[Misc] Rename assets for testing (#17575)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

tests/models/multimodal/processing/test_phi3v.py

@@ -4,7 +4,7 @@ import pytest
 
 from vllm.multimodal import MULTIMODAL_REGISTRY
 
-from ....conftest import _ImageAssets
+from ....conftest import ImageTestAssets
 from ...utils import build_model_context
 
 
@@ -22,7 +22,7 @@ from ...utils import build_model_context
 @pytest.mark.parametrize("num_imgs", [1, 2])
 @pytest.mark.parametrize("kwargs_on_init", [True, False])
 def test_processor_override(
-    image_assets: _ImageAssets,
+    image_assets: ImageTestAssets,
     model_id: str,
     mm_processor_kwargs: dict[str, int],
     expected_toks_per_img: int,

tests/models/multimodal/processing/test_phi4mm.py

@@ -4,7 +4,7 @@ import pytest
 
 from vllm.multimodal import MULTIMODAL_REGISTRY
 
-from ....conftest import _ImageAssets
+from ....conftest import ImageTestAssets
 from ...utils import build_model_context
 
 
@@ -22,7 +22,7 @@ from ...utils import build_model_context
 @pytest.mark.parametrize("num_imgs", [1, 2])
 @pytest.mark.parametrize("kwargs_on_init", [True, False])
 def test_processor_override(
-    image_assets: _ImageAssets,
+    image_assets: ImageTestAssets,
     model_id: str,
     mm_processor_kwargs: dict[str, int],
     expected_toks_per_img: int,

tests/models/multimodal/processing/test_qwen2_vl.py

@@ -4,7 +4,7 @@ import pytest
 
 from vllm.multimodal import MULTIMODAL_REGISTRY
 
-from ....conftest import _ImageAssets
+from ....conftest import ImageTestAssets
 from ...utils import build_model_context
 
 
@@ -19,7 +19,7 @@ from ...utils import build_model_context
 @pytest.mark.parametrize("num_imgs", [1, 2])
 @pytest.mark.parametrize("kwargs_on_init", [True, False])
 def test_processor_override(
-    image_assets: _ImageAssets,
+    image_assets: ImageTestAssets,
     model_id: str,
     mm_processor_kwargs: dict[str, object],
     expected_toks_per_img: int,

tests/models/multimodal/processing/test_smolvlm.py

@@ -5,7 +5,7 @@ from transformers import SmolVLMConfig
 
 from vllm.multimodal import MULTIMODAL_REGISTRY
 
-from ....conftest import _ImageAssets
+from ....conftest import ImageTestAssets
 from ...utils import build_model_context
 
 
@@ -21,7 +21,7 @@ from ...utils import build_model_context
 @pytest.mark.parametrize("num_imgs", [1, 2])
 @pytest.mark.parametrize("kwargs_on_init", [True, False])
 def test_processor_override(
-    image_assets: _ImageAssets,
+    image_assets: ImageTestAssets,
     model_id: str,
     mm_processor_kwargs: dict[str, object],
     expected_toks_per_img: int,

tests/models/quantization/test_awq.py

@@ -7,7 +7,7 @@ import torch
 
 from vllm.multimodal.image import rescale_image_size
 
-from ...conftest import IMAGE_ASSETS, VllmRunner, _ImageAssets
+from ...conftest import IMAGE_ASSETS, ImageTestAssets, VllmRunner
 from ..utils import check_logprobs_close
 
 HF_IMAGE_PROMPTS = IMAGE_ASSETS.prompts({
@@ -20,7 +20,7 @@ HF_IMAGE_PROMPTS = IMAGE_ASSETS.prompts({
 
 def run_awq_test(
     vllm_runner: type[VllmRunner],
-    image_assets: _ImageAssets,
+    image_assets: ImageTestAssets,
     source_model: str,
     quant_model: str,
     *,

vllm/assets/audio.py

@@ -18,19 +18,25 @@ except ImportError:
 
 ASSET_DIR = "multimodal_asset"
 
+AudioAssetName = Literal["winning_call", "mary_had_lamb"]
+
 
 @dataclass(frozen=True)
 class AudioAsset:
-    name: Literal["winning_call", "mary_had_lamb"]
+    name: AudioAssetName
+
+    @property
+    def filename(self) -> str:
+        return f"{self.name}.ogg"
 
     @property
     def audio_and_sample_rate(self) -> tuple[npt.NDArray, float]:
-        audio_path = get_vllm_public_assets(filename=f"{self.name}.ogg",
+        audio_path = get_vllm_public_assets(filename=self.filename,
                                             s3_prefix=ASSET_DIR)
         return librosa.load(audio_path, sr=None)
 
     def get_local_path(self) -> Path:
-        return get_vllm_public_assets(filename=f"{self.name}.ogg",
+        return get_vllm_public_assets(filename=self.filename,
                                       s3_prefix=ASSET_DIR)
 
     @property

vllm/assets/image.py

@@ -10,10 +10,12 @@ from .base import get_vllm_public_assets
 
 VLM_IMAGES_DIR = "vision_model_images"
 
+ImageAssetName = Literal["stop_sign", "cherry_blossom"]
+
 
 @dataclass(frozen=True)
 class ImageAsset:
-    name: Literal["stop_sign", "cherry_blossom"]
+    name: ImageAssetName
 
     @property
     def pil_image(self) -> Image.Image:

vllm/assets/video.py

@@ -2,7 +2,7 @@
 
 from dataclasses import dataclass
 from functools import lru_cache
-from typing import Literal, Optional
+from typing import ClassVar, Literal, Optional
 
 import cv2
 import numpy as np
@@ -76,20 +76,31 @@ def video_to_pil_images_list(path: str,
     ]
 
 
+VideoAssetName = Literal["baby_reading"]
+
+
 @dataclass(frozen=True)
 class VideoAsset:
-    name: Literal["sample_demo_1"]
+    name: VideoAssetName
     num_frames: int = -1
 
+    _NAME_TO_FILE: ClassVar[dict[VideoAssetName, str]] = {
+        "baby_reading": "sample_demo_1.mp4",
+    }
+
+    @property
+    def filename(self) -> str:
+        return self._NAME_TO_FILE[self.name]
+
     @property
     def pil_images(self) -> list[Image.Image]:
-        video_path = download_video_asset(self.name + ".mp4")
+        video_path = download_video_asset(self.filename)
         ret = video_to_pil_images_list(video_path, self.num_frames)
         return ret
 
     @property
     def np_ndarrays(self) -> npt.NDArray:
-        video_path = download_video_asset(self.name + ".mp4")
+        video_path = download_video_asset(self.filename)
         ret = video_to_ndarrays(video_path, self.num_frames)
         return ret
 
@@ -99,5 +110,5 @@ class VideoAsset:
         
         See also: examples/offline_inference/qwen2_5_omni/only_thinker.py
         """
-        video_path = download_video_asset(self.name + ".mp4")
+        video_path = download_video_asset(self.filename)
         return librosa.load(video_path, sr=sampling_rate)[0]