[Misc] mv chat_utils.py to renderers. (#60522)

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

.github/CODEOWNERS

@@ -43,8 +43,6 @@
 /vllm/entrypoints/pooling @noooop
 /vllm/entrypoints/serve/sagemaker @DarkLight1337
 /vllm/entrypoints/serve @njhill
-/vllm/entrypoints/*.py @njhill
-/vllm/entrypoints/chat_utils.py @DarkLight1337
 /vllm/entrypoints/common @DarkLight1337
 /vllm/entrypoints/llm.py @DarkLight1337
 

docs/models/generative_models.md

@@ -127,7 +127,7 @@ If the model doesn't have a chat template or you want to specify another one,
 you can explicitly pass a chat template:
 
 ```python
-from vllm.entrypoints.chat_utils import load_chat_template
+from vllm.renderers.chat_utils import load_chat_template
 
 # You can find a list of existing chat templates under `examples/`
 custom_template = load_chat_template(chat_template="<path_to_template>")

examples/pooling/embed/vision_embedding_offline.py

@@ -223,7 +223,7 @@ def run_vlm2vec_qwen2vl(seed: int):
     from peft import PeftConfig, PeftModel
     from transformers import AutoModelForImageTextToText, AutoProcessor
 
-    from vllm.entrypoints.chat_utils import load_chat_template
+    from vllm.renderers.chat_utils import load_chat_template
 
     model_id = "TIGER-Lab/VLM2Vec-Qwen2VL-2B"
 

rust/src/chat/src/multimodal.rs

@@ -714,7 +714,7 @@ fn input_audio_data_url(data: &str, format: Option<&str>) -> Result<String> {
 ///
 /// Embedding inputs share their base modality's budget rather than getting one
 /// of their own, matching Python's `modality.replace("_embeds", "")` in
-/// `vllm/entrypoints/chat_utils.py`.
+/// `vllm/renderers/chat_utils.py`.
 fn media_part_limit_modality(part: &MediaContentPart) -> Option<MmModality> {
     match part {
         MediaContentPart::Text { .. } => None,

tests/entrypoints/multimodal/llm/test_mm_cache_stats.py

@@ -8,7 +8,7 @@ import regex as re
 
 from tests.entrypoints.multimodal.conftest import TEST_IMAGE_ASSETS
 from vllm import LLM
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.v1.metrics import loggers as stat_loggers
 from vllm.v1.metrics.reader import Counter, Metric
 

tests/entrypoints/multimodal/llm/test_mm_processor_kwargs.py

@@ -7,9 +7,9 @@ from unittest.mock import Mock
 import pytest
 
 from vllm import LLM, SamplingParams
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.inputs import TextPrompt
 from vllm.outputs import RequestOutput
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 
 def _make_mock_llm() -> LLM:

tests/entrypoints/pooling/embed/test_online.py

@@ -18,7 +18,6 @@ from tests.models.language.pooling.embed_utils import (
 )
 from tests.models.utils import check_embeddings_close
 from tests.utils import RemoteOpenAIServer
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.pooling.embed.protocol import EmbeddingResponse
 from vllm.entrypoints.pooling.pooling.protocol import PoolingResponse
 from vllm.entrypoints.pooling.utils import (
@@ -27,6 +26,7 @@ from vllm.entrypoints.pooling.utils import (
     decode_pooling_output,
 )
 from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.tokenizers import get_tokenizer
 from vllm.utils.serial_utils import EMBED_DTYPES, ENDIANNESS, binary2tensor
 

tests/entrypoints/pooling/reward/test_token_reward_online.py

@@ -11,13 +11,13 @@ import torch
 
 from tests.models.utils import check_embeddings_close
 from tests.utils import RemoteOpenAIServer
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.pooling.pooling.protocol import PoolingResponse
 from vllm.entrypoints.pooling.utils import (
     MetadataItem,
     build_metadata_items,
     decode_pooling_output,
 )
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.tokenizers import get_tokenizer
 from vllm.utils.serial_utils import EMBED_DTYPES, ENDIANNESS, binary2tensor
 

tests/entrypoints/pooling/scoring/util.py

@@ -11,12 +11,12 @@ from safetensors.torch import load_file
 from transformers import AutoModel, AutoTokenizer
 
 from tests.conftest import HfRunner
-from vllm.entrypoints.chat_utils import (
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
+from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
+from vllm.renderers.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
-from vllm.entrypoints.pooling.scoring.utils import compute_maxsim_score
 from vllm.transformers_utils.repo_utils import hf_api
 
 

tests/entrypoints/unit_tests/test_chat_billing_header.py

@@ -6,7 +6,7 @@ from unittest.mock import MagicMock
 
 import pytest
 
-from vllm.entrypoints.chat_utils import _parse_chat_message_content
+from vllm.renderers.chat_utils import _parse_chat_message_content
 
 HEADER = (
     "x-anthropic-billing-header: cc_version=2.1.220.300; "

tests/models/multimodal/generation/test_memory_leak.py

@@ -13,8 +13,8 @@ import torch
 from tests.models.registry import HF_EXAMPLE_MODELS
 from vllm import LLM, SamplingParams
 from vllm.distributed import cleanup_dist_env_and_memory
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.utils.mem_constants import KiB_bytes, MiB_bytes
 from vllm.utils.mem_utils import format_mib
 

tests/models/multimodal/pooling/test_colpali.py

@@ -14,12 +14,12 @@ import pytest
 import torch
 from PIL import Image
 
-from vllm.entrypoints.chat_utils import (
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
+from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
-from vllm.platforms import current_platform
 
 from ....conftest import VllmRunner
 

tests/models/multimodal/pooling/test_colqwen3.py

@@ -14,11 +14,11 @@ import pytest
 import torch
 from PIL import Image
 
-from vllm.entrypoints.chat_utils import (
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
+from vllm.renderers.chat_utils import (
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 
 from ....conftest import VllmRunner
 

tests/models/multimodal/pooling/test_jinavl_reranker.py

@@ -5,12 +5,12 @@ import pytest
 from transformers import AutoModel
 
 from vllm.assets.base import VLLM_S3_BUCKET_URL
-from vllm.entrypoints.chat_utils import (
+from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
+from vllm.renderers.chat_utils import (
     ChatCompletionContentPartImageEmbedsParam,
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,
 )
-from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 
 from ....conftest import HfRunner, VllmRunner
 

tests/models/multimodal/pooling/test_llama_nemotron_vl.py

@@ -18,13 +18,13 @@ import torch
 from PIL import Image
 from transformers import AutoModel, AutoModelForSequenceClassification, AutoProcessor
 
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionContentPartImageParam,
-    ChatCompletionContentPartTextParam,
-)
 from vllm.entrypoints.pooling.scoring.typing import ScoreMultiModalParam
 from vllm.model_executor.layers.fusion.mm_input_norm import FusedMMInputNorm
 from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import (
+    ChatCompletionContentPartImageParam,
+    ChatCompletionContentPartTextParam,
+)
 from vllm.transformers_utils.processors.nemotron_vl import (
     SIGLIP_MEAN,
     SIGLIP_STD,

tests/models/multimodal/processing/test_common.py

@@ -23,13 +23,13 @@ from vllm.distributed.ec_transfer.ec_connector.utils import (
     PlaceholderMetadataResolver,
     collect_ec_item_metadata,
 )
-from vllm.entrypoints.chat_utils import _get_embeds_data
 from vllm.inputs import MultiModalDataDict, MultiModalInput
 from vllm.multimodal import MULTIMODAL_REGISTRY
 from vllm.multimodal.cache import MultiModalProcessorOnlyCache
 from vllm.multimodal.inputs import MultiModalFeatureSpec, batched_tensors_equal
 from vllm.multimodal.processing import BaseMultiModalProcessor, InputProcessingContext
 from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import _get_embeds_data
 from vllm.tokenizers import TokenizerLike, cached_tokenizer_from_config
 from vllm.utils.mistral import is_mistral_tokenizer
 

tests/renderers/test_chat_utils.py

@@ -14,7 +14,14 @@ from vllm.assets.audio import AudioAsset
 from vllm.assets.image import ImageAsset
 from vllm.assets.video import VideoAsset
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.exceptions import VLLMValidationError
+from vllm.inputs import MultiModalDataDict, MultiModalUUIDDict
+from vllm.multimodal.utils import (
+    encode_audio_url,
+    encode_image_url,
+    encode_video_url,
+)
+from vllm.renderers.chat_utils import (
     MEDIA_CONNECTOR_REGISTRY,
     AsyncMultiModalItemTracker,
     ChatCompletionMessageParam,
@@ -26,13 +33,6 @@ from vllm.entrypoints.chat_utils import (
     parse_chat_messages_async,
     validate_chat_template,
 )
-from vllm.exceptions import VLLMValidationError
-from vllm.inputs import MultiModalDataDict, MultiModalUUIDDict
-from vllm.multimodal.utils import (
-    encode_audio_url,
-    encode_image_url,
-    encode_video_url,
-)
 from vllm.utils.serial_utils import tensor2base64
 
 KIMI_K2_5_MODEL_ID = "moonshotai/Kimi-K2.5"

tests/renderers/test_chat_utils_prompt_embeds.py

@@ -16,7 +16,9 @@ import regex as re
 import torch
 from transformers import AutoTokenizer
 
-from vllm.entrypoints.chat_utils import (
+from vllm.exceptions import VLLMValidationError
+from vllm.renderers import TokenizeParams
+from vllm.renderers.chat_utils import (
     _ENABLE_PROMPT_EMBEDS_ERROR,
     _PROMPT_EMBEDS_MISSING_DATA_ERROR,
     _RESERVED_PLACEHOLDER_IN_TEXT_ERROR,
@@ -26,8 +28,6 @@ from vllm.entrypoints.chat_utils import (
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.exceptions import VLLMValidationError
-from vllm.renderers import TokenizeParams
 from vllm.renderers.hf import (
     _PROMPT_EMBEDS_PLACEHOLDER_SPAN_MISMATCH_ERROR,
     _build_mixed_prompt_embeds,

tests/renderers/test_hf.py

@@ -5,9 +5,9 @@ import jinja2
 import pytest
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import load_chat_template
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.exceptions import VLLMValidationError
+from vllm.renderers.chat_utils import load_chat_template
 from vllm.renderers.hf import (
     _consolidate_system_messages,
     _convert_developer_to_system,

tests/renderers/test_warmup.py

@@ -28,9 +28,9 @@ from unittest.mock import MagicMock, patch
 
 import pytest
 
-from vllm.entrypoints.chat_utils import ChatTemplateResolutionError
 from vllm.multimodal.processing import BaseMultiModalProcessor
 from vllm.renderers.base import BaseRenderer
+from vllm.renderers.chat_utils import ChatTemplateResolutionError
 from vllm.renderers.params import ChatParams
 
 

tests/tokenizers_/test_deepseek_v4.py

@@ -8,7 +8,7 @@ from types import SimpleNamespace
 import pytest
 import torch
 
-from vllm.entrypoints.chat_utils import parse_chat_messages
+from vllm.renderers.chat_utils import parse_chat_messages
 from vllm.renderers.registry import RENDERER_REGISTRY
 from vllm.tokenizers.deepseek_v4 import get_deepseek_v4_tokenizer
 from vllm.tokenizers.registry import TokenizerRegistry

tests/tokenizers_/test_mistral.py

@@ -10,9 +10,9 @@ from mistral_common.exceptions import InvalidMessageStructureException
 from mistral_common.guidance.grammar_factory import GrammarFactory
 from mistral_common.tokens.tokenizers.base import SpecialTokenPolicy, SpecialTokens
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.exceptions import VLLMValidationError
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.tokenizers.mistral import (
     MistralTokenizer,
     _validate_apply_chat_template_args,

vllm/entrypoints/anthropic/serving.py

@@ -35,7 +35,6 @@ from vllm.entrypoints.anthropic.protocol import (
     AnthropicToolChangeDefinition,
     AnthropicUsage,
 )
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.generate.base.protocol import (
     JsonSchemaResponseFormat,
     ResponseFormat,
@@ -54,6 +53,7 @@ from vllm.entrypoints.serve.engine.protocol import ErrorResponse, UsageInfo
 from vllm.entrypoints.serve.exception_handling.utils import sanitize_message
 from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.renderers.hf import HfRenderer, resolve_chat_template
 from vllm.renderers.online_renderer import OnlineRenderer
 

vllm/entrypoints/cohere/serving.py

@@ -45,7 +45,6 @@ from fastapi import Request
 from pydantic import BaseModel
 
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.cohere.cohere_chat_message import (
     CitationSource as InternalCitationSource,
 )
@@ -91,6 +90,7 @@ from vllm.entrypoints.serve.engine.protocol import ErrorInfo, ErrorResponse
 from vllm.entrypoints.serve.exception_handling.utils import sanitize_message
 from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.parser.abstract_parser import Parser
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.renderers.cohere import (
     MESSAGES_CITATIONS_KEY,
     POSITION_TO_SOURCE_KEY,
@@ -509,7 +509,7 @@ class CohereServingChatV2(OpenAIServingChat):
         # ``TextToolContent`` and ``DocumentToolContent``. The latter
         # produces ``{"type": "document", "document": {...}}`` blocks,
         # which vLLM's shared ``_parse_chat_message_content_parts`` in
-        # ``vllm/entrypoints/chat_utils.py`` rejects because the OpenAI
+        # ``vllm/renderers/chat_utils.py`` rejects because the OpenAI
         # ``ChatCompletionContentPartParam`` union only knows text /
         # image / audio / etc. Emitting them anyway triggers a Pydantic
         # ``ValidatorIterator`` failure at chat-parsing time, so we

vllm/entrypoints/common/offline.py

@@ -15,15 +15,15 @@ from vllm import (
     SamplingParams,
 )
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-)
 from vllm.exceptions import VLLMValidationError
 from vllm.inputs import EngineInput
 from vllm.logger import init_logger
 from vllm.lora.request import LoRARequest
 from vllm.renderers import BaseRenderer, ChatParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+)
 from vllm.renderers.inputs.preprocess import (
     conversation_to_seq,
     parse_model_prompt,

vllm/entrypoints/generate/api_router.py

@@ -72,7 +72,7 @@ async def init_generate_state(
     default_chat_template_kwargs: dict[str, Any],
 ):
     from vllm.entrypoints.anthropic.serving import AnthropicServingMessages
-    from vllm.entrypoints.chat_utils import load_chat_template
+    from vllm.renderers.chat_utils import load_chat_template
 
     # The Cohere serving handler depends on the optional `cohere` SDK for
     # its wire-format protocol models, and is additionally gated on the

vllm/entrypoints/generate/base/protocol.py

@@ -15,10 +15,10 @@ from pydantic import (
 
 import vllm.envs as envs
 from vllm.config.utils import replace
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel, UsageInfo
 from vllm.exceptions import VLLMValidationError
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.sampling_params import StructuredOutputsParams
 
 logger = init_logger(__name__)

vllm/entrypoints/generate/structured_decisions/protocol.py

@@ -8,10 +8,10 @@ from typing import Any, Literal
 from pydantic import Field, model_validator
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.generate.base.protocol import validate_cache_salt
 from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel
 from vllm.renderers import ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.utils import random_uuid
 
 

vllm/entrypoints/generate/structured_decisions/strategies.py

@@ -16,10 +16,10 @@ from typing import Any
 
 from vllm.config import ModelConfig
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.generate.label_reads import next_token_label_reads
 from vllm.inputs import EngineInput, tokens_input
 from vllm.lora.request import LoRARequest
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.renderers.inputs.preprocess import extract_prompt_components
 from vllm.renderers.online_renderer import OnlineRenderer
 from vllm.sampling_params import SamplingParams

vllm/entrypoints/launchers/api_server/app_state.py

@@ -7,7 +7,6 @@ from typing import cast
 from starlette.datastructures import State
 
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import load_chat_template
 from vllm.entrypoints.launchers.cli_args import resolve_default_chat_template_kwargs
 from vllm.entrypoints.mcp.tool_server import init_tool_server
 from vllm.entrypoints.openai.models.protocol import BaseModelPath
@@ -16,6 +15,7 @@ from vllm.entrypoints.serve.tokenize.serving import ServingTokenization
 from vllm.entrypoints.serve.utils.api_utils import process_lora_modules
 from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.plugins.endpoint_plugins.interface import init_endpoint_plugins_state
+from vllm.renderers.chat_utils import load_chat_template
 from vllm.renderers.online_derenderer import OnlineDerenderer
 from vllm.renderers.online_renderer import OnlineRenderer
 from vllm.tasks import FALLBACK_SUPPORTED_TASKS, POOLING_TASKS, SupportedTask

vllm/entrypoints/launchers/cli_args.py

@@ -17,12 +17,12 @@ from vllm.engine.arg_utils import AsyncEngineArgs, optional_type
 from vllm.entrypoints.anthropic.protocol import (
     AnthropicDisabledThinkingEffortOption,
 )
-from vllm.entrypoints.chat_utils import (
+from vllm.entrypoints.openai.models.protocol import LoRAModulePath
+from vllm.logger import init_logger
+from vllm.renderers.chat_utils import (
     ChatTemplateContentFormatOption,
     validate_chat_template,
 )
-from vllm.entrypoints.openai.models.protocol import LoRAModulePath
-from vllm.logger import init_logger
 from vllm.tool_parsers import ToolParserManager
 from vllm.tool_parsers.tool_strict_level import ToolStrictLevelName
 from vllm.utils.argparse_utils import FlexibleArgumentParser

vllm/entrypoints/launchers/render/app_state.py

@@ -5,7 +5,6 @@ from argparse import Namespace
 from starlette.datastructures import State
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import load_chat_template
 from vllm.entrypoints.launchers.cli_args import resolve_default_chat_template_kwargs
 from vllm.entrypoints.mcp.tool_server import init_tool_server
 from vllm.entrypoints.openai.models.protocol import BaseModelPath
@@ -15,6 +14,7 @@ from vllm.entrypoints.serve.tokenize.serving import ServingTokenization
 from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.plugins.endpoint_plugins.interface import init_endpoint_plugins_state
 from vllm.renderers import renderer_from_config
+from vllm.renderers.chat_utils import load_chat_template
 from vllm.renderers.online_derenderer import OnlineDerenderer
 from vllm.renderers.online_renderer import OnlineRenderer
 

vllm/entrypoints/llm.py

@@ -28,11 +28,6 @@ from vllm.config.model import (
 )
 from vllm.config.quantization import QuantizationConfigArgs
 from vllm.engine.arg_utils import EngineArgs
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-    load_chat_template,
-)
 from vllm.entrypoints.common.offline import _O, _R, OfflineInferenceMixin
 from vllm.entrypoints.generate.beam_search.offline import BeamSearchOfflineMixin
 from vllm.entrypoints.pooling.offline import PoolingOfflineMixin
@@ -44,6 +39,11 @@ from vllm.lora.request import LoRARequest
 from vllm.model_executor.layers.quantization import QuantizationMethods
 from vllm.outputs import PoolingRequestOutput, RequestOutput
 from vllm.platforms import current_platform
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+    load_chat_template,
+)
 from vllm.sampling_params import SamplingParams
 from vllm.tokenizers import TokenizerLike
 from vllm.usage.usage_lib import UsageContext

vllm/entrypoints/openai/chat_completion/batch_serving.py

@@ -8,7 +8,6 @@ from http import HTTPStatus
 
 from fastapi import Request
 
-from vllm.entrypoints.chat_utils import ConversationMessage
 from vllm.entrypoints.generate.base.protocol import RequestResponseMetadata
 from vllm.entrypoints.openai.chat_completion.protocol import (
     BatchChatCompletionRequest,
@@ -24,6 +23,7 @@ from vllm.inputs import EngineInput
 from vllm.logger import init_logger
 from vllm.outputs import RequestOutput
 from vllm.parser.abstract_parser import Parser
+from vllm.renderers.chat_utils import ConversationMessage
 from vllm.tokenizers import TokenizerLike
 from vllm.utils.async_utils import merge_async_iterators
 from vllm.utils.collection_utils import as_list

vllm/entrypoints/openai/chat_completion/protocol.py

@@ -19,11 +19,6 @@ from pydantic import (
 )
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-    has_non_text_content,
-)
 from vllm.entrypoints.generate.base.protocol import (
     AnyResponseFormat,
     DeltaMessage,
@@ -44,6 +39,11 @@ from vllm.exceptions import VLLMValidationError
 from vllm.logger import init_logger
 from vllm.logprobs import Logprob
 from vllm.renderers import ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+    has_non_text_content,
+)
 from vllm.sampling_params import (
     BeamSearchParams,
     RepetitionDetectionParams,

vllm/entrypoints/openai/chat_completion/serving.py

@@ -11,11 +11,6 @@ from typing import Any, Final, cast
 from fastapi import Request
 
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import (
-    ChatTemplateContentFormatOption,
-    ConversationMessage,
-    make_tool_call_id,
-)
 from vllm.entrypoints.generate.base.protocol import (
     DeltaMessage,
     FunctionCall,
@@ -61,6 +56,11 @@ from vllm.logprobs import Logprob
 from vllm.outputs import RequestOutput
 from vllm.parser import ParserManager
 from vllm.parser.abstract_parser import Parser
+from vllm.renderers.chat_utils import (
+    ChatTemplateContentFormatOption,
+    ConversationMessage,
+    make_tool_call_id,
+)
 from vllm.renderers.online_renderer import OnlineRenderer
 from vllm.sampling_params import BeamSearchParams, SamplingParams
 from vllm.tokenizers import TokenizerLike

vllm/entrypoints/openai/responses/context.py

@@ -21,9 +21,6 @@ from openai.types.responses.tool import Mcp
 from openai_harmony import Author, Message, Role, TextContent
 
 from vllm import envs
-from vllm.entrypoints.chat_utils import (
-    ChatTemplateContentFormatOption,
-)
 from vllm.entrypoints.mcp.tool import Tool
 from vllm.entrypoints.mcp.tool_server import ToolServer
 from vllm.entrypoints.openai.parser.harmony_utils import render_for_completion
@@ -41,6 +38,9 @@ from vllm.entrypoints.serve.utils.tool_calls_utils import (
 )
 from vllm.outputs import RequestOutput
 from vllm.parser.abstract_parser import Parser
+from vllm.renderers.chat_utils import (
+    ChatTemplateContentFormatOption,
+)
 from vllm.tokenizers import TokenizerLike
 from vllm.utils import random_uuid
 

vllm/entrypoints/openai/responses/protocol.py

@@ -54,10 +54,6 @@ from pydantic import (
 )
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-)
 from vllm.entrypoints.generate.base.protocol import (
     PerRequestMetrics,
     StopParam,
@@ -68,6 +64,10 @@ from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel
 from vllm.exceptions import VLLMValidationError
 from vllm.logger import init_logger
 from vllm.renderers import ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+)
 from vllm.sampling_params import (
     RequestOutputKind,
     SamplingParams,

vllm/entrypoints/openai/responses/serving.py

@@ -23,9 +23,6 @@ from pydantic import TypeAdapter
 from vllm import envs
 from vllm.config.utils import replace
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import (
-    ChatTemplateContentFormatOption,
-)
 from vllm.entrypoints.generate.base.protocol import (
     DeltaMessage,
     RequestResponseMetadata,
@@ -86,6 +83,9 @@ from vllm.lora.request import LoRARequest
 from vllm.outputs import CompletionOutput
 from vllm.parser import Parser, ParserManager
 from vllm.renderers import TokenizeParams
+from vllm.renderers.chat_utils import (
+    ChatTemplateContentFormatOption,
+)
 from vllm.renderers.online_renderer import (
     OnlineRenderer,
     ResponsesPreviousMessages,

vllm/entrypoints/openai/responses/streaming_events.py

@@ -62,7 +62,6 @@ from openai.types.responses.response_reasoning_item import (
 from openai.types.responses.tool import Tool
 from openai_harmony import Message as HarmonyMessage
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import DeltaMessage, DeltaToolCall
 from vllm.entrypoints.mcp.tool_server import ToolServer
 from vllm.entrypoints.openai.parser.harmony_utils import (
@@ -78,6 +77,7 @@ from vllm.entrypoints.openai.responses.utils import (
 )
 from vllm.outputs import CompletionOutput
 from vllm.parser.harmony import Segment
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.utils import random_uuid
 
 TOOL_NAME_TO_MCP_SERVER_LABEL: Final[dict[str, str]] = {

vllm/entrypoints/openai/responses/utils.py

@@ -30,7 +30,6 @@ from openai.types.responses.response_reasoning_item import (
 from openai.types.responses.tool import Tool
 
 from vllm import envs
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import FunctionCall, FunctionDefinition
 from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionMessageParam,
@@ -39,6 +38,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 from vllm.entrypoints.openai.responses.protocol import ResponseInputOutputItem
 from vllm.exceptions import VLLMValidationError
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.utils import (
     build_responses_tool_call_name_map,
     flat_namespace_tool_name,

vllm/entrypoints/pooling/base/io_processor.py

@@ -10,12 +10,12 @@ from vllm import (
     PoolingRequestOutput,
 )
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
-    ChatTemplateConfig,
-)
 from vllm.exceptions import VLLMValidationError
 from vllm.lora.request import LoRARequest
 from vllm.renderers import BaseRenderer, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatTemplateConfig,
+)
 from vllm.renderers.inputs.preprocess import parse_model_prompt, prompt_to_seq
 from vllm.utils.async_utils import make_async
 from vllm.utils.mistral import is_mistral_tokenizer

vllm/entrypoints/pooling/base/protocol.py

@@ -7,14 +7,14 @@ from typing import Annotated, Any, Literal
 from pydantic import Field, model_validator
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-)
 from vllm.entrypoints.generate.base.protocol import validate_cache_salt
 from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel
 from vllm.exceptions import VLLMValidationError
 from vllm.renderers import ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+)
 from vllm.tasks import check_removed_pooling_task
 from vllm.utils import random_uuid
 from vllm.utils.serial_utils import EmbedDType, EncodingFormat, Endianness

vllm/entrypoints/pooling/base/serving.py

@@ -14,7 +14,6 @@ from fastapi.responses import Response
 from vllm import PoolingRequestOutput, envs
 from vllm.config import VllmConfig
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import ChatTemplateConfig
 from vllm.entrypoints.generate.base.protocol import validate_request_mm_kwargs
 from vllm.entrypoints.openai.models.serving import OpenAIServingModels
 from vllm.entrypoints.serve.engine.serving import BaseServing
@@ -23,6 +22,7 @@ from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.exceptions import GenerationError, RetryableRequestError
 from vllm.lora.request import LoRARequest
 from vllm.renderers.base import BaseRenderer
+from vllm.renderers.chat_utils import ChatTemplateConfig
 from vllm.utils.async_utils import make_async, merge_async_iterators
 
 from ...serve.engine.protocol import ErrorResponse

vllm/entrypoints/pooling/embed/io_processor.py

@@ -12,15 +12,15 @@ from openai.types.chat import (
 from openai.types.chat.chat_completion_content_part_image_param import ImageURL
 
 from vllm import PoolingParams
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionContentPartParam,
-    ChatCompletionMessageParam,
-    CustomChatCompletionMessageParam,
-)
 from vllm.inputs import tokens_input
 from vllm.logger import init_logger
 from vllm.outputs import PoolingOutput, PoolingRequestOutput
 from vllm.renderers import merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionContentPartParam,
+    ChatCompletionMessageParam,
+    CustomChatCompletionMessageParam,
+)
 from vllm.renderers.hf import resolve_chat_template
 from vllm.utils.collection_utils import chunk_list
 from vllm.utils.mistral import is_mistral_tokenizer

vllm/entrypoints/pooling/embed/protocol.py

@@ -17,8 +17,8 @@ import pybase64 as base64
 from pydantic import BaseModel, BeforeValidator, Field, model_validator
 
 from vllm import PoolingParams
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel, UsageInfo
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.utils import random_uuid
 
 from ..base.protocol import (

vllm/entrypoints/pooling/factories.py

@@ -6,10 +6,10 @@ from typing import TYPE_CHECKING
 from fastapi import FastAPI
 
 from vllm.config import ModelConfig, VllmConfig
-from vllm.entrypoints.chat_utils import ChatTemplateConfig
 from vllm.logger import init_logger
 from vllm.plugins.io_processors import has_io_processor
 from vllm.renderers import BaseRenderer
+from vllm.renderers.chat_utils import ChatTemplateConfig
 from vllm.tasks import POOLING_TASKS, SCORE_TYPE_MAP, SupportedTask
 
 from .base.io_processor import PoolingIOProcessor
@@ -151,7 +151,7 @@ def init_pooling_state(
     if model_config is None:
         return
 
-    from vllm.entrypoints.chat_utils import load_chat_template
+    from vllm.renderers.chat_utils import load_chat_template
     from vllm.tasks import POOLING_TASKS
 
     from .classify.serving import ServingClassification

vllm/entrypoints/pooling/offline.py

@@ -6,7 +6,6 @@ from typing import Any
 
 from tqdm.auto import tqdm
 
-from vllm.entrypoints.chat_utils import ChatTemplateConfig
 from vllm.entrypoints.common.offline import OfflineInferenceMixin
 from vllm.inputs import DataPrompt, PromptType
 from vllm.logger import init_logger
@@ -18,6 +17,7 @@ from vllm.outputs import (
     ScoringRequestOutput,
 )
 from vllm.pooling_params import PoolingParams
+from vllm.renderers.chat_utils import ChatTemplateConfig
 from vllm.tasks import SCORE_TYPE_MAP, PoolingTask, SupportedTask
 
 from .base.io_processor import PoolingIOProcessor

vllm/entrypoints/pooling/scoring/io_processor.py

@@ -9,6 +9,7 @@ import torch.nn.functional as F
 from vllm import PoolingParams, PoolingRequestOutput, TokensPrompt
 from vllm.logger import init_logger
 from vllm.renderers import TokenizeParams
+from vllm.renderers.chat_utils import ChatTemplateResolutionError
 from vllm.renderers.hf import safe_apply_chat_template
 from vllm.renderers.inputs.preprocess import (
     extract_target_prompt,
@@ -18,7 +19,6 @@ from vllm.renderers.inputs.preprocess import (
 from vllm.tasks import PoolingTask
 from vllm.utils.mistral import is_mistral_tokenizer
 
-from ...chat_utils import ChatTemplateResolutionError
 from ..base.io_processor import PoolingIOProcessor
 from ..pooling.protocol import PoolingCompletionRequest
 from ..typing import (

vllm/entrypoints/pooling/scoring/typing.py

@@ -5,7 +5,7 @@ from typing import Required, TypeAlias
 
 from typing_extensions import TypedDict
 
-from vllm.entrypoints.chat_utils import (
+from vllm.renderers.chat_utils import (
     ChatCompletionContentPartImageEmbedsParam,
     ChatCompletionContentPartImageParam,
     ChatCompletionContentPartTextParam,

vllm/entrypoints/pooling/scoring/utils.py

@@ -7,7 +7,9 @@ import torch
 
 from vllm import PromptType, TextPrompt
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.exceptions import VLLMValidationError
+from vllm.inputs import MultiModalDataDict, MultiModalUUIDDict
+from vllm.renderers.chat_utils import (
     BaseMultiModalItemTracker,
     ChatCompletionContentPartParam,
     ChatCompletionContentPartTextParam,
@@ -15,8 +17,6 @@ from vllm.entrypoints.chat_utils import (
     MultiModalItemTracker,
     _parse_chat_message_content_parts,
 )
-from vllm.exceptions import VLLMValidationError
-from vllm.inputs import MultiModalDataDict, MultiModalUUIDDict
 
 from .typing import (
     ScoreContentPartParam,

vllm/entrypoints/pooling/typing.py

@@ -9,10 +9,10 @@ from fastapi import Request
 from pydantic import ConfigDict
 
 from vllm import PoolingParams, PoolingRequestOutput, PromptType
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.inputs import DataPrompt, EngineInput
 from vllm.lora.request import LoRARequest
 from vllm.renderers import ChatParams, TokenizeParams
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.renderers.inputs import DictPrompt
 
 from ...tasks import PoolingTask

vllm/entrypoints/scale_out/token_in_token_out/serving.py

@@ -13,7 +13,6 @@ import msgspec
 from fastapi import Request
 
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.chat_utils import AsyncMultiModalItemTracker
 from vllm.entrypoints.generate.base.protocol import (
     PerRequestMetrics,
     RequestResponseMetadata,
@@ -47,6 +46,7 @@ from vllm.multimodal.inputs import (
     PlaceholderRange,
 )
 from vllm.outputs import RequestOutput
+from vllm.renderers.chat_utils import AsyncMultiModalItemTracker
 from vllm.renderers.online_renderer import OnlineRenderer
 from vllm.sampling_params import RequestOutputKind, SamplingParams
 from vllm.tokenizers import TokenizerLike

vllm/entrypoints/serve/engine/typing.py

@@ -4,7 +4,6 @@
 from typing import Protocol, TypeAlias
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.openai.chat_completion.protocol import (
     BatchChatCompletionRequest,
     ChatCompletionRequest,
@@ -35,6 +34,7 @@ from vllm.entrypoints.speech_to_text.transcription.protocol import (
 )
 from vllm.entrypoints.speech_to_text.translation.protocol import TranslationRequest
 from vllm.renderers import ChatParams, TokenizeParams
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 
 
 class RendererRequest(Protocol):

vllm/entrypoints/serve/tokenize/protocol.py

@@ -7,14 +7,14 @@ from typing import Annotated, Any, TypeAlias
 from pydantic import ConfigDict, Field, model_validator
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-)
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionToolsParam
 from vllm.entrypoints.serve.engine.protocol import OpenAIBaseModel
 from vllm.exceptions import VLLMValidationError
 from vllm.renderers import ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+)
 
 
 class TokenizeCompletionRequest(OpenAIBaseModel):

vllm/entrypoints/serve/tokenize/serving.py

@@ -5,7 +5,6 @@ from typing import Any, Final
 
 from fastapi import Request
 
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.openai.models.serving import (
     OpenAIModelRegistry,
     OpenAIServingModels,
@@ -23,6 +22,7 @@ from vllm.entrypoints.serve.tokenize.protocol import (
 from vllm.entrypoints.serve.utils.request_logger import RequestLogger
 from vllm.inputs import TokensPrompt, tokens_input
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.renderers.online_renderer import OnlineRenderer
 from vllm.tokenizers import TokenizerLike
 

vllm/parser/abstract_parser.py

@@ -21,10 +21,6 @@ from xgrammar.structural_tag import (
     TriggeredTagsFormat,
 )
 
-from vllm.entrypoints.chat_utils import (
-    get_tool_call_id_type,
-    make_tool_call_id,
-)
 from vllm.entrypoints.generate.base.protocol import (
     DeltaMessage,
     ExtractedToolCallInformation,
@@ -42,6 +38,10 @@ from vllm.parser.engine.adapters import ParserEngineToolAdapter
 from vllm.parser.metrics import record_tool_parser_invocation
 from vllm.parser.utils import count_history_tool_calls
 from vllm.reasoning.abs_reasoning_parsers import ReasoningParser
+from vllm.renderers.chat_utils import (
+    get_tool_call_id_type,
+    make_tool_call_id,
+)
 from vllm.sampling_params import (
     SamplingParams,
     StructuredOutputsParams,

vllm/parser/engine/parser_engine.py

@@ -13,7 +13,6 @@ from typing import TYPE_CHECKING, NamedTuple
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import get_tool_call_id_type, make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -27,6 +26,7 @@ from vllm.parser.abstract_parser import Parser, StreamState
 from vllm.parser.engine.events import EventType, SemanticEvent
 from vllm.parser.engine.parser_engine_config import ParserEngineConfig, ParserState
 from vllm.parser.engine.streaming_parser_engine import StreamingParserEngine
+from vllm.renderers.chat_utils import get_tool_call_id_type, make_tool_call_id
 from vllm.tool_parsers.utils import (
     coerce_to_schema_type,
     extract_types_from_schema,

vllm/parser/harmony.py

@@ -23,7 +23,6 @@ from xgrammar.structural_tag import (
     TagFormat,
 )
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -43,6 +42,7 @@ from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.parser.abstract_parser import DelegatingParser, structured_outputs_to_format
 from vllm.reasoning.gptoss_reasoning_parser import GptOssReasoningParser
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.gptoss_tool_parser import GptOssToolParser
 from vllm.tool_parsers.structural_tag_registry import (
     SimplifiedToolChoice,

vllm/parser/response_template.py

@@ -11,7 +11,6 @@ from typing import TYPE_CHECKING, Any, cast
 
 from openai.types.responses import ToolChoiceFunction
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -30,6 +29,7 @@ from vllm.parser.chat_parsing.response_templates import (
     load_response_template,
 )
 from vllm.reasoning.abs_reasoning_parsers import ReasoningParser
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.abstract_tool_parser import ToolParser
 from vllm.tool_parsers.utils import iter_response_function_tool_dicts
 

vllm/parser/utils.py

@@ -5,12 +5,12 @@ from collections.abc import Iterable, Sequence
 
 from openai.types.responses import ResponseFunctionToolCall
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.entrypoints.openai.responses.protocol import (
     ResponseInputOutputItem,
     ResponsesRequest,
 )
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 
 def count_tool_calls(tool_calls: object) -> int:

vllm/renderers/base.py

@@ -66,7 +66,7 @@ from .params import ChatParams, TokenizeParams
 
 if TYPE_CHECKING:
     from vllm.config import VllmConfig
-    from vllm.entrypoints.chat_utils import (
+    from vllm.renderers.chat_utils import (
         ChatCompletionMessageParam,
         ConversationMessage,
     )
@@ -362,7 +362,7 @@ class BaseRenderer(ABC, Generic[_T]):
         no-op if the background warmup — or a prior reset_mm_cache that joined
         it — already ran it.
         """
-        from vllm.entrypoints.chat_utils import ChatTemplateResolutionError
+        from vllm.renderers.chat_utils import ChatTemplateResolutionError
 
         # prevent MM processor hangs
         with set_default_torch_num_threads(1):

vllm/renderers/cohere.py

@@ -94,13 +94,13 @@ from enum import Enum
 from typing import Any
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.logger import init_logger
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.logger import init_logger
 from vllm.tokenizers.hf import HfTokenizer
 from vllm.utils.async_utils import make_async
 
@@ -165,7 +165,7 @@ POSITION_TO_SOURCE_KEY = "_position_to_source"
 # melody-shape content list the renderer emits.
 #
 # Motivation: OpenAI's ``ChatCompletionContentPartParam`` union (in
-# ``vllm/entrypoints/chat_utils.py``) only allows ``text`` / ``image_url``
+# ``vllm/renderers/chat_utils.py``) only allows ``text`` / ``image_url``
 # / ``audio`` / ``video`` / ``file`` content parts on tool messages, so
 # ``{"type": "document", "document": {...}}`` blocks -- valid in Cohere
 # v2 -- get rejected by vLLM's shared chat validator before they can

vllm/renderers/deepseek_v32.py

@@ -2,7 +2,7 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,

vllm/renderers/deepseek_v4.py

@@ -2,7 +2,7 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,

vllm/renderers/hf.py

@@ -22,13 +22,6 @@ import torch
 from typing_extensions import override
 
 from vllm import envs
-from vllm.entrypoints.chat_utils import (
-    PROMPT_EMBEDS_PLACEHOLDER_TOKEN,
-    ChatTemplateResolutionError,
-    load_chat_template,
-    parse_chat_messages,
-    parse_chat_messages_async,
-)
 from vllm.exceptions import VLLMValidationError
 from vllm.inputs import EmbedsPrompt
 from vllm.inputs.engine import MultiModalInput
@@ -46,6 +39,13 @@ from vllm.multimodal.processing.processor import (
     apply_token_matches,
     find_mm_placeholders,
 )
+from vllm.renderers.chat_utils import (
+    PROMPT_EMBEDS_PLACEHOLDER_TOKEN,
+    ChatTemplateResolutionError,
+    load_chat_template,
+    parse_chat_messages,
+    parse_chat_messages_async,
+)
 from vllm.tokenizers.hf import HfTokenizer, maybe_make_thread_pool
 from vllm.transformers_utils.chat_templates import get_chat_template_fallback_path
 from vllm.transformers_utils.processor import cached_get_processor
@@ -59,18 +59,18 @@ if TYPE_CHECKING:
     from collections.abc import Set
 
     from vllm.config import ModelConfig, VllmConfig
-    from vllm.entrypoints.chat_utils import (
-        ChatCompletionMessageParam,
-        ChatTemplateContentFormat,
-        ChatTemplateContentFormatOption,
-        ConversationMessage,
-    )
     from vllm.inputs import MultiModalDataDict, MultiModalUUIDDict, TokensPrompt
     from vllm.inputs.engine import TokensInput
     from vllm.multimodal.processing.processor import (
         MultiModalPromptUpdates,
         ResolvedPromptUpdate,
     )
+    from vllm.renderers.chat_utils import (
+        ChatCompletionMessageParam,
+        ChatTemplateContentFormat,
+        ChatTemplateContentFormatOption,
+        ConversationMessage,
+    )
 
     from .inputs import DictPrompt
     from .params import ChatParams

vllm/renderers/inkling.py

@@ -14,13 +14,13 @@ input-processing library (mistral-common style) later.
 """
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.logger import init_logger
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.logger import init_logger
 from vllm.tokenizers.hf import HfTokenizer
 from vllm.utils.async_utils import make_async
 

vllm/renderers/inputs/preprocess.py

@@ -21,7 +21,7 @@ if TYPE_CHECKING:
     import torch
 
     from vllm.config import ModelConfig
-    from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+    from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 
 @overload

vllm/renderers/kimi_k3.py

@@ -3,14 +3,14 @@
 from typing import Any, cast
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.exceptions import VLLMValidationError
+from vllm.multimodal.media.connector import merge_media_io_kwargs
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.exceptions import VLLMValidationError
-from vllm.multimodal.media.connector import merge_media_io_kwargs
 from vllm.tokenizers.hf import HfTokenizer
 from vllm.utils.async_utils import make_async
 

vllm/renderers/mistral.py

@@ -2,13 +2,13 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 from vllm.config import VllmConfig
-from vllm.entrypoints.chat_utils import (
+from vllm.logger import init_logger
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.logger import init_logger
 from vllm.tokenizers.mistral import MistralTokenizer
 from vllm.utils.async_utils import make_async
 

vllm/renderers/online_derenderer.py

@@ -4,7 +4,6 @@ from collections.abc import Sequence
 from typing import Any
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
 from vllm.entrypoints.generate.base.protocol import (
     DeltaMessage,
     ToolCall,
@@ -42,6 +41,7 @@ from vllm.entrypoints.serve.utils.tool_calls_utils import (
 from vllm.logger import init_logger
 from vllm.parser import Parser, ParserManager
 from vllm.renderers import BaseRenderer
+from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 from vllm.tokenizers import TokenizerLike
 from vllm.tokenizers.detokenizer_utils import (
     convert_prompt_ids_to_tokens,

vllm/renderers/online_renderer.py

@@ -12,12 +12,6 @@ from openai_harmony import Message as OpenAIMessage
 from openai_harmony import ToolNamespaceConfig
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.chat_utils import (
-    ChatCompletionMessageParam,
-    ChatTemplateContentFormatOption,
-    ConversationMessage,
-    has_non_text_content,
-)
 from vllm.entrypoints.generate.base.protocol import validate_request_mm_kwargs
 from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionNamedToolChoiceParam,
@@ -59,6 +53,12 @@ from vllm.inputs import (
 from vllm.logger import init_logger
 from vllm.parser import Parser, ParserManager
 from vllm.renderers import BaseRenderer, ChatParams, TokenizeParams, merge_kwargs
+from vllm.renderers.chat_utils import (
+    ChatCompletionMessageParam,
+    ChatTemplateContentFormatOption,
+    ConversationMessage,
+    has_non_text_content,
+)
 from vllm.renderers.inputs.preprocess import (
     parse_model_prompt,
     prompt_to_seq,

vllm/renderers/params.py

@@ -13,7 +13,7 @@ from vllm.utils.import_utils import LazyLoader
 if TYPE_CHECKING:
     import torch
 
-    from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption
+    from vllm.renderers.chat_utils import ChatTemplateContentFormatOption
 else:
     torch = LazyLoader("torch", globals(), "torch")
 

vllm/renderers/terratorch.py

@@ -1,13 +1,13 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
-from vllm.entrypoints.chat_utils import (
+from vllm.logger import init_logger
+from vllm.renderers.chat_utils import (
     ChatCompletionMessageParam,
     ConversationMessage,
     parse_chat_messages,
     parse_chat_messages_async,
 )
-from vllm.logger import init_logger
 
 from .base import BaseRenderer
 from .inputs import DictPrompt

vllm/tokenizers/deepseek_v32.py

@@ -5,7 +5,7 @@ from typing import Any
 
 from transformers import TokenizersBackend
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 from .deepseek_v32_encoding import encode_messages
 from .hf import HfTokenizer, get_cached_tokenizer

vllm/tokenizers/deepseek_v4.py

@@ -5,7 +5,7 @@ from typing import Any
 
 from transformers import TokenizersBackend
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 from .deepseek_v4_encoding import encode_messages
 from .hf import HfTokenizer, get_cached_tokenizer

vllm/tokenizers/deepseek_v41.py

@@ -6,7 +6,7 @@ from typing import Any
 
 from transformers import TokenizersBackend
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 from .deepseek_v41_encoding import (
     IMAGE_PLACEHOLDER,

vllm/tokenizers/kimi_audio.py

@@ -13,8 +13,8 @@ import tiktoken
 from transformers import AddedToken, BatchEncoding
 from transformers.utils import chat_template_utils as hf_chat_utils
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.tokenizers.protocol import TokenizerLike
 from vllm.transformers_utils.repo_utils import hf_api
 

vllm/tokenizers/mistral.py

@@ -32,10 +32,10 @@ from mistral_common.tokens.tokenizers.tekken import Tekkenizer
 from pydantic import ValidationError
 from transformers.tokenization_mistral_common import MistralCommonBackend
 
-from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.exceptions import VLLMValidationError
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import ChatCompletionMessageParam
 from vllm.tokenizers.protocol import TokenizerLike
 
 if TYPE_CHECKING:

vllm/tokenizers/protocol.py

@@ -7,7 +7,7 @@ from typing import TYPE_CHECKING, Any, Protocol, overload
 if TYPE_CHECKING:
     from transformers import BatchEncoding
 
-    from vllm.entrypoints.chat_utils import ChatCompletionMessageParam
+    from vllm.renderers.chat_utils import ChatCompletionMessageParam
 
 
 class TokenizerLike(Protocol):

vllm/tool_parsers/apertus_tool_parser.py

@@ -14,7 +14,6 @@ from collections.abc import Sequence
 import regex as re
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -26,6 +25,7 @@ from vllm.entrypoints.generate.base.protocol import (
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/deepseekv31_tool_parser.py

@@ -5,7 +5,6 @@ from collections.abc import Sequence
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -18,6 +17,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
 

vllm/tool_parsers/deepseekv3_tool_parser.py

@@ -5,7 +5,6 @@ from collections.abc import Sequence
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -18,6 +17,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/dots_tool_parser.py

@@ -7,7 +7,6 @@ from typing import Any
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -21,6 +20,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionToolsParam,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
 from vllm.tool_parsers.utils import is_complete_json, partial_tag_overlap

vllm/tool_parsers/functiongemma_tool_parser.py

@@ -6,7 +6,6 @@ from collections.abc import Sequence
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -20,6 +19,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
 

vllm/tool_parsers/gigachat3_tool_parser.py

@@ -6,7 +6,6 @@ from collections.abc import Sequence
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -20,6 +19,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
 

vllm/tool_parsers/granite4_tool_parser.py

@@ -7,7 +7,6 @@ from typing import Any, Protocol, TypeVar
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -21,6 +20,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/granite_20b_fc_tool_parser.py

@@ -9,7 +9,6 @@ import partial_json_parser
 import regex as re
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -22,6 +21,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/hermes_tool_parser.py

@@ -6,7 +6,6 @@ from collections.abc import Sequence
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -20,6 +19,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/hy_v3_tool_parser.py

@@ -7,7 +7,6 @@ from typing import Any
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -21,6 +20,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionToolsParam,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/hy_v4_tool_parser.py

@@ -11,7 +11,6 @@ from typing import TYPE_CHECKING, Any, TypedDict
 import regex as re
 
 import vllm.envs as envs
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -23,6 +22,7 @@ from vllm.entrypoints.generate.base.protocol import (
 from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import ToolParser
 from vllm.tool_parsers.tool_strict_level import ToolStrictLevel

vllm/tool_parsers/internlm2_tool_parser.py

@@ -7,7 +7,6 @@ from collections.abc import Sequence
 import partial_json_parser
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -21,6 +20,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/jamba_tool_parser.py

@@ -8,7 +8,6 @@ import partial_json_parser
 import regex as re
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -22,6 +21,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
 from vllm.tool_parsers.utils import extract_intermediate_diff, is_complete_json

vllm/tool_parsers/llama_tool_parser.py

@@ -10,7 +10,6 @@ from partial_json_parser.core.options import Allow
 from transformers import PreTrainedTokenizerBase
 
 import vllm.envs as envs
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -23,6 +22,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,
     ToolParser,

vllm/tool_parsers/minicpm5xml_tool_parser.py

@@ -7,7 +7,6 @@ from typing import Any
 
 import regex as re
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -22,6 +21,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/muse_glimmer_tool_parser.py

@@ -41,7 +41,6 @@ import regex as re
 from openai.types.responses import ToolChoiceFunction
 from transformers import PreTrainedTokenizerBase
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -56,6 +55,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
 )
 from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,
     ToolParser,

vllm/tool_parsers/phi4mini_tool_parser.py

@@ -8,7 +8,6 @@ from typing import Any
 import regex as re
 from transformers import PreTrainedTokenizerBase
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaMessage,
     ExtractedToolCallInformation,
@@ -19,6 +18,7 @@ from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,
     ToolParser,

vllm/tool_parsers/poolside_v1_tool_parser.py

@@ -19,7 +19,6 @@ import regex as re
 from openai.types.responses import ToolChoiceFunction
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
@@ -36,6 +35,7 @@ from vllm.entrypoints.openai.responses.protocol import (
     ResponsesRequest,
 )
 from vllm.logger import init_logger
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
     Tool,

vllm/tool_parsers/streaming.py

@@ -5,12 +5,12 @@ import json
 from dataclasses import dataclass
 from typing import TYPE_CHECKING
 
-from vllm.entrypoints.chat_utils import make_tool_call_id
 from vllm.entrypoints.generate.base.protocol import (
     DeltaFunctionCall,
     DeltaMessage,
     DeltaToolCall,
 )
+from vllm.renderers.chat_utils import make_tool_call_id
 from vllm.utils.mistral import is_mistral_tokenizer
 
 if TYPE_CHECKING: