Adds automated generation of provider issue to track test progress (#…

…16419) Allows to automatically generate draft of the issue which can be used to track progress of testing released providers.
apache · Jun 20, 2021 · f40d555 · f40d555
1 parent 1e56420
commit f40d555
Show file tree

Hide file tree

Showing 11 changed files with 233 additions and 13 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -320,8 +320,6 @@ jobs:
         uses: actions/checkout@v2
         with:
           persist-credentials: false
-          # Needs to fetch all history to be able to verify changelog generation
-          fetch-depth: 0
       - name: "Setup python"
         uses: actions/setup-python@v2
         with:
@@ -522,6 +520,7 @@ ${{ hashFiles('.pre-commit-config.yaml') }}"
       VERSION_SUFFIX_FOR_PYPI: ".dev0"
       GITHUB_REGISTRY: ${{ needs.ci-images.outputs.githubRegistry }}
       NON_INTERACTIVE: "true"
+      GENERATE_PROVIDERS_ISSUE: "true"
     if: needs.build-info.outputs.image-build == 'true' && needs.build-info.outputs.default-branch == 'main'
     steps:
       - name: "Checkout ${{ github.ref }} ( ${{ github.sha }} )"
@@ -569,6 +568,8 @@ ${{ hashFiles('.pre-commit-config.yaml') }}"
       PYTHON_MAJOR_MINOR_VERSION: ${{needs.build-info.outputs.defaultPythonVersion}}
       VERSION_SUFFIX_FOR_PYPI: ".dev0"
       GITHUB_REGISTRY: ${{ needs.ci-images.outputs.githubRegistry }}
+      NON_INTERACTIVE: "true"
+      GENERATE_PROVIDERS_ISSUE: "true"
     if: needs.build-info.outputs.image-build == 'true' && needs.build-info.outputs.default-branch == 'main'
     steps:
       - name: "Checkout ${{ github.ref }} ( ${{ github.sha }} )"

diff --git a/BREEZE.rst b/BREEZE.rst
@@ -2141,6 +2141,10 @@ This is the current syntax for  `./breeze <./breeze>`_:
 
           Runs the command in non-interactive mode.
 
+  --generate-providers-issue
+
+          Generate providers issue that should be created.
+
   -v, --verbose
           Show verbose information about executed docker, kind, kubectl, helm commands. Useful for
           debugging - when you run breeze with --verbose flags you will be able to see the commands

diff --git a/breeze b/breeze
@@ -1260,6 +1260,13 @@ function breeze::parse_arguments() {
             echo
             shift
             ;;
+        --generate-providers-issue)
+            export GENERATE_PROVIDERS_ISSUE="true"
+            export FORWARD_CREDENTIALS="true"
+            echo "Generating providers issue (includes forwarding credentials)"
+            echo
+            shift
+            ;;
         --installation-method)
             export AIRFLOW_INSTALLATION_METHOD="${2}"
             echo "Airflow installation method: ${AIRFLOW_INSTALLATION_METHOD}"
@@ -1866,7 +1873,7 @@ ${CMDNAME} prepare-provider-documentation [FLAGS] [PACKAGE_ID ...]
 Flags:
 $(breeze::flag_version_suffix)
 $(breeze::flag_packages)
-$(breeze::flag_non_interactive)
+$(breeze::prepare_providers_documentation)
 $(breeze::flag_verbosity)
 "
     readonly DETAILED_USAGE_PREPARE_PROVIDER_DOCUMENTATION
@@ -2596,17 +2603,21 @@ ${FORMATTED_PACKAGE_FORMATS}
 
 #######################################################################################################
 #
-# Prints flags for non-interactive run
+# Prints flags for prepare-providers-documentation
 #
 # Outputs:
 #    Flag information.
 #######################################################################################################
-function breeze::flag_non_interactive() {
+function breeze::prepare_providers_documentation() {
     echo "
 --non-interactive
 
         Runs the command in non-interactive mode.
 
+--generate-providers-issue
+
+        Generate providers issue that should be created.
+
 "
 }
 

diff --git a/breeze-complete b/breeze-complete
@@ -189,7 +189,7 @@ additional-extras: additional-python-deps: disable-pypi-when-building skip-insta
 dev-apt-deps: additional-dev-apt-deps: dev-apt-command: additional-dev-apt-command: additional-dev-apt-env:
 runtime-apt-deps: additional-runtime-apt-deps: runtime-apt-command: additional-runtime-apt-command: additional-runtime-apt-env:
 load-default-connections load-example-dags
-use-packages-from-dist no-rbac-ui package-format: upgrade-to-newer-dependencies installation-method: continue-on-pip-check-failure non-interactive
+use-packages-from-dist no-rbac-ui package-format: upgrade-to-newer-dependencies installation-method: continue-on-pip-check-failure non-interactive generate-providers-issue
 use-airflow-version:
 cleanup-docker-context-files
 test-type: preserve-volumes dry-run-docker

diff --git a/dev/README_RELEASE_PROVIDER_PACKAGES.md b/dev/README_RELEASE_PROVIDER_PACKAGES.md
@@ -31,6 +31,7 @@
   - [Publish the Regular convenience package to PyPI](#publish-the-regular-convenience-package-to-pypi)
   - [Add tags in git](#add-tags-in-git)
   - [Prepare documentation](#prepare-documentation)
+  - [Prepare issue in GitHub to keep status of testing](#prepare-issue-in-github-to-keep-status-of-testing)
   - [Prepare voting email for Providers release candidate](#prepare-voting-email-for-providers-release-candidate)
   - [Verify the release by PMC members](#verify-the-release-by-pmc-members)
   - [Verify by Contributors](#verify-by-contributors)
@@ -109,6 +110,17 @@ are updated, run it in non-interactive mode:
 ./breeze --non-interactive prepare-provider-documentation [packages]
 ```
 
+When you run the command and documentation generation is successful you will get a command that you can run to
+create GitHub issue where you will be tracking status of tests for the providers you release.
+
+You can also trigger automated execution of the issue by running:
+
+```shell script
+./breeze --non-interactive --generate-providers-issue prepare-provider-documentation [packages]
+```
+
+Once you release packages, you should create the issue with the content specified and link to it in
+the email sent to the devlist.
 
 ## Build provider packages for SVN apache upload
 
@@ -338,6 +350,11 @@ git commit -m "Add documentation for packages - $(date "+%Y-%m-%d%n")"
 git push --set-upstream origin "${branch}"
 ```
 
+## Prepare issue in GitHub to keep status of testing
+
+Create GitHub issue with the content generated via prepare-provider-documentation or manual
+execution of the script above. You will use link to that issue in the next step.
+
 ## Prepare voting email for Providers release candidate
 
 Make sure the packages are in https://dist.apache.org/repos/dist/dev/airflow/providers/
@@ -350,7 +367,7 @@ subject:
 
 ```shell script
 cat <<EOF
-[VOTE] Airflow Providers - release prepared $(date "+%Y-%m-%d%n")
+[VOTE] Airflow Providers prepared on $(date "+%B %d, %Y")
 EOF
 ```
 
@@ -363,7 +380,7 @@ which will last for 72 hours - which means that it will end on $(date -d '+3 day
 
 Consider this my (binding) +1.
 
-<PASTE ANY HIGH-LEVEL DESCRIPTION OF THE CHANGES HERE!>
+<ADD ANY HIGH-LEVEL DESCRIPTION OF THE CHANGES HERE!>
 
 Airflow Providers are available at:
 https://dist.apache.org/repos/dist/dev/airflow/providers/
@@ -399,6 +416,9 @@ Please note that the version number excludes the 'rcX' string.
 This will allow us to rename the artifact without modifying
 the artifact checksums when we actually release.
 
+The status of testing the providers by the community is kept here:
+<TODO COPY LINK TO THE ISSUE CREATED>
+
 You can find packages as well as detailed changelog following the below links:
 
 <PASTE TWINE UPLOAD LINKS HERE. SORT THEM BEFORE!>

diff --git a/dev/provider_packages/PROVIDER_ISSUE_TEMPLATE.md.jinja2 b/dev/provider_packages/PROVIDER_ISSUE_TEMPLATE.md.jinja2
@@ -0,0 +1,35 @@
+I have a kind request for all the contributors to the latest provider packages release.
+Could you help us to test the RC versions of the providers and let us know in the comment,
+if the issue is addressed there.
+
+## Providers that need testing
+
+Those are providers that require testing as there were some substantial changes introduced:
+
+{% for provider_id, provider_pr_info in interesting_providers.items()  %}
+### Provider [{{ provider_id }}: {{ provider_pr_info.provider_details.versions[0] }}{{ suffix }}](https://pypi.org/project/{{ provider_pr_info.provider_details.pypi_package_name }}/{{ provider_pr_info.provider_details.versions[0] }}{{ suffix }})
+{%- for pr in provider_pr_info.pr_list %}
+   - [ ] [{{ pr.title }} (#{{ pr.number }})]({{ pr.html_url }}): @{{ pr.user.login }}
+{%- endfor %}
+{%- endfor %}
+
+## Providers that do not need testing
+
+Those are providers that were either doc-only or had changes that do not require testing.
+
+{% for provider_id, provider_pr_info in non_interesting_providers.items()  %}
+* Provider [{{ provider_id }}: {{ provider_pr_info.provider_details.versions[0] }}{{ suffix }}](https://pypi.org/project/{{ provider_pr_info.provider_details.pypi_package_name }}/{{ provider_pr_info.provider_details.versions[0] }}{{ suffix }})
+{%- endfor %}
+
+<!--
+
+NOTE TO RELEASE MANAGER:
+
+Please move here the providers that have doc-only changes or for which changes are trivial and
+you could asses that they are OK. In case
+
+The providers are automatically installed on Airflow 2.1 and latest `main` during the CI, so we know they
+are installable. Also all classes within the providers are imported during the CI run so we know all
+providers can be imported.
+
+-->
diff --git a/dev/provider_packages/prepare_provider_packages.py b/dev/provider_packages/prepare_provider_packages.py
@@ -44,9 +44,11 @@
 import click
 import jsonschema
 import yaml
+from github import Github, PullRequest, UnknownObjectException
 from packaging.version import Version
 from rich import print
 from rich.console import Console
+from rich.progress import Progress
 from rich.syntax import Syntax
 
 try:
@@ -155,6 +157,7 @@ def cli():
 )
 argument_package_id = click.argument('package_id')
 argument_changelog_files = click.argument('changelog_files', nargs=-1)
+argument_package_ids = click.argument('package_ids', nargs=-1)
 
 
 @contextmanager
@@ -200,6 +203,7 @@ class VerifiedEntities(NamedTuple):
 class ProviderPackageDetails(NamedTuple):
     provider_package_id: str
     full_package_name: str
+    pypi_package_name: str
     source_provider_package_path: str
     documentation_provider_package_path: str
     provider_description: str
@@ -1436,6 +1440,7 @@ def get_provider_details(provider_package_id: str) -> ProviderPackageDetails:
     return ProviderPackageDetails(
         provider_package_id=provider_package_id,
         full_package_name=f"airflow.providers.{provider_package_id}",
+        pypi_package_name=f"apache-airflow-providers-{provider_package_id.replace('.', '-')}",
         source_provider_package_path=get_source_package_path(provider_package_id),
         documentation_provider_package_path=get_documentation_package_path(provider_package_id),
         provider_description=provider_info['description'],
@@ -1797,7 +1802,7 @@ def get_all_providers() -> List[str]:
     return list(PROVIDERS_REQUIREMENTS.keys())
 
 
-def verify_provider_package(provider_package_id: str) -> None:
+def verify_provider_package(provider_package_id: str) -> str:
     """
     Verifies if the provider package is good.
     :param provider_package_id: package id to verify
@@ -2246,13 +2251,133 @@ def get_package_from_changelog(changelog_path: str):
 @option_git_update
 @option_verbose
 def update_changelogs(changelog_files: List[str], git_update: bool, verbose: bool):
+    """Updates changelogs for multiple packages."""
     if git_update:
         make_sure_remote_apache_exists_and_fetch(git_update, verbose)
     for changelog_file in changelog_files:
         package_id = get_package_from_changelog(changelog_file)
         _update_changelog(package_id=package_id, verbose=verbose)
 
 
+def get_prs_for_package(package_id: str) -> List[int]:
+    pr_matcher = re.compile(r".*\(#([0-9]*)\)``$")
+    verify_provider_package(package_id)
+    changelog_path = verify_changelog_exists(package_id)
+    provider_details = get_provider_details(package_id)
+    current_release_version = provider_details.versions[0]
+    prs = []
+    with open(changelog_path) as changelog_file:
+        changelog_lines = changelog_file.readlines()
+        extract_prs = False
+        skip_line = False
+        for line in changelog_lines:
+            if skip_line:
+                # Skip first "....." header
+                skip_line = False
+                continue
+            if line.strip() == current_release_version:
+                extract_prs = True
+                skip_line = True
+                continue
+            if extract_prs:
+                if all(c == '.' for c in line):
+                    # Header for next version reached
+                    break
+                if line.startswith('.. Below changes are excluded from the changelog'):
+                    # The reminder of PRs is not important skipping it
+                    break
+                match_result = pr_matcher.match(line.strip())
+                if match_result:
+                    prs.append(int(match_result.group(1)))
+    return prs
+
+
+class ProviderPRInfo(NamedTuple):
+    provider_details: ProviderPackageDetails
+    pr_list: List[PullRequest.PullRequest]
+
+
+@cli.command()
+@click.option('--github-token', envvar='GITHUB_TOKEN')
+@click.option('--suffix', default='rc1')
+@click.option('--excluded-pr-list', type=str, help="Coma-separated list of PRs to exclude from the issue.")
+@argument_package_ids
+def generate_issue_content(package_ids: List[str], github_token: str, suffix: str, excluded_pr_list: str):
+    if not package_ids:
+        package_ids = get_all_providers()
+    """Generates content for issue to test the release."""
+    with with_group("Generates GitHub issue content with people who can test it"):
+        if excluded_pr_list:
+            excluded_prs = [int(pr) for pr in excluded_pr_list.split(",")]
+        else:
+            excluded_prs = []
+        console = Console(width=200, color_system="standard")
+        all_prs: Set[int] = set()
+        provider_prs: Dict[str, List[int]] = {}
+        for package_id in package_ids:
+            console.print(f"Extracting PRs for provider {package_id}")
+            prs = get_prs_for_package(package_id)
+            provider_prs[package_id] = list(filter(lambda pr: pr not in excluded_prs, prs))
+            all_prs.update(provider_prs[package_id])
+        g = Github(github_token)
+        repo = g.get_repo("apache/airflow")
+        pull_requests: Dict[int, PullRequest.PullRequest] = {}
+        with Progress(console=console) as progress:
+            task = progress.add_task(f"Retrieving {len(all_prs)} PRs ", total=len(all_prs))
+            pr_list = list(all_prs)
+            for i in range(len(pr_list)):
+                pr_number = pr_list[i]
+                progress.console.print(
+                    f"Retrieving PR#{pr_number}: " f"https://github.com/apache/airflow/pull/{pr_number}"
+                )
+                try:
+                    pull_requests[pr_number] = repo.get_pull(pr_number)
+                except UnknownObjectException:
+                    # Fallback to issue if PR not found
+                    try:
+                        pull_requests[pr_number] = repo.get_issue(pr_number)  # noqa (same fields as PR)
+                    except UnknownObjectException:
+                        console.print(f"[red]The PR #{pr_number} could not be found[/]")
+                progress.advance(task)
+        interesting_providers: Dict[str, ProviderPRInfo] = {}
+        non_interesting_providers: Dict[str, ProviderPRInfo] = {}
+        for package_id in package_ids:
+            pull_request_list = [pull_requests[pr] for pr in provider_prs[package_id] if pr in pull_requests]
+            provider_details = get_provider_details(package_id)
+            if pull_request_list:
+                interesting_providers[package_id] = ProviderPRInfo(provider_details, pull_request_list)
+            else:
+                non_interesting_providers[package_id] = ProviderPRInfo(provider_details, pull_request_list)
+        context = {
+            'interesting_providers': interesting_providers,
+            'date': datetime.now(),
+            'suffix': suffix,
+            'non_interesting_providers': non_interesting_providers,
+        }
+        issue_content = render_template(template_name="PROVIDER_ISSUE", context=context, extension=".md")
+        console.print()
+        console.print(
+            "[green]Below you can find the issue content that you can use "
+            "to ask contributor to test providers![/]"
+        )
+        console.print()
+        console.print()
+        console.print(
+            "Issue title: [yellow]Status of testing Providers that were "
+            f"prepared on { datetime.now().strftime('%B %d, %Y') }[/]"
+        )
+        console.print()
+        syntax = Syntax(issue_content, "markdown", theme="ansi_dark")
+        console.print(syntax)
+        console.print()
+        users: Set[str] = set()
+        for provider_info in interesting_providers.values():
+            for pr in provider_info.pr_list:
+                users.add("@" + pr.user.login)
+        console.print("All users involved in the PRs:")
+        console.print(" ".join(users))
+
+
 if __name__ == "__main__":
     # The cli exit code is:
     #   * 0 in case of success

diff --git a/scripts/ci/docker-compose/forward-credentials.yml b/scripts/ci/docker-compose/forward-credentials.yml
@@ -24,6 +24,8 @@ services:
     # Everything you install in Docker
     # If you add it here - also add it to "in_container_fix_ownership" method in
     # the _in_container_utils.sh file to make it friendly for Linux users
+    environment:
+      - GITHUB_TOKEN
     volumes:
       - ${HOME}/.aws:/root/.aws:cached
       - ${HOME}/.azure:/root/.azure:cached

diff --git a/scripts/ci/libraries/_runs.sh b/scripts/ci/libraries/_runs.sh
@@ -73,6 +73,8 @@ function runs::run_prepare_provider_documentation() {
         "${term_flag}" \
         -v "${AIRFLOW_SOURCES}:/opt/airflow" \
         -e "NON_INTERACTIVE" \
+        -e "GENERATE_PROVIDERS_ISSUE" \
+        -e "GITHUB_TOKEN" \
         "${AIRFLOW_CI_IMAGE}" \
         "--" "/opt/airflow/scripts/in_container/run_prepare_provider_documentation.sh" "${@}"
 }