Support server based rollout in Verlengine (#4848)

Co-authored-by: Jin Pan <jpan236@wisc.edu>
Co-authored-by: Chayenne <zhaochen20@outlook.com>
Co-authored-by: Jinn <47354855+jhinpan@users.noreply.github.com>

python/sglang/test/test_utils.py

@@ -25,7 +25,12 @@ from sglang.bench_serving import run_benchmark
 from sglang.global_config import global_config
 from sglang.lang.backend.openai import OpenAI
 from sglang.lang.backend.runtime_endpoint import RuntimeEndpoint
-from sglang.srt.utils import get_bool_env_var, kill_process_tree, retry
+from sglang.srt.utils import (
+    get_bool_env_var,
+    is_port_available,
+    kill_process_tree,
+    retry,
+)
 from sglang.test.run_eval import run_eval
 from sglang.utils import get_exception_traceback
 
@@ -98,6 +103,17 @@ def call_generate_lightllm(prompt, temperature, max_tokens, stop=None, url=None)
     return pred
 
 
+def find_available_port(base_port: int):
+    port = base_port + random.randint(100, 1000)
+    while True:
+        if is_port_available(port):
+            return port
+        if port < 60000:
+            port += 42
+        else:
+            port -= 43
+
+
 def call_generate_vllm(prompt, temperature, max_tokens, stop=None, n=1, url=None):
     assert url is not None