2025-02-10 09:36:09 +08:00
|
|
|
[build-system]
|
|
|
|
|
# Should be mirrored in requirements.txt
|
|
|
|
|
requires = [
|
2025-11-28 18:06:39 +08:00
|
|
|
"attrs",
|
2025-04-03 14:52:34 +08:00
|
|
|
"cmake>=3.26",
|
2025-02-10 09:36:09 +08:00
|
|
|
"decorator",
|
2025-05-12 22:04:48 +08:00
|
|
|
"einops",
|
2025-11-28 18:06:39 +08:00
|
|
|
"googleapis-common-protos",
|
2025-04-07 16:07:21 +08:00
|
|
|
"numpy<2.0.0",
|
2025-04-16 09:28:58 +08:00
|
|
|
"packaging",
|
2025-04-12 10:24:53 +08:00
|
|
|
"pip",
|
2025-04-07 16:07:21 +08:00
|
|
|
"pybind11",
|
2025-02-10 09:36:09 +08:00
|
|
|
"pyyaml",
|
|
|
|
|
"scipy",
|
2025-10-22 14:13:32 +08:00
|
|
|
"pandas",
|
|
|
|
|
"pandas-stubs",
|
2025-11-28 18:06:39 +08:00
|
|
|
"psutil",
|
2025-04-07 16:07:21 +08:00
|
|
|
"setuptools>=64",
|
|
|
|
|
"setuptools-scm>=8",
|
2025-11-25 15:32:24 +08:00
|
|
|
"transformers<=4.57.1",
|
2026-01-22 17:59:06 +08:00
|
|
|
"torch-npu==2.9.0",
|
|
|
|
|
"torch==2.9.0",
|
2025-08-05 08:43:24 +08:00
|
|
|
"torchvision",
|
2025-04-27 17:28:29 +08:00
|
|
|
"wheel",
|
2025-06-08 22:33:37 +08:00
|
|
|
"msgpack",
|
|
|
|
|
"quart",
|
|
|
|
|
"numba",
|
2026-01-23 09:52:55 +08:00
|
|
|
"xgrammar>=0.1.30",
|
2025-12-16 15:58:27 +08:00
|
|
|
"fastapi<0.124.0",
|
2025-10-25 18:07:54 +08:00
|
|
|
"opencv-python-headless<=4.11.0.86", # Required to avoid numpy version conflict with vllm
|
2026-01-05 19:15:49 +08:00
|
|
|
"compressed_tensors>=0.11.0",
|
2026-01-22 09:29:50 +08:00
|
|
|
"arctic-inference==0.1.1",
|
|
|
|
|
"triton-ascend==3.2.0"
|
2025-02-10 09:36:09 +08:00
|
|
|
]
|
|
|
|
|
build-backend = "setuptools.build_meta"
|
2025-07-25 22:16:10 +08:00
|
|
|
|
|
|
|
|
[tool.pymarkdown]
|
|
|
|
|
plugins.md004.style = "sublist" # ul-style
|
|
|
|
|
plugins.md007.indent = 4 # ul-indent
|
|
|
|
|
plugins.md007.start_indented = true # ul-indent
|
|
|
|
|
plugins.md013.enabled = false # line-length
|
|
|
|
|
plugins.md041.enabled = false # first-line-h1
|
|
|
|
|
plugins.md033.enabled = false # inline-html
|
|
|
|
|
plugins.md046.enabled = false # code-block-style
|
|
|
|
|
plugins.md024.allow_different_nesting = true # no-duplicate-headers
|
|
|
|
|
plugins.md029.enabled = false # ol-prefix
|
2026-01-13 15:29:34 +08:00
|
|
|
|
|
|
|
|
[tool.ruff]
|
|
|
|
|
# TODO: according to PEP8, there should be 80 characters per line
|
|
|
|
|
line-length = 120
|
|
|
|
|
# Folder to be modified
|
|
|
|
|
exclude = [
|
|
|
|
|
"tests/**",
|
2026-01-19 08:59:46 +08:00
|
|
|
"vllm_ascend/attention/mla_v1.py",
|
|
|
|
|
"vllm_ascend/attention/sfa_v1.py",
|
[Lint]Style: Convert `vllm-ascend/compilation` to ruff format (#5912)
### What this PR does / why we need it?
Convert `vllm-ascend/compilation` to ruff format.
### Does this PR introduce _any_ user-facing change?
During this migration, we encountered some **errors** in our CI and
testing environments, such as:
```
vllm_ascend/utils.py:653: in <module>
def register_ascend_customop(vllm_config: VllmConfig | None = None):
^^^^^^^^^^^^^^^^^
E TypeError: unsupported operand type(s) for |: 'NoneType' and 'NoneType'
```
**1. Root Cause Analysis:**
The project uses a common pattern to break circular dependencies:
```python
if TYPE_CHECKING:
from vllm.config import VllmConfig
else:
VllmConfig = None # Placeholder assigned at runtime
```
When Python parses the function definition `def
register_ascend_customop(vllm_config: VllmConfig | None)`, it attempts
to evaluate the expression `VllmConfig | None`.
Since `VllmConfig` is assigned `None` at runtime, the expression
effectively becomes `None | None`. In Python, `None` is an instance of
`NoneType`. While the `|` operator is implemented for Type objects
(classes), it is not supported for `NoneType` instances, leading to the
`TypeError` shown above.
**2. Solution:**
To maintain the modern `|` syntax required by our new linting standards
while preserving our dependency management strategy, I have introduced:
```python
from __future__ import annotations
```
at the top of the affected files. This enables **Postponed Evaluation of
Annotations (PEP 563)**.
**3. Impact and Benefits:**
- By enabling `annotations`, Python no longer executes the `VllmConfig |
None` operation during module load. Instead, it stores the annotation as
a string literal, completely avoiding the `None | None` calculation.
- We can keep the `VllmConfig = None` placeholders. This ensures that
other modules can still import these symbols without triggering an
`ImportError`, maintaining a stable dependency graph.
- IDEs and static type checkers (MyPy/Pyright) continue to resolve the
types correctly. This allows us to use modern syntax without sacrificing
type safety or runtime stability.
- The only side effect is that `__annotations__` will now return strings
instead of type objects. Since this module does not use runtime type
enforcement or reflection, this change has zero negative impact on
existing functionality.
### How was this patch tested?
- vLLM version: v0.13.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/11b6af5280d6d6dfb8953af16e67b25f819b3be9
---------
Signed-off-by: MrZ20 <2609716663@qq.com>
2026-01-16 20:57:46 +08:00
|
|
|
"vllm_ascend/core",
|
|
|
|
|
"vllm_ascend/distributed",
|
|
|
|
|
"vllm_ascend/eplb",
|
|
|
|
|
"vllm_ascend/kv_offload",
|
|
|
|
|
"vllm_ascend/lora",
|
|
|
|
|
"vllm_ascend/model_loader",
|
2026-01-23 14:59:19 +08:00
|
|
|
"vllm_ascend/ops/fused_moe",
|
|
|
|
|
"vllm_ascend/ops/activation.py",
|
|
|
|
|
"vllm_ascend/ops/flashcomm2_oshard_manager.py",
|
|
|
|
|
"vllm_ascend/ops/layer_shard_linear.py",
|
|
|
|
|
"vllm_ascend/ops/layernorm.py",
|
|
|
|
|
"vllm_ascend/ops/linear_op.py",
|
|
|
|
|
"vllm_ascend/ops/linear.py",
|
|
|
|
|
"vllm_ascend/ops/mla.py",
|
|
|
|
|
"vllm_ascend/ops/mm_encoder_attention.py",
|
|
|
|
|
"vllm_ascend/ops/register_custom_ops.py",
|
|
|
|
|
"vllm_ascend/ops/rotary_embedding.py",
|
|
|
|
|
"vllm_ascend/ops/vocab_parallel_embedding.py",
|
|
|
|
|
"vllm_ascend/ops/weight_prefetch.py",
|
|
|
|
|
"vllm_ascend/ops/__init__.py",
|
[Lint]Style: Convert `vllm-ascend/compilation` to ruff format (#5912)
### What this PR does / why we need it?
Convert `vllm-ascend/compilation` to ruff format.
### Does this PR introduce _any_ user-facing change?
During this migration, we encountered some **errors** in our CI and
testing environments, such as:
```
vllm_ascend/utils.py:653: in <module>
def register_ascend_customop(vllm_config: VllmConfig | None = None):
^^^^^^^^^^^^^^^^^
E TypeError: unsupported operand type(s) for |: 'NoneType' and 'NoneType'
```
**1. Root Cause Analysis:**
The project uses a common pattern to break circular dependencies:
```python
if TYPE_CHECKING:
from vllm.config import VllmConfig
else:
VllmConfig = None # Placeholder assigned at runtime
```
When Python parses the function definition `def
register_ascend_customop(vllm_config: VllmConfig | None)`, it attempts
to evaluate the expression `VllmConfig | None`.
Since `VllmConfig` is assigned `None` at runtime, the expression
effectively becomes `None | None`. In Python, `None` is an instance of
`NoneType`. While the `|` operator is implemented for Type objects
(classes), it is not supported for `NoneType` instances, leading to the
`TypeError` shown above.
**2. Solution:**
To maintain the modern `|` syntax required by our new linting standards
while preserving our dependency management strategy, I have introduced:
```python
from __future__ import annotations
```
at the top of the affected files. This enables **Postponed Evaluation of
Annotations (PEP 563)**.
**3. Impact and Benefits:**
- By enabling `annotations`, Python no longer executes the `VllmConfig |
None` operation during module load. Instead, it stores the annotation as
a string literal, completely avoiding the `None | None` calculation.
- We can keep the `VllmConfig = None` placeholders. This ensures that
other modules can still import these symbols without triggering an
`ImportError`, maintaining a stable dependency graph.
- IDEs and static type checkers (MyPy/Pyright) continue to resolve the
types correctly. This allows us to use modern syntax without sacrificing
type safety or runtime stability.
- The only side effect is that `__annotations__` will now return strings
instead of type objects. Since this module does not use runtime type
enforcement or reflection, this change has zero negative impact on
existing functionality.
### How was this patch tested?
- vLLM version: v0.13.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/11b6af5280d6d6dfb8953af16e67b25f819b3be9
---------
Signed-off-by: MrZ20 <2609716663@qq.com>
2026-01-16 20:57:46 +08:00
|
|
|
"vllm_ascend/patch",
|
|
|
|
|
"vllm_ascend/quantization",
|
|
|
|
|
"vllm_ascend/sample",
|
|
|
|
|
"vllm_ascend/spec_decode",
|
|
|
|
|
"vllm_ascend/worker",
|
|
|
|
|
"vllm_ascend/xlite",
|
2026-01-13 15:29:34 +08:00
|
|
|
]
|
|
|
|
|
|
|
|
|
|
[tool.ruff.lint]
|
|
|
|
|
select = [
|
|
|
|
|
# pycodestyle
|
|
|
|
|
"E",
|
|
|
|
|
# Pyflakes
|
|
|
|
|
"F",
|
|
|
|
|
# pyupgrade
|
|
|
|
|
"UP",
|
|
|
|
|
# flake8-bugbear
|
|
|
|
|
"B",
|
|
|
|
|
# flake8-simplify
|
|
|
|
|
"SIM",
|
|
|
|
|
# isort
|
|
|
|
|
"I",
|
|
|
|
|
# flake8-logging-format
|
|
|
|
|
"G",
|
|
|
|
|
]
|
|
|
|
|
ignore = [
|
|
|
|
|
# star imports
|
|
|
|
|
"F405", "F403",
|
|
|
|
|
# lambda expression assignment
|
|
|
|
|
"E731",
|
|
|
|
|
# zip without `strict=`
|
|
|
|
|
"B905",
|
|
|
|
|
# Loop control variable not used within loop body
|
|
|
|
|
"B007",
|
|
|
|
|
# f-string format
|
|
|
|
|
"UP032",
|
|
|
|
|
# TODO: FIE ME
|
|
|
|
|
"G004",
|
|
|
|
|
"B904",
|
|
|
|
|
"SIM108",
|
|
|
|
|
"SIM102"
|
|
|
|
|
]
|
|
|
|
|
|
|
|
|
|
[tool.ruff.format]
|
|
|
|
|
docstring-code-format = true
|