issue-assessment.yaml

name: Issue Assessment

on:
  issues:
    types: [labeled]

jobs:
  # Comment when need-more-info label is manually added
  need-more-info:
    if: github.event.label.name == 'need-more-info'
    runs-on: ubuntu-latest
    permissions:
      issues: write
    steps:
      - name: Add comment
        uses: actions/github-script@v7
        with:
          script: |
            await github.rest.issues.createComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: context.issue.number,
              body: `## Additional Information Needed

This issue is missing information we need to investigate the problem.

### Please provide:
- [ ] **Reproduction repository** — A public GitHub repo that demonstrates the issue
- [ ] **Prettier extension logs** — Click "Prettier" in the status bar and copy the output
- [ ] **Steps to reproduce** — Clear steps we can follow to see the issue

### Why this matters
Most issues are specific to your project's configuration. Without a reproduction, we usually cannot diagnose the problem.

### Resources
- [Creating a minimal reproduction](https://github.com/prettier/prettier-vscode/blob/main/docs/troubleshooting.md)
- [Writing a good issue](https://github.com/prettier/prettier-vscode/blob/main/docs/writing-an-issue.md)

> **Note:** Issues without the requested information may be closed after 7 days.`
            });

  triage:
    if: github.event.label.name == 'triage-pending'
    runs-on: ubuntu-latest
    permissions:
      issues: write
      models: read
      contents: read
    outputs:
      triage_labels: ${{ steps.triage.outputs.ai_assessments }}
    steps:
      - name: Checkout repository
        uses: actions/checkout@v6

      - name: Run Issue Triage
        id: triage
        uses: github/ai-assessment-comment-labeler@main
        with:
          token: ${{ secrets.GITHUB_TOKEN }}
          issue_number: ${{ github.event.issue.number }}
          issue_body: ${{ github.event.issue.body }}
          ai_review_label: "triage-pending"
          prompts_directory: "./.github/prompts"
          labels_to_prompts_mapping: "triage-pending,issue-triage.prompt.yml"

      - name: Remove pending label
        uses: actions/github-script@v7
        with:
          script: |
            await github.rest.issues.removeLabel({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: context.issue.number,
              name: 'triage-pending'
            }).catch(() => {});

  post-triage:
    runs-on: ubuntu-latest
    needs: triage
    if: always() && needs.triage.result == 'success'
    permissions:
      issues: write
    steps:
      - name: Process triage results and post comments
        uses: actions/github-script@v7
        env:
          TRIAGE_OUTPUT: ${{ needs.triage.outputs.triage_labels }}
        with:
          script: |
            const issueNumber = context.issue.number;
            const triageOutput = process.env.TRIAGE_OUTPUT;

            let triageAssessments = [];
            try {
              triageAssessments = JSON.parse(triageOutput || '[]');
            } catch (e) {
              console.log('Could not parse triage output:', e);
            }

            const triageLabel = triageAssessments[0]?.assessmentLabel || '';
            console.log('Triage assessment label:', triageLabel);

            const responses = {
              'ai:issue-triage:needs-reproduction': {
                addLabel: 'need-more-info',
                comment: `## Reproduction Repository Needed

Thanks for the report! To investigate this issue, we need more information.

### Please provide:
1. **A GitHub repository** that reproduces the issue
2. **Step-by-step instructions** to reproduce the problem
3. **Prettier extension logs** — Click "Prettier" in the status bar and copy the output

### Why this matters
Most issues depend on specific project configurations. A reproduction repository lets us see exactly what you're experiencing.

### Resources
- [Creating a minimal reproduction](https://github.com/prettier/prettier-vscode/blob/main/docs/troubleshooting.md)
- [Writing a good issue](https://github.com/prettier/prettier-vscode/blob/main/docs/writing-an-issue.md)

> **Note:** Issues without reproduction info may be closed after 7 days.`
              },
              'ai:issue-triage:prettier-core': {
                addLabel: 'prettier-core',
                comment: `## This Looks Like a Prettier Core Issue

Thanks for the report! This issue appears to be about **how Prettier formats code**, not about the VS Code extension.

### Understanding the difference
| Prettier Core | VS Code Extension |
|--------------|-------------------|
| Decides *how* code is formatted | Integrates Prettier into VS Code |
| Formatting rules and output | Editor integration, settings UI |
| [prettier/prettier](https://github.com/prettier/prettier) | [prettier/prettier-vscode](https://github.com/prettier/prettier-vscode) |

### To confirm this is a Prettier Core issue
Run in your terminal:
\`\`\`bash
npx prettier --write yourfile.js
\`\`\`
If you see the same formatting behavior, it's a Prettier Core issue.

### Next steps
1. Check the [Prettier Playground](https://prettier.io/playground/) to test formatting
2. Search [existing Prettier issues](https://github.com/prettier/prettier/issues)
3. Open an issue at **[prettier/prettier](https://github.com/prettier/prettier/issues/new)**

### If this IS an extension issue
If the CLI formats differently than the extension, please reply with:
- Output from \`npx prettier --version\`
- Prettier extension logs
- Example showing the difference`
              },
              'ai:issue-triage:performance-info-needed': {
                addLabel: 'need-more-info',
                comment: `## Performance Information Needed

Thanks for reporting this performance issue! To investigate, we need additional details.

### Please provide:

| Information | How to get it |
|------------|---------------|
| Prettier version | Check the Prettier output panel |
| File size | Size of slow-to-format files |
| File types affected | Which languages/extensions |
| Prettier plugins | List any plugins you're using |
| Debug logs | Enable with \`"prettier.enableDebugLogs": true\` |

### Quick test
Compare with the CLI:
\`\`\`bash
time npx prettier --write yourfile.js
\`\`\`
If the CLI is also slow, the issue is with Prettier Core, not this extension.

### Optional: VS Code performance profile
1. Open Command Palette → **Developer: Start Performance Profile**
2. Format a document
3. Stop the profile and attach it here

> **Note:** Issues without the requested information may be closed after 7 days.`
              },
              'ai:issue-triage:incomplete-template': {
                addLabel: 'need-more-info',
                comment: `## Issue Template Incomplete

Thanks for opening this issue! Some required information is missing.

### Please edit your issue to include:
- [ ] **Summary** — Brief description of the problem
- [ ] **Reproduction repository** — Public GitHub repo demonstrating the issue
- [ ] **Steps to reproduce** — Clear steps to see the problem
- [ ] **Expected vs actual result** — What should happen vs what happens
- [ ] **Environment info:**
  - VS Code version (\`Help → About → Copy\`)
  - Prettier extension version
  - Operating system
- [ ] **Prettier logs** — Click "Prettier" in status bar, copy output

### Resources
- [Writing a good issue](https://github.com/prettier/prettier-vscode/blob/main/docs/writing-an-issue.md)
- [Troubleshooting guide](https://github.com/prettier/prettier-vscode/blob/main/docs/troubleshooting.md)

> **Note:** Issues without the requested information may be closed after 7 days.`
              },
              'ai:issue-triage:support-question': {
                comment: `## This Looks Like a Support Question

Thanks for reaching out! This appears to be a question about configuration or usage rather than a bug report.

### Get help here:
| Resource | Best for |
|----------|----------|
| [Stack Overflow](https://stackoverflow.com/questions/tagged/prettier-vscode) | Configuration questions |
| [Prettier Docs](https://prettier.io/docs/en/index.html) | Formatting options |
| [Extension Settings](https://github.com/prettier/prettier-vscode#settings) | VS Code settings |
| [Troubleshooting Guide](https://github.com/prettier/prettier-vscode/blob/main/docs/troubleshooting.md) | Common issues |

### If this is actually a bug
Please reopen with:
1. What you've already tried
2. A minimal reproduction repository
3. Prettier extension logs

We're happy to help with confirmed bugs!`
              }
            };

            if (triageLabel && responses[triageLabel]) {
              const response = responses[triageLabel];

              if (response.comment) {
                await github.rest.issues.createComment({
                  owner: context.repo.owner,
                  repo: context.repo.repo,
                  issue_number: issueNumber,
                  body: response.comment
                });
              }

              if (response.addLabel) {
                await github.rest.issues.addLabels({
                  owner: context.repo.owner,
                  repo: context.repo.repo,
                  issue_number: issueNumber,
                  labels: [response.addLabel]
                });
              }
            }

  # Duplicate detection
  duplicate-search:
    runs-on: ubuntu-latest
    needs: triage
    if: always() && needs.triage.result == 'success'
    permissions:
      issues: write
      models: read
      contents: read
    steps:
      - name: Checkout repository
        uses: actions/checkout@v6

      - name: Search for candidate issues
        id: search
        uses: actions/github-script@v7
        with:
          script: |
            const issueTitle = context.payload.issue.title;
            const issueBody = context.payload.issue.body || '';
            const issueNumber = context.issue.number;

            // Common words to filter out
            const stopWords = new Set([
              'the', 'a', 'an', 'is', 'are', 'was', 'were', 'be', 'been', 'being',
              'have', 'has', 'had', 'do', 'does', 'did', 'will', 'would', 'could',
              'should', 'may', 'might', 'must', 'can', 'need', 'to', 'of', 'in',
              'for', 'on', 'with', 'at', 'by', 'from', 'as', 'into', 'through',
              'during', 'before', 'after', 'then', 'once', 'here', 'there', 'when',
              'where', 'why', 'how', 'all', 'each', 'few', 'more', 'most', 'other',
              'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so',
              'than', 'too', 'very', 'just', 'and', 'but', 'if', 'or', 'because',
              'until', 'while', 'this', 'that', 'these', 'those', 'i', 'me', 'my',
              'we', 'our', 'you', 'your', 'he', 'him', 'she', 'her', 'it', 'its',
              'they', 'them', 'their', 'what', 'which', 'who', 'whom',
              'prettier', 'vscode', 'extension', 'code', 'file', 'files', 'issue',
              'bug', 'error', 'problem', 'help', 'please', 'work', 'working',
              'works', 'doesnt', "doesn't", 'dont', "don't", 'cant', "can't",
              'wont', "won't", 'using', 'use', 'get', 'getting', 'not', 'format',
              'formatting', 'save', 'saving'
            ]);

            const extractKeywords = (text) => {
              return text
                .toLowerCase()
                .replace(/[^\w\s-]/g, ' ')
                .split(/\s+/)
                .filter(word => word.length > 2 && !stopWords.has(word));
            };

            const titleKeywords = extractKeywords(issueTitle);
            console.log('Keywords:', titleKeywords);

            if (titleKeywords.length === 0) {
              console.log('No keywords, skipping');
              core.setOutput('candidates', '[]');
              core.setOutput('has_candidates', 'false');
              return;
            }

            const searchKeywords = titleKeywords.slice(0, 4).join(' ');
            const query = `repo:${context.repo.owner}/${context.repo.repo} is:issue ${searchKeywords}`;

            let candidates = [];
            try {
              const result = await github.rest.search.issuesAndPullRequests({
                q: query,
                sort: 'updated',
                order: 'desc',
                per_page: 10
              });

              for (const issue of result.data.items) {
                if (issue.number === issueNumber) continue;
                candidates.push({
                  number: issue.number,
                  title: issue.title,
                  body: (issue.body || '').substring(0, 500),
                  state: issue.state,
                  url: issue.html_url
                });
              }
            } catch (e) {
              console.log('Search error:', e.message);
            }

            candidates = candidates.slice(0, 8);
            console.log(`Found ${candidates.length} candidates`);

            // Build comparison text for AI
            const newIssue = `## NEW ISSUE #${issueNumber}
            Title: ${issueTitle}
            Body: ${issueBody.substring(0, 1000)}`;

            let candidateText = '';
            for (const c of candidates) {
              candidateText += `\n\n## CANDIDATE ISSUE #${c.number} (${c.state})
            Title: ${c.title}
            Body: ${c.body}`;
            }

            const comparisonInput = newIssue + '\n\n---\n\nCANDIDATE ISSUES TO COMPARE:' + candidateText;

            core.setOutput('candidates', JSON.stringify(candidates));
            core.setOutput('has_candidates', candidates.length > 0 ? 'true' : 'false');
            core.setOutput('comparison_input', comparisonInput);

      - name: Duplicate Analysis
        id: analysis
        if: steps.search.outputs.has_candidates == 'true'
        uses: github/ai-assessment-comment-labeler@main
        with:
          token: ${{ secrets.GITHUB_TOKEN }}
          issue_number: ${{ github.event.issue.number }}
          issue_body: ${{ steps.search.outputs.comparison_input }}
          ai_review_label: "triage-pending"
          prompts_directory: "./.github/prompts"
          labels_to_prompts_mapping: "triage-pending,duplicate-comparison.prompt.yml"
          suppress_comments: true
          suppress_labels: true

      - name: Post duplicate analysis results
        if: steps.search.outputs.has_candidates == 'true'
        uses: actions/github-script@v7
        env:
          AI_OUTPUT: ${{ steps.analysis.outputs.ai_assessments }}
          CANDIDATES: ${{ steps.search.outputs.candidates }}
        with:
          script: |
            const issueNumber = context.issue.number;
            const aiOutput = process.env.AI_OUTPUT;
            const candidatesJson = process.env.CANDIDATES;

            let candidates = [];
            try {
              candidates = JSON.parse(candidatesJson || '[]');
            } catch (e) {
              console.log('Could not parse candidates');
              return;
            }

            if (candidates.length === 0) {
              console.log('No candidates to analyze');
              return;
            }

            // Try to parse AI response
            let aiAssessments = [];
            let analysis = null;

            try {
              aiAssessments = JSON.parse(aiOutput || '[]');
              const response = aiAssessments[0]?.response || '';
              console.log('AI Response:', response);

              // Try to extract JSON from the response
              const jsonMatch = response.match(/\{[\s\S]*\}/);
              if (jsonMatch) {
                analysis = JSON.parse(jsonMatch[0]);
              }
            } catch (e) {
              console.log('Could not parse AI analysis:', e.message);
            }

            // Build comment based on AI analysis
            let comment = '';
            let hasDuplicates = false;
            let hasRelated = false;

            if (analysis && (analysis.duplicates?.length > 0 || analysis.related?.length > 0)) {
              // AI found duplicates or related issues
              if (analysis.duplicates?.length > 0) {
                hasDuplicates = true;
                comment += `### Likely Duplicate Issues Found\n\n`;
                comment += `This issue appears to be a duplicate of:\n\n`;
                comment += `| Issue | Title | Confidence | Reason |\n`;
                comment += `|-------|-------|------------|--------|\n`;

                for (const dup of analysis.duplicates) {
                  const candidate = candidates.find(c => c.number === dup.number);
                  if (candidate) {
                    const title = candidate.title.replace(/\|/g, '\\|').substring(0, 40) + (candidate.title.length > 40 ? '...' : '');
                    const status = candidate.state === 'open' ? '🟢' : '🔴';
                    comment += `| ${status} #${dup.number} | ${title} | ${dup.confidence} | ${dup.reason} |\n`;
                  }
                }

                comment += `\n**Please check these issues.** If your issue is indeed a duplicate, close this issue and add your information to the existing one instead.\n\n`;
              }

              if (analysis.related?.length > 0) {
                hasRelated = true;
                comment += `### Possibly Related Issues\n\n`;
                comment += `| Issue | Title | Reason |\n`;
                comment += `|-------|-------|--------|\n`;

                for (const rel of analysis.related) {
                  const candidate = candidates.find(c => c.number === rel.number);
                  if (candidate) {
                    const title = candidate.title.replace(/\|/g, '\\|').substring(0, 50) + (candidate.title.length > 50 ? '...' : '');
                    const status = candidate.state === 'open' ? '🟢' : '🔴';
                    comment += `| ${status} #${rel.number} | ${title} | ${rel.reason} |\n`;
                  }
                }

                comment += `\nThese issues might provide useful context or workarounds.\n\n`;
              }
            } else {
              // Fallback: just show candidate issues if AI analysis failed
              console.log('AI analysis unavailable, showing keyword matches');

              comment = `### Potentially Related Issues\n\n`;
              comment += `I found some existing issues that might be related based on keywords:\n\n`;
              comment += `| Issue | Title | Status |\n`;
              comment += `|-------|-------|--------|\n`;

              for (const c of candidates.slice(0, 5)) {
                const title = c.title.replace(/\|/g, '\\|').substring(0, 50) + (c.title.length > 50 ? '...' : '');
                const status = c.state === 'open' ? '🟢 Open' : '🔴 Closed';
                comment += `| #${c.number} | ${title} | ${status} |\n`;
              }

              comment += `\nPlease check if any of these address your issue.\n\n`;
            }

            comment += `---\n<sub>This analysis was generated automatically to help identify duplicate issues.</sub>`;

            await github.rest.issues.createComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: issueNumber,
              body: comment
            });