Add typo checker in pre-commit (#6179)

Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca>

test/srt/test_openai_server.py

@@ -433,7 +433,7 @@ class TestOpenAIServer(CustomTestCase):
                             )
 
     def test_completion_stream(self):
-        # parallel sampling adn list input are not supported in streaming mode
+        # parallel sampling and list input are not supported in streaming mode
         for echo in [False, True]:
             for logprobs in [None, 5]:
                 for use_list_input in [True, False]:

test/srt/test_session_control.py

@@ -161,7 +161,7 @@ class TestSessionControl(CustomTestCase):
             ]
         )
 
-        # query with a non-existing rid (the last one should be disappeared becuase of backtrack), should see abort
+        # query with a non-existing rid (the last one should be disappeared because of backtrack), should see abort
         response = requests.post(
             self.base_url + "/generate",
             json={
@@ -668,7 +668,7 @@ class TestSessionControlVision(CustomTestCase):
         ).json()
         outputs_from_session.append(response["text"])
 
-        # query with a non-existing rid (the last one should be disappeared becuase of backtrack), should see abort
+        # query with a non-existing rid (the last one should be disappeared because of backtrack), should see abort
         response = requests.post(
             self.base_url + "/generate",
             json={

test/srt/test_srt_endpoint.py

@@ -295,7 +295,7 @@ class TestSRTEndpoint(CustomTestCase):
         print(f"{output_top_logprobs=}")
 
         # Parse results
-        # This is becaues the grammar constraint allows all prefix tokens
+        # This is because the grammar constraint allows all prefix tokens
         logprobs = [None] * 2
         for i in range(len(output_top_logprobs)):
             try:

test/srt/test_srt_engine_with_quant_args.py

@@ -8,7 +8,7 @@ class TestSRTEngineWithQuantArgs(CustomTestCase):
 
     def test_1_quantization_args(self):
 
-        # we only test fp8 because other methods are currenly depend on vllm. We can add other methods back to test after vllm depency is resolved.
+        # we only test fp8 because other methods are currently dependent on vllm. We can add other methods back to test after vllm dependency is resolved.
         quantization_args_list = [
             # "awq",
             "fp8",

test/srt/test_vlm_accuracy.py

@@ -116,7 +116,7 @@ class VisionLLMLogitsBase(unittest.IsolatedAsyncioTestCase):
         }},
         {{
           "type": "text",
-          "text": "Whats in this picture?"
+          "text": "What's in this picture?"
         }}
       ]
     }}