[Refactor] [2/N] Move tool parsers into the vLLM main directory (#30675)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
This commit is contained in:
@@ -19,9 +19,9 @@ from vllm.entrypoints.openai.protocol import (
|
||||
)
|
||||
from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
|
||||
from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParserManager
|
||||
from vllm.outputs import CompletionOutput, RequestOutput
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers import ToolParserManager
|
||||
from vllm.v1.engine.async_llm import AsyncLLM
|
||||
|
||||
from ...utils import RemoteOpenAIServer
|
||||
|
||||
@@ -10,8 +10,8 @@ from tests.entrypoints.openai.tool_parsers.utils import (
|
||||
run_tool_extraction_streaming,
|
||||
)
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers import ToolParser, ToolParserManager
|
||||
|
||||
SIMPLE_ARGS_DICT = {
|
||||
"action": "create",
|
||||
|
||||
@@ -6,8 +6,8 @@ import json
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.protocol import ChatCompletionRequest
|
||||
from vllm.entrypoints.openai.tool_parsers.hermes_tool_parser import Hermes2ProToolParser
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers.hermes_tool_parser import Hermes2ProToolParser
|
||||
|
||||
from ....utils import RemoteOpenAIServer
|
||||
|
||||
|
||||
@@ -12,7 +12,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
|
||||
run_tool_extraction_streaming,
|
||||
)
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
|
||||
from vllm.tool_parsers import ToolParser, ToolParserManager
|
||||
|
||||
|
||||
def make_tool_call(name, arguments):
|
||||
|
||||
@@ -6,8 +6,8 @@ from unittest.mock import MagicMock, patch
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.protocol import ExtractedToolCallInformation
|
||||
from vllm.entrypoints.openai.tool_parsers.llama_tool_parser import Llama3JsonToolParser
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers.llama_tool_parser import Llama3JsonToolParser
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
|
||||
@@ -10,8 +10,8 @@ from tests.entrypoints.openai.tool_parsers.utils import (
|
||||
run_tool_extraction_streaming,
|
||||
)
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers import ToolParser, ToolParserManager
|
||||
|
||||
# Test cases similar to pythonic parser but with Llama4 specific format
|
||||
SIMPLE_FUNCTION_OUTPUT = "[get_weather(city='LA', metric='C')]"
|
||||
|
||||
@@ -10,8 +10,8 @@ from tests.entrypoints.openai.tool_parsers.utils import (
|
||||
run_tool_extraction_streaming,
|
||||
)
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers import ToolParser, ToolParserManager
|
||||
|
||||
# https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/text_prompt_format.md#model-response-format-1
|
||||
SIMPLE_FUNCTION_OUTPUT = "get_weather(city='San Francisco', metric='celsius')"
|
||||
|
||||
@@ -10,8 +10,8 @@ from tests.entrypoints.openai.tool_parsers.utils import (
|
||||
run_tool_extraction_streaming,
|
||||
)
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers import ToolParser, ToolParserManager
|
||||
|
||||
# https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/text_prompt_format.md#model-response-format-1
|
||||
SIMPLE_FUNCTION_OUTPUT = "get_weather(city='San Francisco', metric='celsius')"
|
||||
|
||||
@@ -10,8 +10,8 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers import ToolParser
|
||||
from vllm.tokenizers import TokenizerLike
|
||||
from vllm.tool_parsers import ToolParser
|
||||
|
||||
|
||||
class StreamingToolReconstructor:
|
||||
|
||||
@@ -5,12 +5,12 @@ import json
|
||||
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.tool_parsers.mistral_tool_parser import (
|
||||
from vllm.sampling_params import SamplingParams
|
||||
from vllm.tokenizers.mistral import MistralTokenizer
|
||||
from vllm.tool_parsers.mistral_tool_parser import (
|
||||
MistralToolCall,
|
||||
MistralToolParser,
|
||||
)
|
||||
from vllm.sampling_params import SamplingParams
|
||||
from vllm.tokenizers.mistral import MistralTokenizer
|
||||
|
||||
from ...utils import check_logprobs_close
|
||||
|
||||
|
||||
@@ -3,10 +3,10 @@
|
||||
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.tool_parsers.deepseekv31_tool_parser import (
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers.deepseekv31_tool_parser import (
|
||||
DeepSeekV31ToolParser,
|
||||
)
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
|
||||
MODEL = "deepseek-ai/DeepSeek-V3.1"
|
||||
|
||||
|
||||
@@ -13,9 +13,9 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.ernie45_tool_parser import Ernie45ToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tool_parsers.ernie45_tool_parser import Ernie45ToolParser
|
||||
|
||||
# Use a common model that is likely to be available
|
||||
MODEL = "baidu/ERNIE-4.5-21B-A3B-Thinking"
|
||||
|
||||
@@ -7,10 +7,10 @@ import json
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers.glm4_moe_tool_parser import (
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers.glm4_moe_tool_parser import (
|
||||
Glm4MoeModelToolParser,
|
||||
)
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -9,9 +9,9 @@ import pytest
|
||||
from partial_json_parser.core.options import Allow
|
||||
|
||||
from vllm.entrypoints.openai.protocol import DeltaMessage, FunctionCall, ToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers.jamba_tool_parser import JambaToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tool_parsers.jamba_tool_parser import JambaToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -7,8 +7,8 @@ import json
|
||||
import pytest
|
||||
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers.kimi_k2_tool_parser import KimiK2ToolParser
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers.kimi_k2_tool_parser import KimiK2ToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -12,8 +12,8 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.minimax_tool_parser import MinimaxToolParser
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers.minimax_tool_parser import MinimaxToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -12,10 +12,10 @@ from mistral_common.protocol.instruct.tool_calls import FunctionCall, ToolCall
|
||||
from partial_json_parser.core.options import Allow
|
||||
|
||||
from vllm.entrypoints.openai.protocol import DeltaMessage, DeltaToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers.mistral_tool_parser import MistralToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tokenizers.mistral import MistralTokenizer
|
||||
from vllm.tool_parsers.mistral_tool_parser import MistralToolParser
|
||||
|
||||
|
||||
@pytest.fixture(scope="module")
|
||||
|
||||
@@ -15,8 +15,8 @@ from openai_harmony import (
|
||||
)
|
||||
|
||||
from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
|
||||
from vllm.entrypoints.openai.tool_parsers.openai_tool_parser import OpenAIToolParser
|
||||
from vllm.tokenizers import get_tokenizer
|
||||
from vllm.tool_parsers.openai_tool_parser import OpenAIToolParser
|
||||
|
||||
MODEL = "gpt2"
|
||||
|
||||
|
||||
@@ -13,12 +13,12 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.qwen3coder_tool_parser import (
|
||||
Qwen3CoderToolParser,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.qwen3xml_tool_parser import Qwen3XMLToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tool_parsers.qwen3coder_tool_parser import (
|
||||
Qwen3CoderToolParser,
|
||||
)
|
||||
from vllm.tool_parsers.qwen3xml_tool_parser import Qwen3XMLToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -14,9 +14,9 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.seed_oss_tool_parser import SeedOssToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tool_parsers.seed_oss_tool_parser import SeedOssToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -12,7 +12,7 @@ from vllm.entrypoints.openai.protocol import (
|
||||
ChatCompletionToolsParam,
|
||||
)
|
||||
from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
|
||||
from vllm.entrypoints.openai.tool_parsers.utils import get_json_schema_from_tools
|
||||
from vllm.tool_parsers.utils import get_json_schema_from_tools
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
@@ -12,9 +12,9 @@ from vllm.entrypoints.openai.protocol import (
|
||||
FunctionCall,
|
||||
ToolCall,
|
||||
)
|
||||
from vllm.entrypoints.openai.tool_parsers.xlam_tool_parser import xLAMToolParser
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
|
||||
from vllm.tool_parsers.xlam_tool_parser import xLAMToolParser
|
||||
|
||||
pytestmark = pytest.mark.cpu_test
|
||||
|
||||
|
||||
Reference in New Issue
Block a user