You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@maven.apache.org by gn...@apache.org on 2026/10/14 07:14:35 UTC
(maven-gh-actions-shared) branch feature/pr-labels created (now 6ba24f7)
This is an automated email from the ASF dual-hosted git repository.
gnodet pushed a change to branch feature/pr-labels
in repository https://gitbox.apache.org/repos/asf/maven-gh-actions-shared.git
at 6ba24f7 feat: add pr-labels reusable workflow for automatic PR label assignment
This branch includes the following new commits:
new 6ba24f7 feat: add pr-labels reusable workflow for automatic PR label assignment
The 1 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "add" were already present in the repository and have only
been added to this reference.
(maven-gh-actions-shared) 01/01: feat: add pr-labels reusable workflow for automatic PR label assignment
Posted by gn...@apache.org.
This is an automated email from the ASF dual-hosted git repository.
gnodet pushed a commit to branch feature/pr-labels
in repository https://gitbox.apache.org/repos/asf/maven-gh-actions-shared.git
commit 6ba24f76311c1b4b6efa926e0699bf7eeb0de284
Author: Guillaume Nodet <gn...@gmail.com>
AuthorDate: Wed Oct 14 07:14:11 2026 +0000
feat: add pr-labels reusable workflow for automatic PR label assignment
Introduces a new reusable workflow that assigns labels to a pull request
when it is opened or reopened.
Labels are assigned in two passes:
1. Deterministic: a version label (mvn4/mvn40/mvn3) is derived from the
base branch name with no external dependencies.
2. LLM-assisted: a type label (bug/enhancement/maintenance/documentation)
and backport signals (backport-to-4.0.x, backport-to-3.10.x) are
inferred from the PR title and description via llm.apache.org.
This pass requires the LLMAO_TOKEN secret and is skipped gracefully
when the secret is absent.
The LLM is given a strict allowlist so it can never assign arbitrary
labels. The LLMAO_TOKEN secret is declared optional so repos that have
not yet set it still benefit from the deterministic version labeling.
Usage in a consuming repo:
on:
pull_request_target:
types: [opened, reopened]
jobs:
pr-labels:
uses: apache/maven-gh-actions-shared/.github/workflows/pr-labels.yml@v5
secrets:
LLMAO_TOKEN: ${{ secrets.LLMAO_TOKEN }}
---
.github/workflows/pr-labels.yml | 196 ++++++++++++++++++++++++++++++++++++++++
1 file changed, 196 insertions(+)
diff --git a/.github/workflows/pr-labels.yml b/.github/workflows/pr-labels.yml
new file mode 100644
index 0000000..550bdb6
--- /dev/null
+++ b/.github/workflows/pr-labels.yml
@@ -0,0 +1,196 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements. See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership. The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License. You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied. See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Reusable workflow that assigns labels to a pull request on open/reopen.
+# Labels are assigned in two passes:
+# 1. Deterministic: version label (mvn3/mvn40/mvn4) from the base branch name.
+# 2. LLM-assisted: type label (bug/enhancement/maintenance/documentation/…)
+# and backport signals (backport-to-4.0.x / backport-to-3.10.x) inferred
+# from the PR title and description via llm.apache.org (OpenAI-compatible).
+# Requires the LLMAO_TOKEN secret to be set; skipped gracefully if absent.
+
+name: PR Labels
+
+on:
+ workflow_call:
+ inputs:
+ llmao-model:
+ description: |
+ Model ID to use on llm.apache.org.
+ Defaults to gpt-4o-mini which is well suited for structured classification.
+ required: false
+ type: string
+ default: 'gpt-4o-mini'
+ secrets:
+ LLMAO_TOKEN:
+ description: |
+ Personal access token for llm.apache.org (https://llm.apache.org).
+ Create one at llm.apache.org after signing in with your ASF account.
+ If not provided, only the deterministic version label is applied.
+ required: false
+
+# All permissions are declared per-job.
+permissions: {}
+
+jobs:
+ assign-labels:
+ name: Assign labels
+ runs-on: ubuntu-slim
+ permissions:
+ pull-requests: write
+ issues: write # label writes go through the issues API
+ # Skip bots (Dependabot, Renovate, …)
+ if: github.event.pull_request.user.type != 'Bot'
+ # Prevent races when opened + synchronize fire together
+ concurrency:
+ group: pr-labels-${{ github.repository }}-${{ github.event.pull_request.number }}
+ cancel-in-progress: false
+ steps:
+ - name: Assign labels
+ uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
+ env:
+ LLMAO_TOKEN: ${{ secrets.LLMAO_TOKEN }}
+ LLMAO_MODEL: ${{ inputs.llmao-model }}
+ with:
+ script: |
+ const { owner, repo } = context.repo;
+ const pr = context.payload.pull_request;
+ const prNumber = pr.number;
+ const base = pr.base.ref;
+
+ // ── 1. Deterministic: version label from base branch ──────────────────
+ const VERSION_LABELS = {
+ mvn4: /^(master|main|feature\/.+)$/,
+ mvn40: /^maven-4\.0\.x$/,
+ mvn3: /^maven-3\.\d+\.x$/,
+ };
+
+ const existingLabels = pr.labels.map(l => l.name);
+ core.info(`Existing labels: ${existingLabels.join(', ') || '(none)'}`);
+
+ const labelsToAdd = new Set();
+
+ for (const [label, pattern] of Object.entries(VERSION_LABELS)) {
+ if (pattern.test(base) && !existingLabels.includes(label)) {
+ labelsToAdd.add(label);
+ core.info(`Version label: ${label} (base branch: ${base})`);
+ break; // exactly one version label
+ }
+ }
+
+ // ── 2. LLM-assisted: type + backport labels ───────────────────────────
+ const token = process.env.LLMAO_TOKEN;
+ if (!token) {
+ core.info('LLMAO_TOKEN not set — skipping LLM classification.');
+ } else {
+ const model = process.env.LLMAO_MODEL || 'gpt-4o-mini';
+
+ // Fetch the full PR (payload body may be truncated for large descriptions)
+ const { data: fullPr } = await github.rest.pulls.get({
+ owner, repo, pull_number: prNumber,
+ });
+
+ const prompt = [
+ 'You are a Maven project maintainer. Classify the following pull request and assign labels.',
+ '',
+ 'Available labels and when to use them:',
+ ' bug - fixes incorrect or broken behavior',
+ ' enhancement - adds new functionality or meaningfully improves existing behavior',
+ ' maintenance - refactoring, cleanup, test additions, comment removal, dependency bumps',
+ ' documentation - pure documentation or comment changes',
+ ' breaking - introduces a breaking API or behavior change',
+ ' backport-to-4.0.x - this bug fix on master/main should also be applied to the maven-4.0.x branch',
+ ' backport-to-3.10.x - this bug fix on master/main should also be applied to the maven-3.10.x branch',
+ '',
+ 'Rules:',
+ '- Pick exactly ONE type label (bug / enhancement / maintenance / documentation).',
+ '- Add "breaking" only if the change breaks existing API or behavior.',
+ '- Add backport labels ONLY for bug fixes whose base branch is master or main,',
+ ' and only when the buggy code plausibly exists in the older branch too.',
+ '- Respond with ONLY a JSON array of label strings, no explanation.',
+ ' Example: ["bug","backport-to-4.0.x"]',
+ '',
+ `PR title: ${fullPr.title}`,
+ `Base branch: ${base}`,
+ `Body:\n${(fullPr.body || '').slice(0, 3000)}`,
+ ].join('\n');
+
+ try {
+ const resp = await fetch('https://llm.apache.org/v1/chat/completions', {
+ method: 'POST',
+ headers: {
+ 'Authorization': `Bearer ${token}`,
+ 'Content-Type': 'application/json',
+ },
+ body: JSON.stringify({
+ model,
+ messages: [{ role: 'user', content: prompt }],
+ max_tokens: 80,
+ temperature: 0,
+ }),
+ });
+
+ if (!resp.ok) {
+ core.warning(`llm.apache.org returned HTTP ${resp.status} — skipping LLM labels.`);
+ } else {
+ const data = await resp.json();
+ const text = data.choices?.[0]?.message?.content?.trim() ?? '';
+ core.info(`LLM raw response: ${text}`);
+
+ // Strip markdown code fences if the model wraps the JSON
+ const cleaned = text.replace(/^```[a-z]*\n?/, '').replace(/\n?```$/, '').trim();
+
+ let suggested;
+ try {
+ suggested = JSON.parse(cleaned);
+ } catch (_) {
+ core.warning(`Could not parse LLM response as JSON: ${text}`);
+ suggested = [];
+ }
+
+ // Allowlist — never let the LLM assign arbitrary labels
+ const ALLOWED = new Set([
+ 'bug', 'enhancement', 'maintenance', 'documentation', 'breaking',
+ 'backport-to-4.0.x', 'backport-to-3.10.x',
+ ]);
+
+ for (const label of suggested) {
+ if (ALLOWED.has(label) && !existingLabels.includes(label)) {
+ labelsToAdd.add(label);
+ } else if (!ALLOWED.has(label)) {
+ core.warning(`Ignoring label not on allowlist: ${label}`);
+ }
+ }
+ }
+ } catch (err) {
+ core.warning(`LLM call failed: ${err.message} — skipping LLM labels.`);
+ }
+ }
+
+ // ── 3. Apply ──────────────────────────────────────────────────────────
+ if (labelsToAdd.size === 0) {
+ core.info('No new labels to add.');
+ return;
+ }
+
+ const labels = [...labelsToAdd];
+ core.info(`Applying labels: ${labels.join(', ')}`);
+ await github.rest.issues.addLabels({
+ owner, repo,
+ issue_number: prNumber,
+ labels,
+ });