Add script to remove duplicate issues on declarations repository (#1115)

Author: Ndpnt

CHANGELOG.md

@@ -2,6 +2,14 @@
 
 All changes that impact users of this module are documented in this file, in the [Common Changelog](https://common-changelog.org) format with some additional specifications defined in the CONTRIBUTING file. This codebase adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## Unreleased [minor]
+
+> Development of this release was supported by the [French Ministry for Foreign Affairs](https://www.diplomatie.gouv.fr/fr/politique-etrangere-de-la-france/diplomatie-numerique/) through its ministerial [State Startups incubator](https://beta.gouv.fr/startups/open-terms-archive.html) under the aegis of the Ambassador for Digital Affairs.
+
+### Added
+
+- Add script to remove duplicate issues in GitHub reports
+
 ## 2.4.0 - 2024-10-24
 
 _Full changeset and discussions: [#1114](https://github.com/OpenTermsArchive/engine/pull/1114)._

scripts/reporter/duplicate/README.md

@@ -0,0 +1,37 @@
+# Duplicate issues removal script
+
+This script helps remove duplicate issues from a GitHub repository by closing issues that have the same title as any older issue.
+
+## Prerequisites
+
+1. Set up environment variables:
+   - Create a `.env` file in the root directory
+   - Add the GitHub personal access token of the bot that manages issues on your collection, with `repo` permissions:
+
+```shell
+OTA_ENGINE_GITHUB_TOKEN=your_github_token
+```
+
+2. Configure the target repository in your chosen configuration file within the `config` folder:
+
+```json
+{
+  "@opentermsarchive/engine": {
+    "reporter": {
+      "githubIssues": {
+        "repositories": {
+          "declarations": "owner/repository"
+        }
+      }
+    }
+  }
+}
+```
+
+## Usage
+
+Run the script using:
+
+```shell
+node scripts/reporter/duplicate/index.js
+```

scripts/reporter/duplicate/index.js

@@ -0,0 +1,73 @@
+import 'dotenv/config';
+import config from 'config';
+import { Octokit } from 'octokit';
+
+async function removeDuplicateIssues() {
+  const repository = config.get('@opentermsarchive/engine.reporter.githubIssues.repositories.declarations');
+
+  if (!repository.includes('/') || repository.includes('https://')) {
+    throw new Error(`Configuration entry "reporter.githubIssues.repositories.declarations" is expected to be a string in the format <owner>/<repo>, but received: "${repository}"`);
+  }
+
+  const [ owner, repo ] = repository.split('/');
+
+  const octokit = new Octokit({ auth: process.env.OTA_ENGINE_GITHUB_TOKEN });
+
+  console.log(`Getting issues from repository ${repository}…`);
+
+  const issues = await octokit.paginate('GET /repos/{owner}/{repo}/issues', {
+    owner,
+    repo,
+    state: 'open',
+    per_page: 100,
+  });
+
+  const onlyIssues = issues.filter(issue => !issue.pull_request);
+  const issuesByTitle = new Map();
+  let counter = 0;
+
+  console.log(`Found ${onlyIssues.length} issues`);
+
+  for (const issue of onlyIssues) {
+    if (!issuesByTitle.has(issue.title)) {
+      issuesByTitle.set(issue.title, [issue]);
+    } else {
+      issuesByTitle.get(issue.title).push(issue);
+    }
+  }
+
+  for (const [ title, duplicateIssues ] of issuesByTitle) {
+    if (duplicateIssues.length === 1) continue;
+
+    const originalIssue = duplicateIssues.reduce((oldest, current) => (new Date(current.created_at) < new Date(oldest.created_at) ? current : oldest));
+
+    console.log(`\nFound ${duplicateIssues.length - 1} duplicates for issue #${originalIssue.number} "${title}"`);
+
+    for (const issue of duplicateIssues) {
+      if (issue.number === originalIssue.number) {
+        continue;
+      }
+
+      await octokit.request('PATCH /repos/{owner}/{repo}/issues/{issue_number}', { /* eslint-disable-line no-await-in-loop */
+        owner,
+        repo,
+        issue_number: issue.number,
+        state: 'closed',
+      });
+
+      await octokit.request('POST /repos/{owner}/{repo}/issues/{issue_number}/comments', { /* eslint-disable-line no-await-in-loop */
+        owner,
+        repo,
+        issue_number: issue.number,
+        body: `This issue is detected as duplicate as it has the same title as #${originalIssue.number}. It most likely was created accidentally by an engine older than [v2.3.2](https://github.com/OpenTermsArchive/engine/releases/tag/v2.3.2). Closing automatically.`,
+      });
+
+      counter++;
+      console.log(`Closed issue #${issue.number}: ${issue.html_url}`);
+    }
+  }
+
+  console.log(`\nDuplicate removal process completed; ${counter} issues closed`);
+}
+
+removeDuplicateIssues();