mirror of
https://github.com/OpenHands/OpenHands.git
synced 2026-10-07 16:19:05 +08:00
Add issue duplicate automation workflow (#14444)
Co-authored-by: Engel Nyst <engel.nyst@gmail.com> Co-authored-by: openhands <openhands@all-hands.dev>
This commit is contained in:
🐾 smolpaws
co-authored by
Engel Nyst
openhands
Engel Nyst
openhands
parent
68083f275b
commit
e3d9abfd01
@@ -0,0 +1,468 @@
|
||||
---
|
||||
name: Issue Duplicate Check via OpenHands Cloud
|
||||
|
||||
on:
|
||||
issues:
|
||||
types: [opened]
|
||||
schedule:
|
||||
- cron: 0 9 * * *
|
||||
workflow_dispatch:
|
||||
inputs:
|
||||
mode:
|
||||
description: Which workflow path to run
|
||||
required: true
|
||||
type: choice
|
||||
options:
|
||||
- smoke-clone
|
||||
- issue-check
|
||||
- auto-close
|
||||
default: smoke-clone
|
||||
issue_number:
|
||||
description: Existing issue number to analyze when mode is issue-check
|
||||
required: false
|
||||
type: number
|
||||
close_after_days:
|
||||
description: Days to wait before auto-closing duplicate candidates in auto-close mode
|
||||
required: false
|
||||
type: number
|
||||
default: 3
|
||||
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
issues: write
|
||||
|
||||
jobs:
|
||||
smoke-clone:
|
||||
if: github.event_name == 'workflow_dispatch' && inputs.mode == 'smoke-clone'
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Checkout repository
|
||||
uses: actions/checkout@v6
|
||||
|
||||
- name: Clone repository
|
||||
run: |
|
||||
git clone --depth 1 "https://github.com/${{ github.repository }}.git" /tmp/repo-clone
|
||||
echo "Repository HEAD: $(git -C /tmp/repo-clone rev-parse --short HEAD)"
|
||||
|
||||
- name: Summarize smoke test
|
||||
run: |
|
||||
{
|
||||
echo "## Smoke clone completed"
|
||||
echo
|
||||
echo "- Repository cloned to /tmp/repo-clone"
|
||||
} >> "$GITHUB_STEP_SUMMARY"
|
||||
|
||||
issue-duplicate-check:
|
||||
if: |
|
||||
github.event_name == 'issues' ||
|
||||
(github.event_name == 'workflow_dispatch' && inputs.mode == 'issue-check' && inputs.issue_number != null)
|
||||
runs-on: ubuntu-latest
|
||||
timeout-minutes: 35
|
||||
concurrency:
|
||||
group: issue-duplicate-check-${{ github.repository }}-${{ github.event.issue.number || inputs.issue_number }}
|
||||
cancel-in-progress: false
|
||||
steps:
|
||||
- name: Checkout repository
|
||||
uses: actions/checkout@v6
|
||||
|
||||
- name: Set up Python
|
||||
uses: actions/setup-python@v6
|
||||
with:
|
||||
python-version: '3.13'
|
||||
|
||||
- name: Validate duplicate check inputs
|
||||
env:
|
||||
OPENHANDS_API_KEY: ${{ secrets.OPENHANDS_API_KEY }}
|
||||
ISSUE_NUMBER: ${{ github.event.issue.number || inputs.issue_number }}
|
||||
run: |
|
||||
if [ -z "$OPENHANDS_API_KEY" ]; then
|
||||
echo "Error: OPENHANDS_API_KEY secret is required"
|
||||
exit 1
|
||||
fi
|
||||
if [ -z "$ISSUE_NUMBER" ]; then
|
||||
echo "Error: ISSUE_NUMBER is required"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
- name: Run OpenHands duplicate check conversation
|
||||
id: run_check
|
||||
env:
|
||||
OPENHANDS_API_KEY: ${{ secrets.OPENHANDS_API_KEY }}
|
||||
GITHUB_TOKEN: ${{ secrets.OPENHANDS_BOT_GITHUB_PAT_PUBLIC || github.token }}
|
||||
ISSUE_NUMBER: ${{ github.event.issue.number || inputs.issue_number }}
|
||||
OUTPUT_PATH: ${{ runner.temp }}/issue-duplicate-check-result.json
|
||||
run: |
|
||||
python scripts/issue_duplicate_check_openhands.py \
|
||||
--repository "${{ github.repository }}" \
|
||||
--issue-number "$ISSUE_NUMBER" \
|
||||
--output "$OUTPUT_PATH"
|
||||
test -f "$OUTPUT_PATH" || {
|
||||
echo "Error: Output file not created"
|
||||
exit 1
|
||||
}
|
||||
echo "result_path=$OUTPUT_PATH" >> "$GITHUB_OUTPUT"
|
||||
|
||||
- name: Parse duplicate check result
|
||||
id: parsed_result
|
||||
env:
|
||||
RESULT_PATH: ${{ steps.run_check.outputs.result_path }}
|
||||
run: |
|
||||
python - <<'PY'
|
||||
import json
|
||||
import os
|
||||
import sys
|
||||
from pathlib import Path
|
||||
|
||||
try:
|
||||
result = json.loads(Path(os.environ['RESULT_PATH']).read_text())
|
||||
except (FileNotFoundError, json.JSONDecodeError) as exc:
|
||||
print(
|
||||
f"Error: Failed to read duplicate check result: {exc}",
|
||||
file=sys.stderr,
|
||||
)
|
||||
raise SystemExit(1) from exc
|
||||
output_path = Path(os.environ['GITHUB_OUTPUT'])
|
||||
summary_path = Path(os.environ['GITHUB_STEP_SUMMARY'])
|
||||
|
||||
def write_multiline(name: str, value: str) -> None:
|
||||
delimiter = f"EOF_{os.urandom(8).hex()}"
|
||||
with output_path.open('a', encoding='utf-8') as fh:
|
||||
fh.write(f"{name}<<{delimiter}\n{value}\n{delimiter}\n")
|
||||
|
||||
canonical_issue_number = result.get('canonical_issue_number')
|
||||
with output_path.open('a', encoding='utf-8') as fh:
|
||||
fh.write(f"should_comment={'true' if result.get('should_comment') else 'false'}\n")
|
||||
fh.write(f"is_duplicate={'true' if result.get('is_duplicate') else 'false'}\n")
|
||||
fh.write(
|
||||
f"auto_close_candidate={'true' if result.get('auto_close_candidate') else 'false'}\n"
|
||||
)
|
||||
fh.write(f"confidence={result.get('confidence', '')}\n")
|
||||
fh.write(f"classification={result.get('classification', '')}\n")
|
||||
fh.write(
|
||||
f"canonical_issue_number={canonical_issue_number if canonical_issue_number is not None else ''}\n"
|
||||
)
|
||||
fh.write(f"conversation_url={result.get('conversation_url', '')}\n")
|
||||
fh.write(f"app_conversation_id={result.get('app_conversation_id', '')}\n")
|
||||
|
||||
write_multiline('summary', str(result.get('summary', '')).strip())
|
||||
write_multiline(
|
||||
'candidate_issues_json',
|
||||
json.dumps(result.get('candidate_issues', []), ensure_ascii=False),
|
||||
)
|
||||
|
||||
candidate_lines = []
|
||||
for candidate in result.get('candidate_issues', []):
|
||||
candidate_lines.append(
|
||||
f"- #{candidate.get('number')}: {candidate.get('title')} ({candidate.get('url')}) — {candidate.get('similarity_reason', '')}"
|
||||
)
|
||||
|
||||
summary_path.write_text(
|
||||
"\n".join(
|
||||
[
|
||||
"## Duplicate check result",
|
||||
"",
|
||||
f"- Repository: {result.get('repository')}",
|
||||
f"- Issue: #{result.get('issue_number')}",
|
||||
f"- Should comment: {result.get('should_comment')}",
|
||||
f"- Exact duplicate: {result.get('is_duplicate')}",
|
||||
f"- Auto-close candidate: {result.get('auto_close_candidate')}",
|
||||
f"- Classification: {result.get('classification')}",
|
||||
f"- Confidence: {result.get('confidence')}",
|
||||
f"- Canonical issue: {canonical_issue_number}",
|
||||
f"- Conversation: {result.get('conversation_url')}",
|
||||
"",
|
||||
"### Summary",
|
||||
result.get('summary', ''),
|
||||
"",
|
||||
"### Candidate issues",
|
||||
*(candidate_lines or ["- None"]),
|
||||
]
|
||||
)
|
||||
+ "\n",
|
||||
encoding='utf-8',
|
||||
)
|
||||
PY
|
||||
|
||||
- name: Post duplicate overlap notice
|
||||
if: steps.parsed_result.outputs.should_comment == 'true'
|
||||
uses: actions/github-script@v7
|
||||
env:
|
||||
ISSUE_NUMBER: ${{ github.event.issue.number || inputs.issue_number }}
|
||||
SUMMARY: ${{ steps.parsed_result.outputs.summary }}
|
||||
CANDIDATE_ISSUES_JSON: ${{ steps.parsed_result.outputs.candidate_issues_json }}
|
||||
CLASSIFICATION: ${{ steps.parsed_result.outputs.classification }}
|
||||
AUTO_CLOSE_CANDIDATE: ${{ steps.parsed_result.outputs.auto_close_candidate }}
|
||||
CANONICAL_ISSUE_NUMBER: ${{ steps.parsed_result.outputs.canonical_issue_number }}
|
||||
CLOSE_AFTER_DAYS: ${{ inputs.close_after_days || '3' }}
|
||||
with:
|
||||
github-token: ${{ secrets.OPENHANDS_BOT_GITHUB_PAT_PUBLIC || github.token }}
|
||||
script: |
|
||||
const issueNumber = Number(process.env.ISSUE_NUMBER);
|
||||
const summary = (process.env.SUMMARY || '').trim();
|
||||
const classification = process.env.CLASSIFICATION || 'no-match';
|
||||
const autoClose = process.env.AUTO_CLOSE_CANDIDATE === 'true';
|
||||
const closeAfterDays = process.env.CLOSE_AFTER_DAYS || '3';
|
||||
let candidates = [];
|
||||
try {
|
||||
candidates = JSON.parse(process.env.CANDIDATE_ISSUES_JSON || '[]');
|
||||
} catch (error) {
|
||||
core.setFailed(`Invalid candidate JSON: ${error.message}`);
|
||||
return;
|
||||
}
|
||||
if (!Array.isArray(candidates)) {
|
||||
core.setFailed('CANDIDATE_ISSUES_JSON is not an array');
|
||||
return;
|
||||
}
|
||||
if (candidates.length === 0) {
|
||||
core.setFailed(`No candidate issues were returned for issue #${issueNumber}.`);
|
||||
return;
|
||||
}
|
||||
const canonicalIssueRaw = process.env.CANONICAL_ISSUE_NUMBER || candidates[0].number;
|
||||
const canonicalIssueNumber = canonicalIssueRaw ? Number(canonicalIssueRaw) : Number.NaN;
|
||||
const candidateLabel = 'duplicate-candidate';
|
||||
|
||||
function parseDuplicateCheckMarker(body) {
|
||||
if (!body) {
|
||||
return null;
|
||||
}
|
||||
const match = body.match(/<!-- openhands-duplicate-check canonical=(\d+) auto-close=(true|false) -->/);
|
||||
if (!match) {
|
||||
return null;
|
||||
}
|
||||
return {
|
||||
canonicalIssueNumber: Number(match[1]),
|
||||
autoClose: match[2] === 'true',
|
||||
};
|
||||
}
|
||||
|
||||
async function ensureCanonicalIssueIsOpenIssue() {
|
||||
let canonicalIssue;
|
||||
try {
|
||||
({ data: canonicalIssue } = await github.rest.issues.get({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: canonicalIssueNumber,
|
||||
}));
|
||||
} catch (error) {
|
||||
if (error.status === 404) {
|
||||
core.setFailed(`Canonical issue #${canonicalIssueNumber} does not exist.`);
|
||||
return false;
|
||||
}
|
||||
throw error;
|
||||
}
|
||||
if (canonicalIssue.pull_request) {
|
||||
core.setFailed(`Canonical issue #${canonicalIssueNumber} is a pull request, not an issue.`);
|
||||
return false;
|
||||
}
|
||||
if (canonicalIssue.state !== 'open' || canonicalIssue.locked) {
|
||||
core.setFailed(`Canonical issue #${canonicalIssueNumber} must be an open, unlocked issue.`);
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
async function ensureCandidateLabelOnIssue() {
|
||||
try {
|
||||
await github.rest.issues.getLabel({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
name: candidateLabel,
|
||||
});
|
||||
} catch (error) {
|
||||
if (error.status !== 404) {
|
||||
throw error;
|
||||
}
|
||||
await github.rest.issues.createLabel({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
name: candidateLabel,
|
||||
color: 'FBCA04',
|
||||
description: 'Potential duplicate awaiting auto-close or maintainer review',
|
||||
});
|
||||
}
|
||||
|
||||
const { data: issue } = await github.rest.issues.get({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
});
|
||||
const labelNames = (issue.labels || []).map((label) => (
|
||||
typeof label === 'string' ? label : label.name
|
||||
));
|
||||
if (!labelNames.includes(candidateLabel)) {
|
||||
await github.rest.issues.addLabels({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
labels: [candidateLabel],
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
async function removeCandidateLabelFromIssue() {
|
||||
try {
|
||||
await github.rest.issues.removeLabel({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
name: candidateLabel,
|
||||
});
|
||||
} catch (error) {
|
||||
if (error.status !== 404) {
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (!Number.isInteger(canonicalIssueNumber) || canonicalIssueNumber <= 0) {
|
||||
core.setFailed(`No canonical issue number was returned for issue #${issueNumber}.`);
|
||||
return;
|
||||
}
|
||||
|
||||
if (!(await ensureCanonicalIssueIsOpenIssue())) {
|
||||
return;
|
||||
}
|
||||
|
||||
const marker = `<!-- openhands-duplicate-check canonical=${canonicalIssueNumber} auto-close=${autoClose ? 'true' : 'false'} -->`;
|
||||
const header = candidates.length === 1
|
||||
? 'Found 1 possible duplicate issue:'
|
||||
: `Found ${candidates.length} possible duplicate issues:`;
|
||||
const candidateLines = candidates.map((candidate, index) => (
|
||||
`${index + 1}. [#${candidate.number}](${candidate.url}) — ${candidate.title}`
|
||||
));
|
||||
|
||||
const sections = [];
|
||||
if (summary) {
|
||||
sections.push(summary, '');
|
||||
}
|
||||
sections.push(header, '', ...candidateLines);
|
||||
|
||||
if (classification === 'overlapping-scope') {
|
||||
sections.push(
|
||||
'',
|
||||
'These may not be exact duplicates, but the scope appears to overlap enough that keeping discussion in one place may be more useful.'
|
||||
);
|
||||
}
|
||||
|
||||
if (autoClose) {
|
||||
sections.push(
|
||||
'',
|
||||
`This issue will be automatically closed as a duplicate in ${closeAfterDays} days.`,
|
||||
'',
|
||||
'- If your issue is a duplicate, please close it and 👍 the existing issue instead',
|
||||
'- To prevent auto-closure, add a comment or 👎 this comment'
|
||||
);
|
||||
}
|
||||
|
||||
sections.push(
|
||||
'',
|
||||
marker,
|
||||
'_This comment was created by an AI assistant (OpenHands) on behalf of the repository maintainer._'
|
||||
);
|
||||
const body = sections.join('\n').trim();
|
||||
|
||||
const MAX_COMMENT_PAGES = 50;
|
||||
let allComments = [];
|
||||
let page = 1;
|
||||
while (page <= MAX_COMMENT_PAGES) {
|
||||
const { data: comments } = await github.rest.issues.listComments({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
per_page: 100,
|
||||
page,
|
||||
});
|
||||
if (!comments || comments.length === 0) {
|
||||
break;
|
||||
}
|
||||
allComments = allComments.concat(comments);
|
||||
if (comments.length < 100) {
|
||||
break;
|
||||
}
|
||||
page += 1;
|
||||
}
|
||||
if (page > MAX_COMMENT_PAGES) {
|
||||
core.setFailed(
|
||||
`Stopped loading comments for issue #${issueNumber} after ${MAX_COMMENT_PAGES} pages.`
|
||||
);
|
||||
return;
|
||||
}
|
||||
|
||||
const existing = allComments.find((comment) => comment.body && comment.body.includes('<!-- openhands-duplicate-check '));
|
||||
if (existing) {
|
||||
const existingMarker = parseDuplicateCheckMarker(existing.body);
|
||||
if (existingMarker) {
|
||||
if (existingMarker.autoClose) {
|
||||
await ensureCandidateLabelOnIssue();
|
||||
} else {
|
||||
await removeCandidateLabelFromIssue();
|
||||
}
|
||||
if (
|
||||
existingMarker.canonicalIssueNumber !== canonicalIssueNumber ||
|
||||
existingMarker.autoClose !== autoClose
|
||||
) {
|
||||
core.setFailed(
|
||||
`Duplicate check comment already exists on issue #${issueNumber} with different canonical/auto-close metadata; manual reconciliation is required.`
|
||||
);
|
||||
return;
|
||||
}
|
||||
} else {
|
||||
core.warning(
|
||||
`Duplicate check comment already exists on issue #${issueNumber} but its marker could not be parsed; leaving label state unchanged.`
|
||||
);
|
||||
}
|
||||
core.info(`Duplicate check comment already exists on issue #${issueNumber}; skipping.`);
|
||||
return;
|
||||
}
|
||||
|
||||
await github.rest.issues.createComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
body,
|
||||
});
|
||||
|
||||
if (autoClose) {
|
||||
await ensureCandidateLabelOnIssue();
|
||||
}
|
||||
|
||||
auto-close-duplicates:
|
||||
if: |
|
||||
github.event_name == 'schedule' ||
|
||||
(github.event_name == 'workflow_dispatch' && inputs.mode == 'auto-close')
|
||||
runs-on: ubuntu-latest
|
||||
timeout-minutes: 20
|
||||
concurrency:
|
||||
group: auto-close-duplicates-${{ github.repository }}
|
||||
cancel-in-progress: false
|
||||
steps:
|
||||
- name: Checkout repository
|
||||
uses: actions/checkout@v6
|
||||
|
||||
- name: Set up Python
|
||||
uses: actions/setup-python@v6
|
||||
with:
|
||||
python-version: '3.13'
|
||||
|
||||
- name: Auto-close aged duplicate candidates
|
||||
env:
|
||||
GITHUB_TOKEN: ${{ secrets.OPENHANDS_BOT_GITHUB_PAT_PUBLIC || github.token }}
|
||||
CLOSE_AFTER_DAYS: ${{ inputs.close_after_days || '3' }}
|
||||
run: |
|
||||
python scripts/auto_close_duplicate_issues.py \
|
||||
--repository "${{ github.repository }}" \
|
||||
--close-after-days "$CLOSE_AFTER_DAYS" | tee "$RUNNER_TEMP/auto-close-summary.json"
|
||||
status=${PIPESTATUS[0]}
|
||||
if [ "$status" -ne 0 ]; then
|
||||
echo "::error::Auto-close script failed with exit code $status"
|
||||
exit "$status"
|
||||
fi
|
||||
|
||||
- name: Summarize auto-close run
|
||||
run: |
|
||||
{
|
||||
echo "## Auto-close duplicate candidates"
|
||||
echo
|
||||
cat "$RUNNER_TEMP/auto-close-summary.json"
|
||||
} >> "$GITHUB_STEP_SUMMARY"
|
||||
@@ -0,0 +1,64 @@
|
||||
---
|
||||
name: Remove duplicate candidate label on activity
|
||||
|
||||
on:
|
||||
issue_comment:
|
||||
types: [created]
|
||||
|
||||
permissions:
|
||||
issues: write
|
||||
|
||||
concurrency:
|
||||
group: remove-duplicate-${{ github.repository }}-${{ github.event.issue.number }}
|
||||
cancel-in-progress: false
|
||||
|
||||
jobs:
|
||||
remove-duplicate-candidate:
|
||||
if: |
|
||||
github.event.issue.state == 'open' &&
|
||||
github.event.issue.pull_request == null &&
|
||||
contains(github.event.issue.labels.*.name, 'duplicate-candidate') &&
|
||||
github.event.comment.user.type != 'Bot' &&
|
||||
!startsWith(github.event.comment.body || '', '<!-- openhands-duplicate-check') &&
|
||||
!startsWith(github.event.comment.body || '', '<!-- openhands-duplicate-veto')
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Remove duplicate-candidate label
|
||||
uses: actions/github-script@v7
|
||||
with:
|
||||
github-token: ${{ secrets.OPENHANDS_BOT_GITHUB_PAT_PUBLIC || github.token }}
|
||||
script: |
|
||||
const issueNumber = context.issue.number;
|
||||
const commenter = context.payload.comment.user.login || '';
|
||||
const normalizedCommenter = commenter.toLowerCase();
|
||||
|
||||
if (
|
||||
normalizedCommenter.endsWith('[bot]') ||
|
||||
normalizedCommenter === 'all-hands-bot'
|
||||
) {
|
||||
core.info(
|
||||
`Skipping duplicate-candidate label removal for bot comment from ${commenter || 'unknown'}`
|
||||
);
|
||||
return;
|
||||
}
|
||||
|
||||
core.info(
|
||||
`Removing duplicate-candidate label from issue #${issueNumber} after comment from ${commenter}`
|
||||
);
|
||||
|
||||
try {
|
||||
await github.rest.issues.removeLabel({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: issueNumber,
|
||||
name: 'duplicate-candidate',
|
||||
});
|
||||
} catch (error) {
|
||||
if (error.status === 404) {
|
||||
core.info(
|
||||
`duplicate-candidate label was already removed from issue #${issueNumber}`
|
||||
);
|
||||
return;
|
||||
}
|
||||
throw error;
|
||||
}
|
||||
Reference in New Issue
Block a user