chore: Check whether version in PR is already released (#157)

We have a recurring issue that you often forget to bump the package
version number in a PR, and then when you merge it to `master`, the
package publication fails because the version you're trying to publish
already exists.

This adds a check to CI which detects these issues early and fails the
CI, so you can't merge those PRs without bumping the package version
number.

It gets skipped if the title of the PR is `docs: ...`, because such PRs
don't trigger package publication when merged.

Same as apify/apify-sdk-python#127.

Author: fnesveda

.github/workflows/check_version_availability.yaml

@@ -0,0 +1,25 @@
+name: Check package version availability
+
+on:
+  workflow_call:
+
+jobs:
+  check_version_availability:
+    name: Check version availability
+    runs-on: ubuntu-latest
+    if: (!startsWith(github.event.pull_request.title, 'docs:'))
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.8"
+
+      - name: Install dependencies
+        run: make install-dev
+
+      - name: Check version availability
+        run: make check-version-availability

.github/workflows/run_checks.yaml

@@ -4,6 +4,10 @@ on:
   pull_request:
 
 jobs:
+  check_version_availability:
+    name: Check version availability
+    uses: ./.github/workflows/check_version_availability.yaml
+
   lint_and_type_checks:
     name: Run lint and type checks
     uses: ./.github/workflows/lint_and_type_checks.yaml

Makefile

@@ -1,4 +1,4 @@
-.PHONY: clean install-dev build publish twine-check lint unit-tests integration-tests type-check check-code format check-async-docstrings fix-async-docstrings check-changelog-entry build-api-reference
+.PHONY: clean install-dev build publish twine-check lint unit-tests integration-tests type-check check-code format check-async-docstrings fix-async-docstrings check-version-availability check-changelog-entry build-api-reference
 
 # This is default for local testing, but GitHub workflows override it to a higher value in CI
 INTEGRATION_TESTS_CONCURRENCY = 1
@@ -44,6 +44,9 @@ check-async-docstrings:
 fix-async-docstrings:
 	python3 scripts/fix_async_docstrings.py
 
+check-version-availability:
+	python3 scripts/check_version_availability.py
+
 check-changelog-entry:
 	python3 scripts/check_version_in_changelog.py
 

scripts/check_version_availability.py

@@ -0,0 +1,13 @@
+#!/usr/bin/env python3
+from utils import get_current_package_version, get_published_package_versions
+
+# Checks whether the current package version number was not already used in a published release.
+if __name__ == '__main__':
+    current_version = get_current_package_version()
+
+    # Load the version numbers of the currently published versions from PyPI
+    published_versions = get_published_package_versions()
+
+    # We don't want to try to publish a version with the same version number as an already released stable version
+    if current_version in published_versions:
+        raise RuntimeError(f'The current version {current_version} was already released!')

scripts/update_version_for_prerelease.py

@@ -1,11 +1,9 @@
 #!/usr/bin/env python3
 
-import json
 import re
 import sys
-import urllib.request
 
-from utils import PACKAGE_NAME, get_current_package_version, set_current_package_version
+from utils import get_current_package_version, get_published_package_versions, set_current_package_version
 
 # Checks whether the current package version number was not already used in a published release,
 # and if not, modifies the package version number in pyproject.toml
@@ -32,16 +30,7 @@
         raise RuntimeError(f'The current version {current_version} does not match the proper semver format for stable releases (X.Y.Z)')
 
     # Load the version numbers of the currently published versions from PyPI
-    # If the URL returns 404, it means the package has no releases yet (which is okay in our case)
-    package_info_url = f'https://pypi.org/pypi/{PACKAGE_NAME}/json'
-    try:
-        conn = urllib.request.urlopen(package_info_url)
-        package_data = json.load(urllib.request.urlopen(package_info_url))
-        published_versions = list(package_data['releases'].keys())
-    except urllib.error.HTTPError as e:
-        if e.code != 404:
-            raise e
-        published_versions = []
+    published_versions = get_published_package_versions()
 
     # We don't want to publish a prerelease version with the same version number as an already released stable version
     if current_version in published_versions:

scripts/utils.py

@@ -1,5 +1,7 @@
+import json
 import pathlib
 import re
+import urllib.request
 
 PACKAGE_NAME = 'apify_client'
 REPO_ROOT = pathlib.Path(__file__).parent.resolve() / '..'
@@ -47,3 +49,17 @@ def sync_to_async_docstring(docstring: str) -> str:
     for (pattern, replacement) in substitutions:
         res = re.sub(pattern, replacement, res, flags=re.M)
     return res
+
+
+# Load the version numbers of the currently published versions from PyPI
+def get_published_package_versions() -> list:
+    package_info_url = f'https://pypi.org/pypi/{PACKAGE_NAME}/json'
+    try:
+        package_data = json.load(urllib.request.urlopen(package_info_url))
+        published_versions = list(package_data['releases'].keys())
+    # If the URL returns 404, it means the package has no releases yet (which is okay in our case)
+    except urllib.error.HTTPError as e:
+        if e.code != 404:
+            raise e
+        published_versions = []
+    return published_versions