[Frontend][3/n] Improve pooling entrypoints | scoring. (#28631)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>

tests/models/multimodal/pooling/test_colmodernvbert.py

@@ -10,7 +10,7 @@ embeddings for visual document retrieval.
 import pytest
 import torch
 
-from vllm.entrypoints.pooling.score.utils import compute_maxsim_score
+from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
 
 MODEL_NAME = "ModernVBERT/colmodernvbert-merged"
 COLBERT_DIM = 128

tests/models/multimodal/pooling/test_colpali.py

@@ -18,7 +18,7 @@ from vllm.entrypoints.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.score.utils import ScoreMultiModalParam
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 
 from ....conftest import VllmRunner
 
@@ -114,7 +114,7 @@ def _run_late_interaction_test(
     dtype: str,
 ) -> None:
     """Verify MaxSim scoring matches manual computation."""
-    from vllm.entrypoints.pooling.score.utils import compute_maxsim_score
+    from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
 
     with vllm_runner(
         model,

tests/models/multimodal/pooling/test_colqwen3.py

@@ -18,7 +18,7 @@ from vllm.entrypoints.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.score.utils import ScoreMultiModalParam
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 
 from ....conftest import VllmRunner
 
@@ -125,7 +125,7 @@ def _run_late_interaction_test(
     dtype: str,
 ) -> None:
     """Verify MaxSim scoring matches manual computation."""
-    from vllm.entrypoints.pooling.score.utils import compute_maxsim_score
+    from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
 
     with vllm_runner(
         model,

tests/models/multimodal/pooling/test_colqwen3_5.py

@@ -73,7 +73,7 @@ def _run_late_interaction_test(
     dtype: str,
 ) -> None:
     """Verify MaxSim scoring matches manual computation."""
-    from vllm.entrypoints.pooling.score.utils import compute_maxsim_score
+    from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
 
     with vllm_runner(
         model,

tests/models/multimodal/pooling/test_jinavl_reranker.py

@@ -11,7 +11,7 @@ from vllm.entrypoints.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.score.utils import ScoreMultiModalParam
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 
 from ....conftest import HfRunner, VllmRunner
 

tests/models/multimodal/pooling/test_llama_nemotron_vl.py

@@ -21,7 +21,7 @@ from vllm.entrypoints.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.score.utils import ScoreMultiModalParam
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 from vllm.platforms import current_platform
 
 from ....conftest import IMAGE_ASSETS, HfRunner, PromptImageInput, VllmRunner