[CI/Build] use setuptools-scm to set __version__ (#4738)

Co-authored-by: youkaichao <youkaichao@126.com>

.gitignore

@@ -196,5 +196,8 @@ _build/
 *_hip*
 hip_compat.h
 
+# version file generated by setuptools-scm
+/vllm/_version.py
+
 # Benchmark dataset
 benchmarks/*.json

Dockerfile

@@ -79,15 +79,13 @@ ENV MAX_JOBS=${max_jobs}
 ARG nvcc_threads=8
 ENV NVCC_THREADS=$nvcc_threads
 
-ARG buildkite_commit
-ENV BUILDKITE_COMMIT=${buildkite_commit}
-
 ARG USE_SCCACHE
 ARG SCCACHE_BUCKET_NAME=vllm-build-sccache
 ARG SCCACHE_REGION_NAME=us-west-2
 ARG SCCACHE_S3_NO_CREDENTIALS=0
 # if USE_SCCACHE is set, use sccache to speed up compilation
 RUN --mount=type=cache,target=/root/.cache/pip \
+    --mount=type=bind,source=.git,target=.git \
     if [ "$USE_SCCACHE" = "1" ]; then \
         echo "Installing sccache..." \
         && curl -L -o sccache.tar.gz https://github.com/mozilla/sccache/releases/download/v0.8.1/sccache-v0.8.1-x86_64-unknown-linux-musl.tar.gz \
@@ -107,6 +105,7 @@ RUN --mount=type=cache,target=/root/.cache/pip \
 ENV CCACHE_DIR=/root/.cache/ccache
 RUN --mount=type=cache,target=/root/.cache/ccache \
     --mount=type=cache,target=/root/.cache/pip \
+    --mount=type=bind,source=.git,target=.git  \
     if [ "$USE_SCCACHE" != "1" ]; then \
         python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38; \
     fi

Dockerfile.cpu

@@ -62,8 +62,10 @@ ENV VLLM_CPU_DISABLE_AVX512=${VLLM_CPU_DISABLE_AVX512}
 
 RUN --mount=type=cache,target=/root/.cache/pip \
     --mount=type=cache,target=/root/.cache/ccache \
+    --mount=type=bind,source=.git,target=.git \
     VLLM_TARGET_DEVICE=cpu python3 setup.py bdist_wheel && \
-    pip install dist/*.whl
+    pip install dist/*.whl && \
+    rm -rf dist
 
 WORKDIR /workspace/
 

Dockerfile.neuron

@@ -6,9 +6,12 @@ FROM $BASE_IMAGE
 RUN echo "Base image is $BASE_IMAGE"
 
 # Install some basic utilities
-RUN apt-get update \
-    && apt-get install python3 python3-pip -y \
-    && apt-get install -y ffmpeg libsm6 libxext6 libgl1 
+RUN apt-get update && \
+    apt-get install -y \
+        git \
+        python3 \
+        python3-pip \
+        ffmpeg libsm6 libxext6 libgl1
 
 ### Mount Point ###
 # When launching the container, mount the code directory to /app
@@ -22,17 +25,17 @@ RUN python3 -m pip install sentencepiece transformers==4.36.2 -U
 RUN python3 -m pip install transformers-neuronx --extra-index-url=https://pip.repos.neuron.amazonaws.com -U
 RUN python3 -m pip install --pre neuronx-cc==2.15.* --extra-index-url=https://pip.repos.neuron.amazonaws.com -U
 
-COPY ./vllm /app/vllm/vllm
-COPY ./setup.py /app/vllm/setup.py
-COPY ./requirements-common.txt /app/vllm/requirements-common.txt
-COPY ./requirements-neuron.txt /app/vllm/requirements-neuron.txt
+COPY . /app/vllm
 
 RUN cd /app/vllm \
-    && python3 -m pip install -U -r requirements-neuron.txt
+    && python3 -m pip install -U \
+        cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
+        -r requirements-neuron.txt
 
 ENV VLLM_TARGET_DEVICE neuron
-RUN cd /app/vllm \
-    && pip install -e . \
+RUN --mount=type=bind,source=.git,target=.git \
+    cd /app/vllm \
+    && pip install --no-build-isolation -v -e . \
     && cd ..
 
 CMD ["/bin/bash"]

Dockerfile.openvino

@@ -4,8 +4,9 @@
 FROM ubuntu:22.04 AS dev
 
 RUN apt-get update -y && \
-    apt-get install -y python3-pip git && \
-    apt-get install -y ffmpeg libsm6 libxext6 libgl1 
+    apt-get install -y \
+        git python3-pip \
+        ffmpeg libsm6 libxext6 libgl1
 WORKDIR /workspace
 
 # copy requirements

Dockerfile.ppc64le

@@ -16,9 +16,15 @@ COPY ./ /workspace/vllm
 WORKDIR /workspace/vllm
 
 # These packages will be in rocketce eventually
-RUN pip install -v cmake xformers torch==2.3.1 uvloop==0.20.0 -r requirements-cpu.txt --prefer-binary --extra-index-url https://repo.fury.io/mgiessing
-
-RUN VLLM_TARGET_DEVICE=cpu python3 setup.py install
+RUN --mount=type=cache,target=/root/.cache/pip  \
+    pip install -v --prefer-binary --extra-index-url https://repo.fury.io/mgiessing \
+        cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
+        torch==2.3.1 \
+        -r requirements-cpu.txt \
+        xformers uvloop==0.20.0
+
+RUN --mount=type=bind,source=.git,target=.git \
+    VLLM_TARGET_DEVICE=cpu python3 setup.py install
 
 WORKDIR /workspace/
 

Dockerfile.rocm

@@ -51,13 +51,15 @@ RUN python3 -m pip install --upgrade pip
 # TODO: implement sccache support across components
 RUN apt-get purge -y sccache; python3 -m pip uninstall -y sccache; rm -f "$(which sccache)"
 # Install torch == 2.5.0 on ROCm
-RUN case "$(ls /opt | grep -Po 'rocm-[0-9]\.[0-9]')" in \
+RUN --mount=type=cache,target=/root/.cache/pip \
+    case "$(ls /opt | grep -Po 'rocm-[0-9]\.[0-9]')" in \
         *"rocm-6.1"*) \
             python3 -m pip uninstall -y torch torchvision \
-            && python3 -m pip install --no-cache-dir --pre \
+            && python3 -m pip install --pre \
                 torch==2.5.0.dev20240726 \
+                cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
                 torchvision==0.20.0.dev20240726 \
-               --index-url https://download.pytorch.org/whl/nightly/rocm6.1;; \
+               --extra-index-url https://download.pytorch.org/whl/nightly/rocm6.1 ;; \
         *) ;; esac
 
 ENV LLVM_SYMBOLIZER_PATH=/opt/rocm/llvm/bin/llvm-symbolizer
@@ -138,6 +140,7 @@ ENV RAY_EXPERIMENTAL_NOSET_ROCR_VISIBLE_DEVICES=1
 ENV TOKENIZERS_PARALLELISM=false
 
 RUN --mount=type=cache,target=${CCACHE_DIR} \
+    --mount=type=bind,source=.git,target=.git \
     --mount=type=cache,target=/root/.cache/pip \
     python3 -m pip install -Ur requirements-rocm.txt \
     && case "$(ls /opt | grep -Po 'rocm-[0-9]\.[0-9]')" in \

Dockerfile.tpu

@@ -5,16 +5,25 @@ FROM $BASE_IMAGE
 WORKDIR /workspace
 
 # Install some basic utilities
-RUN apt-get update && apt-get install -y ffmpeg libsm6 libxext6 libgl1 
+RUN apt-get update && apt-get install -y \
+    git \
+    ffmpeg libsm6 libxext6 libgl1
 
 # Install the TPU and Pallas dependencies.
-RUN python3 -m pip install torch_xla[tpu] -f https://storage.googleapis.com/libtpu-releases/index.html
-RUN python3 -m pip install torch_xla[pallas] -f https://storage.googleapis.com/jax-releases/jax_nightly_releases.html -f https://storage.googleapis.com/jax-releases/jaxlib_nightly_releases.html
+RUN --mount=type=cache,target=/root/.cache/pip \
+    python3 -m pip install torch_xla[tpu] -f https://storage.googleapis.com/libtpu-releases/index.html
+RUN --mount=type=cache,target=/root/.cache/pip \
+    python3 -m pip install torch_xla[pallas] -f https://storage.googleapis.com/jax-releases/jax_nightly_releases.html -f https://storage.googleapis.com/jax-releases/jaxlib_nightly_releases.html
 
 # Build vLLM.
 COPY . /workspace/vllm
 ENV VLLM_TARGET_DEVICE="tpu"
-RUN cd /workspace/vllm && python3 -m pip install -r requirements-tpu.txt
+RUN --mount=type=cache,target=/root/.cache/pip \
+    --mount=type=bind,source=.git,target=.git \
+     cd /workspace/vllm && \
+    python3 -m pip install \
+        cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
+        -r requirements-tpu.txt
 RUN cd /workspace/vllm && python3 setup.py develop
 
 CMD ["/bin/bash"]

Dockerfile.xpu

@@ -7,15 +7,20 @@ RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRO
     echo "deb [arch=amd64,i386 signed-by=/usr/share/keyrings/intel-graphics.gpg] https://repositories.intel.com/graphics/ubuntu jammy arc" | tee /etc/apt/sources.list.d/intel.gpu.jammy.list && \
     chmod 644 /usr/share/keyrings/intel-graphics.gpg
 
-RUN apt-get update  -y \
-&& apt-get install -y curl libicu70 lsb-release git wget vim numactl python3 python3-pip ffmpeg libsm6 libxext6 libgl1 
+RUN apt-get update  -y && \
+    apt-get install -y curl libicu70 lsb-release git wget vim numactl python3 python3-pip ffmpeg libsm6 libxext6 libgl1
 
 COPY ./ /workspace/vllm
 
 WORKDIR /workspace/vllm
 
-RUN pip install -v -r requirements-xpu.txt --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
+RUN --mount=type=cache,target=/root/.cache/pip \
+    pip install -v --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/ \
+        cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
+        -r requirements-xpu.txt
 
-RUN VLLM_TARGET_DEVICE=xpu python3 setup.py install
+RUN --mount=type=cache,target=/root/.cache/pip \
+    --mount=type=bind,source=.git,target=.git \
+    VLLM_TARGET_DEVICE=xpu python3 setup.py install
 
 CMD ["/bin/bash"]

docs/source/getting_started/cpu-installation.rst

@@ -56,7 +56,7 @@ Build from source
 .. code-block:: console
 
     $ pip install --upgrade pip
-    $ pip install wheel packaging ninja "setuptools>=49.4.0" numpy
+    $ pip install cmake>=3.26 wheel packaging ninja "setuptools-scm>=8" numpy
     $ pip install -v -r requirements-cpu.txt --extra-index-url https://download.pytorch.org/whl/cpu
 
 - Third, build and install oneDNN library from source:

pyproject.toml

@@ -4,7 +4,8 @@ requires = [
     "cmake>=3.26",
     "ninja",
     "packaging",
-    "setuptools >= 49.4.0",
+    "setuptools>=61",
+    "setuptools-scm>=8.0",
     "torch == 2.4.0",
     "wheel",
     "jinja2",
@@ -19,6 +20,10 @@ exclude = [
     "examples/fp8/quantizer/quantize.py"
 ]
 
+[tool.ruff.lint.per-file-ignores]
+"vllm/version.py" = ["F401"]
+"vllm/_version.py" = ["ALL"]
+
 [tool.ruff.lint]
 select = [
     # pycodestyle
@@ -46,6 +51,9 @@ ignore = [
     "UP032",
 ]
 
+[tool.setuptools_scm]
+version_file = "vllm/_version.py"
+
 [tool.mypy]
 python_version = "3.8"
 

requirements-build.txt

@@ -2,7 +2,8 @@
 cmake>=3.26
 ninja
 packaging
-setuptools>=49.4.0
+setuptools>=61
+setuptools-scm>=8
 torch==2.4.0
 wheel
 jinja2

setup.py

@@ -5,7 +5,6 @@ import os
 import re
 import subprocess
 import sys
-import warnings
 from pathlib import Path
 from shutil import which
 from typing import Dict, List
@@ -14,6 +13,7 @@ import torch
 from packaging.version import Version, parse
 from setuptools import Extension, find_packages, setup
 from setuptools.command.build_ext import build_ext
+from setuptools_scm import get_version
 from torch.utils.cpp_extension import CUDA_HOME
 
 
@@ -28,34 +28,6 @@ def load_module_from_path(module_name, path):
 ROOT_DIR = os.path.dirname(__file__)
 logger = logging.getLogger(__name__)
 
-
-def embed_commit_hash():
-    try:
-        if "BUILDKITE_COMMIT" in os.environ:
-            # ci build
-            commit_id = os.environ["BUILDKITE_COMMIT"]
-        else:
-            commit_id = subprocess.check_output(["git", "rev-parse", "HEAD"],
-                                                encoding="utf-8").strip()
-
-        commit_contents = f'__commit__ = "{commit_id}"\n'
-
-        version_file = os.path.join(ROOT_DIR, "vllm", "commit_id.py")
-        with open(version_file, "w", encoding="utf-8") as f:
-            f.write(commit_contents)
-
-    except subprocess.CalledProcessError as e:
-        warnings.warn(f"Failed to get commit hash:\n{e}",
-                      RuntimeWarning,
-                      stacklevel=2)
-    except Exception as e:
-        warnings.warn(f"Failed to embed commit hash:\n{e}",
-                      RuntimeWarning,
-                      stacklevel=2)
-
-
-embed_commit_hash()
-
 # cannot import envs directly because it depends on vllm,
 #  which is not installed yet
 envs = load_module_from_path('envs', os.path.join(ROOT_DIR, 'vllm', 'envs.py'))
@@ -381,21 +353,9 @@ def get_path(*filepath) -> str:
     return os.path.join(ROOT_DIR, *filepath)
 
 
-def find_version(filepath: str) -> str:
-    """Extract version information from the given filepath.
-
-    Adapted from https://github.com/ray-project/ray/blob/0b190ee1160eeca9796bc091e07eaebf4c85b511/python/setup.py
-    """
-    with open(filepath) as fp:
-        version_match = re.search(r"^__version__ = ['\"]([^'\"]*)['\"]",
-                                  fp.read(), re.M)
-        if version_match:
-            return version_match.group(1)
-        raise RuntimeError("Unable to find version string.")
-
-
 def get_vllm_version() -> str:
-    version = find_version(get_path("vllm", "version.py"))
+    version = get_version()
+    sep = "+" if "+" not in version else "."  # dev versions might contain +
 
     if _no_device():
         if envs.VLLM_TARGET_DEVICE == "empty":
@@ -406,27 +366,27 @@ def get_vllm_version() -> str:
             cuda_version_str = cuda_version.replace(".", "")[:3]
             # skip this for source tarball, required for pypi
             if "sdist" not in sys.argv:
-                version += f"+cu{cuda_version_str}"
+                version += f"{sep}cu{cuda_version_str}"
     elif _is_hip():
         # Get the HIP version
         hipcc_version = get_hipcc_rocm_version()
         if hipcc_version != MAIN_CUDA_VERSION:
             rocm_version_str = hipcc_version.replace(".", "")[:3]
-            version += f"+rocm{rocm_version_str}"
+            version += f"{sep}rocm{rocm_version_str}"
     elif _is_neuron():
         # Get the Neuron version
         neuron_version = str(get_neuronxcc_version())
         if neuron_version != MAIN_CUDA_VERSION:
             neuron_version_str = neuron_version.replace(".", "")[:3]
-            version += f"+neuron{neuron_version_str}"
+            version += f"{sep}neuron{neuron_version_str}"
     elif _is_openvino():
-        version += "+openvino"
+        version += f"{sep}openvino"
     elif _is_tpu():
-        version += "+tpu"
+        version += f"{sep}tpu"
     elif _is_cpu():
-        version += "+cpu"
+        version += f"{sep}cpu"
     elif _is_xpu():
-        version += "+xpu"
+        version += f"{sep}xpu"
     else:
         raise RuntimeError("Unknown runtime environment")
 

tests/test_embedded_commit.py

@@ -2,6 +2,7 @@ import vllm
 
 
 def test_embedded_commit_defined():
-    assert vllm.__commit__ != "COMMIT_HASH_PLACEHOLDER"
-    # 7 characters is the length of a short commit hash
-    assert len(vllm.__commit__) >= 7
+    assert hasattr(vllm, "__version__")
+    assert hasattr(vllm, "__version_tuple__")
+    assert vllm.__version__ != "dev"
+    assert vllm.__version_tuple__ != (0, 0, "dev")

vllm/__init__.py

@@ -12,11 +12,11 @@ from vllm.outputs import (CompletionOutput, EmbeddingOutput,
 from vllm.pooling_params import PoolingParams
 from vllm.sampling_params import SamplingParams
 
-from .version import __commit__, __version__
+from .version import __version__, __version_tuple__
 
 __all__ = [
-    "__commit__",
     "__version__",
+    "__version_tuple__",
     "LLM",
     "ModelRegistry",
     "PromptType",

vllm/version.py

-import warnings
-
 try:
-    import vllm.commit_id
-
-    __commit__ = vllm.commit_id.__commit__
+    from ._version import __version__, __version_tuple__
 except Exception as e:
+    import warnings
+
     warnings.warn(f"Failed to read commit hash:\n{e}",
                   RuntimeWarning,
                   stacklevel=2)
-    __commit__ = "COMMIT_HASH_PLACEHOLDER"
 
-__version__ = "0.6.1.post2"
+    __version__ = "dev"
+    __version_tuple__ = (0, 0, __version__)