[3/N] CI refactor: move some manually triggered tests. (#13448)

test/lang/test_litellm_backend.py

@@ -1,24 +0,0 @@
-import unittest
-
-from sglang import LiteLLM, set_default_backend
-from sglang.test.test_programs import test_mt_bench, test_stream
-from sglang.test.test_utils import CustomTestCase
-
-
-class TestAnthropicBackend(CustomTestCase):
-    chat_backend = None
-
-    @classmethod
-    def setUpClass(cls):
-        cls.chat_backend = LiteLLM("gpt-3.5-turbo")
-        set_default_backend(cls.chat_backend)
-
-    def test_mt_bench(self):
-        test_mt_bench()
-
-    def test_stream(self):
-        test_stream()
-
-
-if __name__ == "__main__":
-    unittest.main()

test/srt/run_suite.py

@@ -235,13 +235,6 @@ suites = {
         TestFile("debug_utils/test_log_parser.py", 5),
         TestFile("test_deepseek_v3_cutedsl_4gpu.py"),
         TestFile("entrypoints/http_server/test_abort_request.py"),
-        TestFile("ep/test_deepep_internode.py"),
-        TestFile("ep/test_deepep_intranode.py"),
-        TestFile("ep/test_deepep_low_latency.py"),
-        TestFile("ep/test_eplb.py"),
-        TestFile("ep/test_hybrid_dp_ep_tp_mtp.py"),
-        TestFile("ep/test_moe_deepep.py"),
-        TestFile("ep/test_moe_deepep_eval_accuracy_large.py"),
         TestFile("hicache/test_disaggregation_hicache.py"),
         TestFile("layers/attention/nsa/test_act_quant_triton.py"),
         TestFile("layers/moe/test_moe_runners.py"),

test/srt/test_session_control.py

@@ -594,8 +594,8 @@ class TestSessionControlVision(CustomTestCase):
             "<|im_start|>user\nDescribe this image in a very short sentence.<|im_end|>\nassistant:",
         ]
         image_chunks = [
-            "https://raw.githubusercontent.com/sgl-project/sglang/main/test/lang/example_image.png",
-            "https://raw.githubusercontent.com/sgl-project/sglang/main/test/lang/example_image.png",
+            "https://raw.githubusercontent.com/sgl-project/sglang/main/examples/assets/example_image.png",
+            "https://raw.githubusercontent.com/sgl-project/sglang/main/examples/assets/example_image.png",
             "https://raw.githubusercontent.com/sgl-project/sglang/main/assets/logo.png",
         ]
 

test/srt/test_vlm_accuracy.py

@@ -30,7 +30,7 @@ from sglang.srt.server_args import ServerArgs
 class VisionLLMLogitsBase(unittest.IsolatedAsyncioTestCase):
     @classmethod
     def setUpClass(cls):
-        cls.image_url = "https://github.com/sgl-project/sglang/blob/main/test/lang/example_image.png?raw=true"
+        cls.image_url = "https://github.com/sgl-project/sglang/blob/main/examples/assets/example_image.png?raw=true"
         cls.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
         cls.model_path = ""
         cls.chat_template = ""

test/srt/test_vlm_input_format.py

@@ -16,7 +16,7 @@ from sglang import Engine
 from sglang.srt.entrypoints.openai.protocol import ChatCompletionRequest
 from sglang.srt.parser.conversation import generate_chat_conv
 
-TEST_IMAGE_URL = "https://github.com/sgl-project/sglang/blob/main/test/lang/example_image.png?raw=true"
+TEST_IMAGE_URL = "https://github.com/sgl-project/sglang/blob/main/examples/assets/example_image.png?raw=true"
 
 
 class VLMInputTestBase: