_anthropic_request(request_data)
|
139 |
+
if status_code != 200:
|
140 |
+
self.send_json_response(response, status_code)
|
141 |
+
return
|
142 |
+
|
143 |
+
chat_response = {
|
144 |
+
"id": "chatcmpl-" + response.get('id', 'default'),
|
145 |
+
"object": "chat.completion",
|
146 |
+
"created": response.get('created', 0),
|
147 |
+
"choices": [{
|
148 |
+
"index": 0,
|
149 |
+
"message": {
|
150 |
+
"role": "assistant",
|
151 |
+
"content": response['content'][0]['text']
|
152 |
+
},
|
153 |
+
"finish_reason": "stop"
|
154 |
+
}],
|
155 |
+
"usage": {
|
156 |
+
"prompt_tokens": -1,
|
157 |
+
"completion_tokens": -1,
|
158 |
+
"total_tokens": -1
|
159 |
+
}
|
160 |
+
}
|
161 |
+
self.send_json_response(chat_response)
|
162 |
+
|
163 |
+
else:
|
164 |
+
self.send_json_response(
|
165 |
+
{"error": "Invalid endpoint"},
|
166 |
+
status=404
|
167 |
+
)
|
168 |
+
|
169 |
+
def run_server(port: int = 7860):
|
170 |
+
"""Start the server on the specified port."""
|
171 |
+
server = ThreadingHTTPServer(("", port), AnthropicProxyHandler)
|
172 |
+
logger.info(f"Starting server on port {port}")
|
173 |
+
server.serve_forever()
|
174 |
|
175 |
+
if __name__ == "__main__":
|
176 |
+
run_server()
|