MCPcopy Create free account
hub / github.com/NVIDIA/OpenShell / call_chat_completions

Function call_chat_completions

e2e/python/test_inference_routing.py:202–229  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

200 spec = datamodel_pb2.SandboxSpec(policy=_baseline_policy())
201
202 def call_chat_completions() -> str:
203 import json
204 import ssl
205 import urllib.request
206
207 body = json.dumps(
208 {
209 "model": "test-model",
210 "messages": [{"role": "user", "content": "hello"}],
211 }
212 ).encode()
213
214 req = urllib.request.Request(
215 "https://inference.local/v1/chat/completions",
216 data=body,
217 headers={
218 "Content-Type": "application/json",
219 "Authorization": "Bearer dummy-key",
220 },
221 method="POST",
222 )
223 # The proxy will TLS-terminate, so we need to accept its cert.
224 ctx = ssl.create_default_context()
225 ctx.check_hostname = False
226 ctx.verify_mode = ssl.CERT_NONE
227
228 resp = urllib.request.urlopen(req, timeout=30, context=ctx)
229 return resp.read().decode()
230
231 with sandbox(spec=spec, delete_on_exit=True) as sb:
232 result = sb.exec_python(call_chat_completions, timeout_seconds=60)

Callers

nothing calls this directly

Calls 1

readMethod · 0.45

Tested by

no test coverage detected