[Refactor] Benchmark: Add typed DatasetArgs/Loader registry and CPU dataset unit tests (#19147)

Co-authored-by: Liangsheng Yin <lsyincs@gmail.com>

benchmark/hicache/bench_mix.py

@@ -13,7 +13,7 @@ from functools import wraps
 import aiohttp
 
 from sglang.bench_serving import RequestFuncOutput
-from sglang.benchmark.datasets import sample_random_requests
+from sglang.benchmark.datasets.random import sample_random_requests
 from sglang.benchmark.utils import get_tokenizer, remove_prefix
 
 # Set up logger

benchmark/hicache/bench_multiturn.py

@@ -11,7 +11,7 @@ import numpy as np
 import requests
 from tqdm.asyncio import tqdm
 
-from sglang.benchmark.datasets import sample_random_requests
+from sglang.benchmark.datasets.random import sample_random_requests
 from sglang.benchmark.utils import get_tokenizer
 from sglang.test.kits.cache_hit_kit import async_request_sglang_generate, gen_payload
 

benchmark/hicache/data_processing.py

@@ -11,12 +11,12 @@ from nextqa import NExTQALoader
 from tqdm.asyncio import tqdm
 from transformers import PreTrainedTokenizerBase
 
-from sglang.benchmark.datasets import (
+from sglang.benchmark.datasets.common import (
     SHAREGPT_FILENAME,
     SHAREGPT_REPO_ID,
     gen_prompt,
-    get_gen_prefix_cache_path,
 )
+from sglang.benchmark.datasets.generated_shared_prefix import get_gen_prefix_cache_path
 from sglang.benchmark.utils import download_and_cache_hf_file
 from sglang.lang.chat_template import get_chat_template, get_chat_template_by_model_path
 from sglang.srt.entrypoints.openai.protocol import ChatCompletionMessageContentPart