Add Python 3.9 support (#15515) (#16883)

This includes several things:

* added per-provider support for python version. Each provider
  can now declare python versions it does not support
* excluded ldap core extra from Python 3.9.
* skip relevant tests in Python 3.9

(cherry picked from commit ce44b62)

BREEZE.rst

@@ -1263,7 +1263,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   -a, --install-airflow-version INSTALL_AIRFLOW_VERSION
           Uses different version of Airflow when building PROD image.
@@ -1493,7 +1493,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   -I, --production-image
           Use production image for entering the environment and builds (not for tests).
@@ -1560,7 +1560,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   -v, --verbose
           Show verbose information about executed docker, kind, kubectl, helm commands. Useful for
@@ -1683,7 +1683,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
 
   ####################################################################################################
@@ -1878,7 +1878,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   -b, --backend BACKEND
           Backend to use for tests - it determines which database is used.
@@ -1942,7 +1942,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   -F, --force-build-images
           Forces building of the local docker images. The images are rebuilt
@@ -2349,7 +2349,7 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           One of:
 
-                 3.6 3.7 3.8
+                 3.6 3.7 3.8 3.9
 
   ****************************************************************************************************
    Choose backend to run for Airflow

CI.rst

@@ -57,7 +57,7 @@ Container Registry used as cache
 For the CI builds of our we are using Container Registry to store results of the "Build Image" workflow
 and pass it to the "CI Build" workflow.
 
-Currently in main version of Airflow we run tests in 3 different versions of Python (3.6, 3.7, 3.8)
+Currently in main version of Airflow we run tests in 4 different versions of Python (3.6, 3.7, 3.8, 3.9)
 which means that we have to build 6 images (3 CI ones and 3 PROD ones). Yet we run around 12 jobs
 with each of the CI images. That is a lot of time to just build the environment to run. Therefore
 we are utilising ``pull_request_target`` feature of GitHub Actions.
@@ -779,7 +779,7 @@ The image names follow the patterns:
 +--------------+----------------------------+--------------------------------+--------------------------------------------------------------------------------------------+
 
 * <BRANCH> might be either "main" or "v1-10-test" or "v2-*-test"
-* <X.Y> - Python version (Major + Minor). For "main" and "v2-*-test" should be in ["3.6", "3.7", "3.8"].
+* <X.Y> - Python version (Major + Minor). For "main" and "v2-*-test" should be in ["3.6", "3.7", "3.8", "3.9"].
 * <COMMIT_SHA> - for images that get merged to "main", "v2-*-test" of "v1-10-test", or built as part of a
   pull request the images are tagged with the (full lenght) commit SHA of that particular branch. For pull
   requests the SHA used is the tip of the pull request branch.

CONTRIBUTING.rst

@@ -873,7 +873,7 @@ This can be done by running this (it utilizes parallel preparation of the constr
 
 .. code-block:: bash
 
-    export CURRENT_PYTHON_MAJOR_MINOR_VERSIONS_AS_STRING="3.6 3.7 3.8"
+    export CURRENT_PYTHON_MAJOR_MINOR_VERSIONS_AS_STRING="3.6 3.7 3.8 3.9"
     for python_version in $(echo "${CURRENT_PYTHON_MAJOR_MINOR_VERSIONS_AS_STRING}")
     do
       ./breeze build-image --upgrade-to-newer-dependencies --python ${python_version} --build-cache-local

CONTRIBUTORS_QUICK_START.rst

@@ -119,7 +119,7 @@ Pyenv and setting up virtual-env
       libreadline-dev libsqlite3-dev wget curl llvm libncurses5-dev libncursesw5-dev \
       xz-utils tk-dev libffi-dev liblzma-dev python-openssl git
 
-  $ sudo apt install build-essentials python3.6-dev python3.7-dev python3.8-dev python-dev openssl \
+  $ sudo apt install build-essentials python3.6-dev python3.7-dev python3.8-dev python3.9-dev python-dev openssl \
        sqlite sqlite-dev default-libmysqlclient-dev libmysqld-dev postgresql
 
 2. Install pyenv

Dockerfile

@@ -44,7 +44,7 @@ ARG AIRFLOW_GID="50000"
 
 ARG PYTHON_BASE_IMAGE="python:3.6-slim-buster"
 
-ARG AIRFLOW_PIP_VERSION=21.1.1
+ARG AIRFLOW_PIP_VERSION=21.1.2
 
 # By default PIP has progress bar but you can disable it.
 ARG PIP_PROGRESS_BAR="on"
@@ -232,11 +232,10 @@ ARG INSTALL_FROM_DOCKER_CONTEXT_FILES=""
 ARG INSTALL_FROM_PYPI="true"
 # Those are additional constraints that are needed for some extras but we do not want to
 # Force them on the main Airflow package.
-# * chardet<4 and certifi<2021.0.0 required to keep snowflake happy
-# * urllib3 - required to keep boto3 happy
+# * certifi<2021.0.0 required to keep snowflake happy
 # * pyjwt<2.0.0: flask-jwt-extended requires it
 # * dill<0.3.3 required by apache-beam
-ARG EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS="chardet<4 urllib3<1.26 pyjwt<2.0.0 dill<0.3.3 certifi<2021.0.0"
+ARG EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS="pyjwt<2.0.0 dill<0.3.3 certifi<2021.0.0"
 ARG CONTINUE_ON_PIP_CHECK_FAILURE="false"
 
 

Dockerfile.ci

@@ -207,7 +207,7 @@ ARG AIRFLOW_PRE_CACHED_PIP_PACKAGES="true"
 # By default in the image, we are installing all providers when installing from sources
 ARG INSTALL_PROVIDERS_FROM_SOURCES="true"
 ARG INSTALL_FROM_PYPI="true"
-ARG AIRFLOW_PIP_VERSION=21.1.1
+ARG AIRFLOW_PIP_VERSION=21.1.2
 # Setup PIP
 # By default PIP install run without cache to make image smaller
 ARG PIP_NO_CACHE_DIR="true"
@@ -251,13 +251,11 @@ ENV AIRFLOW_REPO=${AIRFLOW_REPO}\
 
 # Those are additional constraints that are needed for some extras but we do not want to
 # force them on the main Airflow package. Those limitations are:
-# * chardet<4 and certifi<2021.0.0: required by snowflake provider
+# * certifi<2021.0.0: required by snowflake provider
 # * lazy-object-proxy<1.5.0: required by astroid
-# * pyOpenSSL: required by snowflake provider https://github.com/snowflakedb/snowflake-connector-python/blob/v2.3.6/setup.py#L201
-# * urllib3<1.26: Required to keep boto3 happy
 # * pyjwt<2.0.0: flask-jwt-extended requires it
 # * dill<0.3.3 required by apache-beam
-ARG EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS="chardet<4 lazy-object-proxy<1.5.0 pyOpenSSL<20.0.0 urllib3<1.26 pyjwt<2.0.0 dill<0.3.3 certifi<2021.0.0"
+ARG EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS="lazy-object-proxy<1.5.0 pyjwt<2.0.0 dill<0.3.3 certifi<2021.0.0"
 ARG UPGRADE_TO_NEWER_DEPENDENCIES="false"
 ENV EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS=${EAGER_UPGRADE_ADDITIONAL_REQUIREMENTS} \
     UPGRADE_TO_NEWER_DEPENDENCIES=${UPGRADE_TO_NEWER_DEPENDENCIES}

IMAGES.rst

@@ -81,7 +81,7 @@ where:
   built from branches so they change over time. The ``2.*.*`` labels are built from git tags
   and they are "fixed" once built.
 * ``PYTHON_MAJOR_MINOR_VERSION`` - version of Python used to build the image. Examples: ``3.6``, ``3.7``,
-  ``3.8``
+  ``3.8``, ``3.9``
 * The ``-ci`` suffix is added for CI images
 * The ``-manifest`` is added for manifest images (see below for explanation of manifest images)
 

LOCAL_VIRTUALENV.rst

@@ -51,8 +51,8 @@ Required Software Packages
 Use system-level package managers like yum, apt-get for Linux, or
 Homebrew for macOS to install required software packages:
 
-* Python (One of: 3.6, 3.7, 3.8)
-* MySQL
+* Python (One of: 3.6, 3.7, 3.8, 3.9)
+* MySQL 5.7+
 * libxml
 
 Refer to the `Dockerfile.ci <Dockerfile.ci>`__ for a comprehensive list
@@ -102,7 +102,7 @@ Creating a Local virtualenv
 
 To use your IDE for Airflow development and testing, you need to configure a virtual
 environment. Ideally you should set up virtualenv for all Python versions that Airflow
-supports (3.6, 3.7, 3.8).
+supports (3.6, 3.7, 3.8, 3.9).
 
 To create and initialize the local virtualenv:
 

PULL_REQUEST_WORKFLOW.rst

@@ -58,7 +58,7 @@ We approached the problem by:
 3) Even more optimisation came from limiting the scope of tests to only "default" matrix parameters. So far
    in Airflow we always run all tests for all matrix combinations. The primary matrix components are:
 
-   * Python versions (currently 3.6, 3.7, 3.8)
+   * Python versions (currently 3.6, 3.7, 3.8, 3.9)
    * Backend types (currently MySQL/Postgres)
    * Backed version (currently MySQL 5.7, MySQL 8, Postgres 9.6, Postgres 13
 

README.md

@@ -78,9 +78,9 @@ Airflow is not a streaming solution, but it is often used to process real-time d
 
 Apache Airflow is tested with:
 
-|                      | Main version (dev)        | Stable version (2.0.2)   |
+|                      | Main version (dev)        | Stable version (2.1.0)   |
 | -------------------- | ------------------------- | ------------------------ |
-| Python               | 3.6, 3.7, 3.8             | 3.6, 3.7, 3.8            |
+| Python               | 3.6, 3.7, 3.8, 3.9        | 3.6, 3.7, 3.8            |
 | Kubernetes           | 1.20, 1.19, 1.18          | 1.20, 1.19, 1.18         |
 | PostgreSQL           | 9.6, 10, 11, 12, 13       | 9.6, 10, 11, 12, 13      |
 | MySQL                | 5.7, 8                    | 5.7, 8                   |

airflow/__init__.py

@@ -36,7 +36,7 @@
 
 __version__ = version.version
 
-__all__ = ['__version__', 'login', 'DAG']
+__all__ = ['__version__', 'login', 'DAG', 'PY36', 'PY37', 'PY38', 'PY39']
 
 # Make `airflow` an namespace package, supporting installing
 # airflow.providers.* in different locations (i.e. one in site, and one in user
@@ -50,6 +50,7 @@
 PY36 = sys.version_info >= (3, 6)
 PY37 = sys.version_info >= (3, 7)
 PY38 = sys.version_info >= (3, 8)
+PY39 = sys.version_info >= (3, 9)
 
 
 def __getattr__(name):

airflow/provider.yaml.schema.json

@@ -28,6 +28,13 @@
         "type": "string"
       }
     },
+    "excluded-python-versions": {
+      "description": "List of python versions excluded for that provider",
+      "type": "array",
+      "items": {
+          "type": "string"
+      }
+    },
     "integrations": {
       "description": "List of integrations supported by the provider.",
       "type": "array",

airflow/providers/apache/hive/provider.yaml

@@ -31,6 +31,9 @@ versions:
 additional-dependencies:
   - apache-airflow>=2.1.0
 
+excluded-python-versions:
+  - "3.9"
+
 integrations:
   - integration-name: Apache Hive
     external-doc-url: https://hive.apache.org/

breeze

@@ -3507,7 +3507,7 @@ function breeze::run_breeze_command() {
         docker_engine_resources::check_all_resources
         if [[ ${PRODUCTION_IMAGE} == "true" ]]; then
             ${run_command} "${dc_run_file}" run --service-ports --rm airflow "${@}"
-            ${run_command} "${SCRIPTS_CI_DIR}/tools/ci_fix_ownership.sh"
+            ${run_command} "${SCRIPTS_CI_DIR}/tools/ci_fix_ownership.sh"  || true
         else
             ${run_command} "${dc_run_file}" run --service-ports --rm airflow "${@}"
         fi
@@ -3621,7 +3621,7 @@ function breeze::run_breeze_command() {
 # We have different versions of images depending on the python version used. We keep up with the
 # Latest patch-level changes in Python (this is done automatically during CI builds) so we have
 # To only take into account MAJOR and MINOR version of python. This variable keeps the major/minor
-# version of python in X.Y format (3.6, 3.7, 3.8 etc).
+# version of python in X.Y format (3.6, 3.7, 3.8, 3.9).
 #
 # In Breeze the precedence of setting the version is as follows:
 #      1. --python flag (if set, it will explicitly override it in the next step)

breeze-complete

@@ -23,7 +23,7 @@
 # by the BATS tests automatically during pre-commit and CI
 # Those cannot be made read-only as the breeze-complete must be re-sourceable
 
-_breeze_allowed_python_major_minor_versions="3.6 3.7 3.8"
+_breeze_allowed_python_major_minor_versions="3.6 3.7 3.8 3.9"
 _breeze_allowed_backends="sqlite mysql postgres"
 _breeze_allowed_integrations="cassandra kerberos mongo openldap pinot rabbitmq redis statsd trino all"
 _breeze_allowed_generate_constraints_modes="source-providers pypi-providers no-providers"

dev/provider_packages/PROVIDER_README_TEMPLATE.rst.jinja2

@@ -47,6 +47,8 @@ Installation
 You can install this package on top of an existing airflow 2.1+ installation via
 ``pip install {{PACKAGE_PIP_NAME}}``
 
+The package supports the following python versions: {{ ",".join(SUPPORTED_PYTHON_VERSIONS) }}
+
 {%- if PIP_REQUIREMENTS %}
 
 PIP requirements

dev/provider_packages/SETUP_TEMPLATE.py.jinja2

@@ -68,16 +68,16 @@ def do_setup():
             'Intended Audience :: Developers',
             'Intended Audience :: System Administrators',
             'License :: OSI Approved :: Apache Software License',
-            'Programming Language :: Python :: 3.6',
-            'Programming Language :: Python :: 3.7',
-            'Programming Language :: Python :: 3.8',
+{%- for python_version in SUPPORTED_PYTHON_VERSIONS %}
+            'Programming Language :: Python :: {{ python_version }}',
+{%- endfor %}
             'Topic :: System :: Monitoring',
         ],
         author='Apache Software Foundation',
         author_email='dev@airflow.apache.org',
         url='https://airflow.apache.org/',
         download_url='https://archive.apache.org/dist/airflow/{{ PROVIDERS_FOLDER }}',
-        python_requires='~=3.6',
+        python_requires='{{ PYTHON_REQUIRES }}',
         project_urls={
             'Documentation': 'https://airflow.apache.org/docs/{{ PACKAGE_PIP_NAME }}/{{RELEASE}}/',
             'Bug Tracker': 'https://github.com/apache/airflow/issues',

dev/provider_packages/prepare_provider_packages.py

@@ -51,6 +51,8 @@
 from rich.progress import Progress
 from rich.syntax import Syntax
 
+ALL_PYTHON_VERSIONS = ["3.6", "3.7", "3.8", "3.9"]
+
 try:
     from yaml import CSafeLoader as SafeLoader
 except ImportError:
@@ -208,6 +210,7 @@ class ProviderPackageDetails(NamedTuple):
     documentation_provider_package_path: str
     provider_description: str
     versions: List[str]
+    excluded_python_versions: List[str]
 
 
 ENTITY_NAMES = {
@@ -1445,6 +1448,7 @@ def get_provider_details(provider_package_id: str) -> ProviderPackageDetails:
         documentation_provider_package_path=get_documentation_package_path(provider_package_id),
         provider_description=provider_info['description'],
         versions=provider_info['versions'],
+        excluded_python_versions=provider_info.get("excluded-python-versions") or [],
     )
 
 
@@ -1483,6 +1487,12 @@ def get_provider_jinja_context(
     )
     with open(changelog_path) as changelog_file:
         changelog = changelog_file.read()
+    supported_python_versions = [
+        p for p in ALL_PYTHON_VERSIONS if p not in provider_details.excluded_python_versions
+    ]
+    python_requires = "~=3.6"
+    for p in provider_details.excluded_python_versions:
+        python_requires += f", !={p}"
     context: Dict[str, Any] = {
         "ENTITY_TYPES": list(EntityType),
         "README_FILE": "README.rst",
@@ -1517,6 +1527,8 @@ def get_provider_jinja_context(
             provider_details.documentation_provider_package_path,
         ),
         "CHANGELOG": changelog,
+        "SUPPORTED_PYTHON_VERSIONS": supported_python_versions,
+        "PYTHON_REQUIRES": python_requires,
     }
     return context
 

dev/retag_docker_images.py

@@ -47,7 +47,7 @@
 
 import click
 
-PYTHON_VERSIONS = ["3.6", "3.7", "3.8"]
+PYTHON_VERSIONS = ["3.6", "3.7", "3.8", "3.9"]
 
 DOCKERHUB_IMAGES = [
     "{prefix}:python{python_version}-{branch}",

scripts/ci/libraries/_build_images.sh

@@ -495,7 +495,6 @@ function build_images::rebuild_ci_image_if_needed() {
         push_pull_remove_images::pull_ci_images_if_needed
         return
     fi
-
     local needs_docker_build="false"
     md5sum::check_if_docker_build_is_needed
     build_images::get_local_build_cache_hash
@@ -526,7 +525,7 @@ function build_images::rebuild_ci_image_if_needed() {
                 local root_files_count
                 root_files_count=$(find "airflow" "tests" -user root | wc -l | xargs)
                 if [[ ${root_files_count} != "0" ]]; then
-                    ./scripts/ci/tools/ci_fix_ownership.sh
+                    ./scripts/ci/tools/ci_fix_ownership.sh || true
                 fi
             fi
             verbosity::print_info

scripts/ci/libraries/_initialization.sh

@@ -87,11 +87,11 @@ function initialization::initialize_base_variables() {
     export PRODUCTION_IMAGE="false"
 
     # All supported major/minor versions of python in all versions of Airflow
-    ALL_PYTHON_MAJOR_MINOR_VERSIONS+=("3.6" "3.7" "3.8")
+    ALL_PYTHON_MAJOR_MINOR_VERSIONS+=("3.6" "3.7" "3.8" "3.9")
     export ALL_PYTHON_MAJOR_MINOR_VERSIONS
 
     # Currently supported major/minor versions of python
-    CURRENT_PYTHON_MAJOR_MINOR_VERSIONS+=("3.6" "3.7" "3.8")
+    CURRENT_PYTHON_MAJOR_MINOR_VERSIONS+=("3.6" "3.7" "3.8" "3.9")
     export CURRENT_PYTHON_MAJOR_MINOR_VERSIONS
 
     # Currently supported versions of Postgres

scripts/ci/libraries/_push_pull_remove_images.sh

@@ -144,7 +144,7 @@ function push_pull_remove_images::pull_base_python_image() {
         push_pull_remove_images::pull_image_github_dockerhub "${AIRFLOW_PYTHON_BASE_IMAGE}" \
             "${GITHUB_REGISTRY_PYTHON_BASE_IMAGE}${python_tag_suffix}"
     else
-        docker_v pull "${AIRFLOW_PYTHON_BASE_IMAGE}"
+        docker_v pull "${AIRFLOW_PYTHON_BASE_IMAGE}" || true
     fi
 }
 
@@ -161,7 +161,7 @@ function push_pull_remove_images::pull_ci_images_if_needed() {
             push_pull_remove_images::pull_image_github_dockerhub "${AIRFLOW_CI_IMAGE}" \
                 "${GITHUB_REGISTRY_AIRFLOW_CI_IMAGE}:${GITHUB_REGISTRY_PULL_IMAGE_TAG}"
         else
-            push_pull_remove_images::pull_image_if_not_present_or_forced "${AIRFLOW_CI_IMAGE}"
+            push_pull_remove_images::pull_image_if_not_present_or_forced "${AIRFLOW_CI_IMAGE}" || true
         fi
     fi
 }