[ez] move harmony utils to parser folder (#30117)

Signed-off-by: Andrew Xia <axia@fb.com>
Co-authored-by: Andrew Xia <axia@fb.com>

Author: qandrew

tests/entrypoints/openai/parser/__init__.py

@@ -4,7 +4,7 @@
 from openai.types.responses import ResponseFunctionToolCall, ResponseReasoningItem
 from openai_harmony import Author, Message, Role, TextContent
 
-from vllm.entrypoints.harmony_utils import (
+from vllm.entrypoints.openai.parser.harmony_utils import (
     has_custom_tools,
     parse_input_to_harmony_message,
     parse_output_message,

tests/entrypoints/test_harmony_utils.py renamed to tests/entrypoints/openai/parser/test_harmony_utils.py

@@ -726,7 +726,7 @@ async def test_function_calling_required(client: OpenAI, model_name: str):
 @pytest.mark.asyncio
 @pytest.mark.parametrize("model_name", [MODEL_NAME])
 async def test_system_message_with_tools(client: OpenAI, model_name: str):
-    from vllm.entrypoints.harmony_utils import get_system_message
+    from vllm.entrypoints.openai.parser.harmony_utils import get_system_message
 
     # Test with custom tools enabled - commentary channel should be available
     sys_msg = get_system_message(with_custom_tools=True)

tests/entrypoints/openai/test_response_api_with_harmony.py

@@ -19,7 +19,7 @@
 from vllm.entrypoints.chat_utils import (
     ChatTemplateContentFormatOption,
 )
-from vllm.entrypoints.harmony_utils import (
+from vllm.entrypoints.openai.parser.harmony_utils import (
     get_encoding,
     get_streamable_parser_for_assistant,
     render_for_completion,

vllm/entrypoints/context.py

@@ -21,7 +21,8 @@
     get_history_tool_calls_cnt,
     make_tool_call_id,
 )
-from vllm.entrypoints.harmony_utils import (
+from vllm.entrypoints.logger import RequestLogger
+from vllm.entrypoints.openai.parser.harmony_utils import (
     get_developer_message,
     get_stop_tokens_for_assistant_actions,
     get_streamable_parser_for_assistant,
@@ -30,7 +31,6 @@
     parse_input_to_harmony_message,
     render_for_completion,
 )
-from vllm.entrypoints.logger import RequestLogger
 from vllm.entrypoints.openai.protocol import (
     ChatCompletionLogProb,
     ChatCompletionLogProbs,

vllm/entrypoints/harmony_utils.py renamed to vllm/entrypoints/openai/parser/harmony_utils.py

@@ -64,7 +64,8 @@
     SimpleContext,
     StreamingHarmonyContext,
 )
-from vllm.entrypoints.harmony_utils import (
+from vllm.entrypoints.logger import RequestLogger
+from vllm.entrypoints.openai.parser.harmony_utils import (
     construct_harmony_previous_input_messages,
     get_developer_message,
     get_stop_tokens_for_assistant_actions,
@@ -76,7 +77,6 @@
     parse_response_input,
     render_for_completion,
 )
-from vllm.entrypoints.logger import RequestLogger
 from vllm.entrypoints.openai.protocol import (
     DeltaMessage,
     ErrorResponse,

vllm/entrypoints/openai/serving_chat.py

@@ -4,7 +4,7 @@
 from collections.abc import Sequence
 from typing import TYPE_CHECKING
 
-from vllm.entrypoints.harmony_utils import parse_output_into_messages
+from vllm.entrypoints.openai.parser.harmony_utils import parse_output_into_messages
 from vllm.entrypoints.openai.protocol import (
     ChatCompletionRequest,
     DeltaMessage,

vllm/entrypoints/openai/serving_responses.py

@@ -5,7 +5,7 @@
 
 from transformers import PreTrainedTokenizerBase
 
-from vllm.entrypoints.harmony_utils import parse_chat_output
+from vllm.entrypoints.openai.parser.harmony_utils import parse_chat_output
 from vllm.entrypoints.openai.protocol import ChatCompletionRequest, DeltaMessage
 from vllm.entrypoints.tool_server import ToolServer
 from vllm.logger import init_logger