[Ascend]support xgrammar backend for ascend npu

[ash-sigh]

Motivation

support xgrammar backend for ascend npu

Modifications

support vocab mask for ascend npu

Accuracy Tests

test1:

export PYTHONPATH=/home/w30027501/sglang_code/sglang/python:$PYTHONPATH
export HCCL_OP_EXPANSION_MODE="AIV"
export OMP_PROC_BIND=false
export STREAMS_PER_DEVICE=32

python -m sglang.launch_server --model-path /data/weights/Qwen3-VL-4B-Instruct --host 127.0.0.1 --port 8022 --device npu --attention-backend ascend --tp 1 --grammar-backend xgrammar --base-gpu-id 7 --mm-attention-backend ascend_attn --trust-remote-code --enable-multimodal --cuda-graph-bs 8

# benchmark
python3 -m sglang.test.few_shot_gsm8k --num-questions 200 --port 8022

result:
Accuracy: 0.930
Invalid: 0.000
Latency: 364.722 s
Output throughput: 72.389 token/s

test2:

import openai
import os

from sglang.test.doc_patch import launch_server_cmd
from sglang.utils import wait_for_server, print_highlight, terminate_process
import json

client = openai.Client(
    base_url=f"http://127.0.0.1:8022/v1",
    api_key="None"
)

json_schema = json.dumps(
    {
        "type": "object",
        "properties": {
            "name": {"type": "string", "pattern": "^[\\w]+$"},
            "population": {"type": "integer"},
        },
        "required": ["name", "population"],
    }
)

response = client.chat.completions.create(
    model="/data/weights/Qwen3-VL-4B-Instruct",
    messages=[
        {
            "role": "user",
            "content": "Give me the information of the capital of France in the JSON format.",
        },
    ],
    temperature=0,
    max_tokens=128,
    response_format={
        "type": "json_schema",
        "json_schema": {"name": "foo", "schema": json.loads(json_schema)},
    },
)

print_highlight(response.choices[0].message.content)

result:
{
"name": "Paris",
"population": 2161000
}

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.