feat: drop python 3.8 for the coming EOL (#4948)

.github/workflows/ci.yml

@@ -28,7 +28,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: [3.8, 3.11, 3.12]
+        python-version: [3.9, 3.11, 3.12]
     name: unit-tests (python${{ matrix.python-version }}.${{ matrix.os }})
     runs-on: ${{ matrix.os }}
     steps:
@@ -86,14 +86,14 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: '3.8'
+          python-version: '3.9'
           cache: pip
       - name: Install dependencies
         run: pipx install pdm && pipx install nox
       - name: Run framework integration tests
         run: nox --session "framework-integration(framework='${{ matrix.framework }}')"
       - name: Disambiguate coverage filename
-        run: mv .coverage ".coverage.integrations.ubuntu-latest.3.8.${{ matrix.framework }}"
+        run: mv .coverage ".coverage.integrations.ubuntu-latest.3.9.${{ matrix.framework }}"
       - name: Upload coverage data
         uses: actions/upload-artifact@v3
         with:
@@ -104,7 +104,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: [3.8, 3.11, 3.12]
+        python-version: [3.9, 3.11, 3.12]
     name: monitoring-e2e-tests (python${{ matrix.python-version }}.${{ matrix.os }})
     runs-on: ${{ matrix.os }}
     timeout-minutes: 90
@@ -142,7 +142,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest]
-        python-version: [3.8, 3.11, 3.12]
+        python-version: [3.9, 3.11, 3.12]
         suite: ['bento_server_http', 'bento_new_sdk']
         exclude:
           - os: macos-latest

.python-version-default

@@ -1 +1 @@
-3.9
+3.11

DEVELOPMENT.md

@@ -12,7 +12,7 @@ If you are interested in proposing a new feature, make sure to create a new feat
 
 1. Make sure to have [Git](https://git-scm.com/),
    [pip](https://pip.pypa.io/en/stable/installation/),
-   [Python3.8+](https://www.python.org/downloads/), and
+   [Python3.9+](https://www.python.org/downloads/), and
    [PDM](https://pdm.fming.dev/latest/) installed.
 
    Optionally, make sure to have [GNU Make](https://www.gnu.org/software/make/)
@@ -75,7 +75,7 @@ If you are interested in proposing a new feature, make sure to create a new feat
 
 1. Confirm that you have the following installed:
 
-   - [Python3.8+](https://www.python.org/downloads/)
+   - [Python3.9+](https://www.python.org/downloads/)
    - VS Code with the [Python](https://marketplace.visualstudio.com/items?itemName=ms-python.python) and [Pylance](https://marketplace.visualstudio.com/items?itemName=ms-python.vscode-pylance) extensions
 
 2. Fork the BentoML project on [GitHub](https://github.com/bentoml/BentoML).
@@ -264,10 +264,10 @@ pip install git+https://github.com/{YOUR_GITHUB_USERNAME}/bentoml@{YOUR_REVISION
 And in your BentoML projects' `bentofile.yaml`, force the Bento to install this distribution, e.g.:
 
 ```yaml
-service: "service:svc"
-description: "file: ./README.md"
+service: 'service:svc'
+description: 'file: ./README.md'
 include:
-  - "*.py"
+  - '*.py'
 python:
   packages:
     - pandas

README.md

@@ -8,7 +8,6 @@
 
 🍱 Build model inference APIs and multi-model serving systems with any open-source or custom AI models. 👉 [Join our Slack community!](https://l.bentoml.com/join-slack)
 
-
 [![License: Apache-2.0](https://img.shields.io/badge/License-Apache%202-green.svg)](https://github.com/bentoml/BentoML?tab=Apache-2.0-1-ov-file)
 [![Releases](https://img.shields.io/github/v/release/bentoml/bentoml.svg)](https://github.com/bentoml/bentoml/releases)
 [![CI](https://github.com/bentoml/bentoml/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/bentoml/BentoML/actions/workflows/ci.yml?query=branch%3Amain)
@@ -25,13 +24,12 @@ BentoML is a Python library for building online serving systems optimized for AI
 - **👩‍💻 Fully customizable.** Easily implement your own APIs or task queues, with custom business logic, model inference and multi-model composition. Supports any ML framework, modality, and inference runtime.
 - **🚀 Ready for Production.** Develop, run and debug locally. Seamlessly deploy to production with Docker containers or [BentoCloud](https://www.bentoml.com/).
 
-
 ## Getting started
 
 Install BentoML:
 
 ```
-# Requires Python≥3.8
+# Requires Python≥3.9
 pip install -U bentoml
 ```
 
@@ -82,13 +80,13 @@ with bentoml.SyncHTTPClient('http://localhost:3000') as client:
 To deploy your BentoML Service code, first create a `bentofile.yaml` file to define its dependencies and environments. Find the full list of bentofile options [here](https://docs.bentoml.com/en/latest/guides/build-options.html).
 
 ```yaml
-service: "service:Summarization" # Entry service import path
+service: 'service:Summarization' # Entry service import path
 include:
-  - "*.py" # Include all .py files in current directory
+  - '*.py' # Include all .py files in current directory
 python:
   packages: # Python dependencies to include
-  - torch
-  - transformers
+    - torch
+    - transformers
 docker:
   python_version: 3.11
 ```

docs/source/get-started/installation.rst

@@ -7,7 +7,7 @@ BentoML is distributed as a Python package available on `PyPI <https://pypi.org/
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 - (Optional) `Git <https://git-scm.com/>`_ installed for `editable install <https://docs.bentoml.com/en/latest/quickstarts/install-bentoml.html#editable-install>`_.
 

docs/source/get-started/quickstart.rst

@@ -13,7 +13,7 @@ You can find all the project files in the `quickstart <https://github.com/bentom
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 
 Install dependencies

docs/source/guides/build-options.rst

@@ -398,7 +398,7 @@ The following ``docker`` field contains some basic Docker configurations:
 
     docker:
       distro: debian
-      python_version: "3.8.12"
+      python_version: "3.11"
       cuda_version: "11.6.2"
       system_packages:
         - libblas-dev

docs/source/use-cases/audio/whisperx.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoWhisperX GitHub r
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - If you want to test this project locally, install FFmpeg on your system.
 - Gain access to the model used in this project: `pyannote/segmentation-3.0 <https://huggingface.co/pyannote/segmentation-3.0>`_.

docs/source/use-cases/custom-models/mlflow.rst

@@ -9,7 +9,7 @@ This document explains how to serve and deploy an MLflow model with BentoML. You
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 

docs/source/use-cases/custom-models/xgboost.rst

@@ -9,7 +9,7 @@ This document explains how to serve and deploy an XGBoost model for predicting b
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 

docs/source/use-cases/diffusion-models/sdxl-lcm-lora.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoDiffusion GitHub
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - To run this BentoML Service locally, you need a Nvidia GPU with at least 12G VRAM.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.

docs/source/use-cases/diffusion-models/sdxl-turbo.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoDiffusion GitHub
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - To run this BentoML Service locally, you need a Nvidia GPU with at least 12G VRAM.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.

docs/source/use-cases/embeddings/clip-embeddings.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoClip GitHub repos
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 

docs/source/use-cases/embeddings/sentence-transformer.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoSentenceTransform
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 

docs/source/use-cases/large-language-models/vllm.rst

@@ -11,7 +11,7 @@ All the source code in this tutorial is available in the `BentoVLLM GitHub repos
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - If you want to test the project locally, you need a Nvidia GPU with least 16G VRAM.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.

docs/source/use-cases/multimodality/blip.rst

@@ -9,7 +9,7 @@ All the source code in this tutorial is available in the `BentoBlip GitHub repos
 Prerequisites
 -------------
 
-- Python 3.8+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
+- Python 3.9+ and ``pip`` installed. See the `Python downloads page <https://www.python.org/downloads/>`_ to learn more.
 - You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read :doc:`/get-started/quickstart` first.
 - (Optional) We recommend you create a virtual environment for dependency isolation. See the `Conda documentation <https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html>`_ or the `Python documentation <https://docs.python.org/3/library/venv.html>`_ for details.
 

examples/quickstart/README.md

@@ -4,7 +4,7 @@ This quickstart demonstrates how to build a text summarization application with
 
 ## Prerequisites
 
-Python 3.8+ and `pip` installed. See the [Python downloads page](https://www.python.org/downloads/) to learn more.
+Python 3.9+ and `pip` installed. See the [Python downloads page](https://www.python.org/downloads/) to learn more.
 
 ## Get started
 

noxfile.py

@@ -12,16 +12,15 @@ TEST_ARGS = [
     "-vv",
 ]
 
-PYTHON_VERSIONS = ["3.8", "3.9", "3.10", "3.11", "3.12"]
+PYTHON_VERSIONS = ["3.9", "3.10", "3.11", "3.12"]
 
 FRAMEWORK_DEPENDENCIES = {
-    "catboost": ["catboost"],
+    "catboost": ["catboost", "numpy<2"],
     "diffusers": ["diffusers", "transformers", "tokenizer"],
     "easyocr": ["easyocr"],
     "fastai": ["fastai"],
     "flax": [
-        "tensorflow~=2.13.1",
-        "pydantic<2",
+        "tensorflow",
         "flax; platform_system!='Windows'",
         "jax[cpu]; platform_system!='Windows'",
         "jaxlib; platform_system!='Windows'",
@@ -34,7 +33,7 @@ FRAMEWORK_DEPENDENCIES = {
     "pytorch": [],
     "pytorch_lightning": ["lightning"],
     "sklearn": ["scikit-learn"],
-    "tensorflow": ["tensorflow~=2.13.1", "pydantic<2"],
+    "tensorflow": ["tensorflow"],
     "torchscript": [],
     "xgboost": ["xgboost"],
     "detectron": ["detectron2"],

pyproject.toml

@@ -2,7 +2,7 @@
 name = "bentoml"
 description = "BentoML: The easiest way to serve AI apps and models"
 readme = { file = "README.md", content-type = "text/markdown" }
-requires-python = ">=3.8"
+requires-python = ">=3.9"
 keywords = ["BentoML", "Compound AI Systems", "MLOps", "LLMOps", "Model Serving", "Model Inference", "Model Deployment"]
 license = { text = "Apache-2.0" }
 authors = [{ name = "BentoML Team", email = "[email protected]" }]
@@ -14,7 +14,6 @@ classifiers = [
     "Operating System :: OS Independent",
     "Programming Language :: Python :: 3",
     "Programming Language :: Python :: 3 :: Only",
-    "Programming Language :: Python :: 3.8",
     "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
@@ -71,7 +70,6 @@ dependencies = [
     "aiosqlite>=0.20.0",
     "InquirerPy>=0.3.4",
     "uv",
-    "eval-type-backport>=0.2.0; python_version < \"3.9\"",
 ]
 dynamic = ["version"]
 [project.urls]

src/bentoml/_internal/external_typing/tensorflow.py

@@ -1,11 +1,11 @@
 import typing as t
 
+from tensorflow import Tensor
 from tensorflow.python.client.session import Session
 from tensorflow.python.eager.def_function import Function
 from tensorflow.python.eager.function_spec import FunctionSpec
 from tensorflow.python.framework.composite_tensor import CompositeTensor
 from tensorflow.python.framework.indexed_slices import IndexedSlices
-from tensorflow.python.framework.ops import Tensor
 from tensorflow.python.framework.sparse_tensor import SparseTensorSpec
 from tensorflow.python.framework.tensor_spec import TensorSpec
 from tensorflow.python.framework.type_spec import TypeSpec

src/bentoml/_internal/frameworks/tensorflow.py

@@ -411,7 +411,7 @@ class TensorflowTensorContainer(
 # NOTE: we only register the TensorflowTensorContainer as if eager execution is enabled.
 if tf.executing_eagerly():
     DataContainerRegistry.register_container(
-        LazyType("tensorflow.python.framework.ops", "Tensor"),
-        LazyType("tensorflow.python.framework.ops", "Tensor"),
+        LazyType("tensorflow", "Tensor"),
+        LazyType("tensorflow", "Tensor"),
         TensorflowTensorContainer,
     )

tests/integration/frameworks/mlflow/MNIST/conda.yaml

@@ -1,7 +1,7 @@
 channels:
   - conda-forge
 dependencies:
-  - python=3.8.2
+  - python=3.9
   - pip
   - pip:
       - mlflow

tests/integration/frameworks/models/tensorflow.py

@@ -242,44 +242,7 @@ native_multi_output_model2 = FrameworkTestModel(
     ],
 )
 
-keras_models = [
-    FrameworkTestModel(
-        name="tf2",
-        model=model,
-        save_kwargs={"signatures": {"__call__": {"batchable": True, "batch_dim": 0}}},
-        configurations=[
-            Config(
-                test_inputs={
-                    "__call__": [
-                        Input(
-                            input_args=[inp],
-                            expected=lambda out: np.isclose(out, [[15.0]]).all(),
-                        ),
-                        Input(
-                            input_args=[inp],
-                            expected=lambda out: np.isclose(out, [[15.0]]).all(),
-                        ),
-                        Input(
-                            input_args=[inp],
-                            expected=lambda out: np.isclose(out, [[15.0]]).all(),
-                        ),
-                    ],
-                },
-            ),
-        ],
-    )
-    for model in [
-        make_keras_functional_model(),
-        make_keras_sequential_model(),
-    ]
-    for inp in [
-        input_tensor,
-        input_array,
-        input_data,
-    ]
-]
-
-models: list[FrameworkTestModel] = keras_models + [
+models: list[FrameworkTestModel] = [
     native_model,
     native_multi_input_model,
     native_multi_output_model,

tests/integration/frameworks/test_tensorflow_unit.py

@@ -80,11 +80,11 @@ def test_register_container():
     )
 
     assert (
-        LazyType("tensorflow.python.framework.ops", "Tensor")
+        LazyType("tensorflow", "Tensor")
         not in DataContainerRegistry.CONTAINER_BATCH_TYPE_MAP
     )
 
     assert (
-        LazyType("tensorflow.python.framework.ops", "Tensor")
+        LazyType("tensorflow", "Tensor")
         not in DataContainerRegistry.CONTAINER_SINGLE_TYPE_MAP
     )