Rewrite changelog helper to Python and use markdown for changelog (pa…

…ntsbuild#11224)

Some changes:

* You now specify the `--prior` and `--new` versions as CLI args, e.g. `--prior 2.1.0 --new 2.2.0.dev0`. Before, it asked for the prior Git  tag as an `input()` once the script was running.
* We use markdown, which is simpler to work with.
    * We don't need to worry about compat with the v1 site.
* It says which file to modify.
* The version header is generated.
* We represent the date as `(Nov 20, 2020)`, rather than `(11/20/20)`. This avoids any confusion for non-US users.
* We no longer include related PRs and issues. It was a bit confusing what those meant, and the user should open the original PR to get more info anyways.

[ci skip-rust]
[ci skip-build-wheels]

build-support/bin/BUILD

@@ -17,6 +17,11 @@ pex_binary(
    sources = ['bootstrap_and_deploy_ci_pants_pex.py'],
 )
 
+pex_binary(
+  name="changelog",
+  sources=["changelog.py"],
+)
+
 pex_binary(
   name = 'check_banned_imports',
   sources = ['check_banned_imports.py'],
@@ -80,5 +85,3 @@ pex_binary(
   name = "packages",
   sources = ["packages.py"],
 )
-
-python_tests()

build-support/bin/changelog.py

@@ -0,0 +1,109 @@
+#!/usr/bin/env python3
+# Copyright 2020 Pants project contributors (see CONTRIBUTORS.md).
+# Licensed under the Apache License, Version 2.0 (see LICENSE).
+
+import argparse
+import datetime
+import re
+import subprocess
+from textwrap import dedent
+from typing import List
+
+
+def create_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description="Prepare the changelog for a release.")
+    parser.add_argument(
+        "--prior",
+        required=True,
+        type=str,
+        help="The version of the prior release, e.g. `2.0.0.dev0` or `2.0.0rc1`.",
+    )
+    parser.add_argument(
+        "--new",
+        required=True,
+        type=str,
+        help="The version for the new release, e.g. `2.0.0.dev0` or `2.0.0rc1`.",
+    )
+    return parser
+
+
+def relevant_shas(prior: str) -> List[str]:
+    prior_tag = f"release_{prior}"
+    return (
+        subprocess.run(
+            ["git", "log", "--format=format:%H", "HEAD", f"^{prior_tag}"],
+            check=True,
+            stdout=subprocess.PIPE,
+        )
+        .stdout.decode()
+        .splitlines()
+    )
+
+
+def prepare_sha(sha: str) -> str:
+    subject = (
+        subprocess.run(
+            ["git", "log", "-1", "--format=format:%s", sha],
+            check=True,
+            stdout=subprocess.PIPE,
+        )
+        .stdout.decode()
+        .strip()
+    )
+    pr_num_match = re.search(r"\(#(\d{4,5})\)\s*$", subject)
+    if not pr_num_match:
+        return f"* {subject}"
+    pr_num = pr_num_match.groups()[0]
+    pr_url = f"https://github.com/pantsbuild/pants/pull/{pr_num}"
+    subject_with_url = subject.replace(f"(#{pr_num})", f"([#{pr_num}]({pr_url}))")
+    return f"* {subject_with_url.capitalize()}"
+
+
+def instructions(new_version: str) -> str:
+    date = datetime.date.today().strftime("%b %d, %Y")
+    version_components = new_version.split(".", maxsplit=4)
+    major, minor = version_components[0], version_components[1]
+    return dedent(
+        f"""\
+        Copy the below headers into `src/python/notes/{major}.{minor}.x.md`. Then, put each commit
+        into the relevant category. You can tweak descriptions to be more descriptive or to fix
+        typos, and you can reorder based on relative importance to end users. Delete any unused
+        headers.
+
+        ---------------------------------------------------------------------
+
+        # {new_version} ({date})
+
+        ## New Features
+
+
+        ## User API Changes
+
+
+        ## Plugin API Changes
+
+
+        ## Bug fixes
+
+
+        ## Documentation
+
+
+        ## Internal only (Copy into the PR description, rather than the release notes)
+
+
+        --------------------------------------------------------------------
+
+        """
+    )
+
+
+def main() -> None:
+    args = create_parser().parse_args()
+    print(instructions(args.new))
+    entries = [prepare_sha(sha) for sha in relevant_shas(args.prior)]
+    print("\n\n".join(entries))
+
+
+if __name__ == "__main__":
+    main()