coderabbitai-mcp/dist/tools/get-review-details.js

MCP server for interacting with CodeRabbit AI reviews on GitHub pull requests. Enables LLMs to analyze, implement, and resolve CodeRabbit suggestions programmatically.

import { z } from 'zod';
const GetReviewDetailsSchema = z.object({
    owner: z.string().min(1, "Repository owner is required"),
    repo: z.string().min(1, "Repository name is required"),
    pullNumber: z.number().int().positive("Pull request number must be positive"),
    reviewId: z.number().int().positive("Review ID must be positive")
});
/**
 * Parse CodeRabbit review body to extract detailed structured information
 */
function parseCoderabbitReviewBodyDetailed(body) {
    const lines = body.split('\n');
    // Extract actionable comments count
    const actionableMatch = body.match(/\*\*Actionable comments posted: (\d+)\*\*/);
    const actionableCount = actionableMatch ? parseInt(actionableMatch[1]) : 0;
    // Extract duplicate comments count  
    const duplicateMatch = body.match(/♻️ Duplicate comments \((\d+)\)/);
    const duplicateCount = duplicateMatch ? parseInt(duplicateMatch[1]) : 0;
    // Extract nitpick comments count
    const nitpickMatch = body.match(/🧹 Nitpick comments \((\d+)\)/);
    const nitpickCount = nitpickMatch ? parseInt(nitpickMatch[1]) : 0;
    // Extract configuration information
    const configMatch = body.match(/\*\*Configuration used: (.+?)\*\*/);
    const configuration = configMatch ? configMatch[1] : 'Unknown';
    // Extract review profile
    const profileMatch = body.match(/\*\*Review profile: (.+?)\*\*/);
    const reviewProfile = profileMatch ? profileMatch[1] : 'Unknown';
    // Extract files reviewed from the "Files selected for processing" section
    const filesSection = body.match(/📒 Files selected for processing.*?\n(.*?)(?=\n\n|\n<|$)/s);
    const filesReviewed = [];
    if (filesSection) {
        const fileLines = filesSection[1].split('\n');
        for (const line of fileLines) {
            const fileMatch = line.match(/\* `([^`]+)`/);
            if (fileMatch) {
                filesReviewed.push(fileMatch[1]);
            }
        }
    }
    // Extract summary from review details section
    let summary = "";
    const summaryMatch = body.match(/📜 Review details[\s\S]*?(?=\n\n|\n<|$)/);
    if (summaryMatch) {
        summary = summaryMatch[0].replace(/📜 Review details/, '').trim();
    }
    // Parse detailed comments from the review body
    const comments = [];
    // Look for comment blocks in details sections
    const commentSections = body.split('<details>');
    for (const section of commentSections) {
        if (!section.includes('</details>'))
            continue;
        // Extract category from summary
        const summaryMatch = section.match(/<summary>(.*?)<\/summary>/);
        if (!summaryMatch)
            continue;
        const category = summaryMatch[1].trim();
        // Skip non-code-review sections
        if (category.includes('Commits') || category.includes('Files ignored'))
            continue;
        // Parse individual comments within this section
        const commentBlocks = section.split('</blockquote>');
        for (const block of commentBlocks) {
            if (!block.trim())
                continue;
            // Determine severity from emoji/markers
            let severity = 'info';
            if (block.includes('⚠️ Potential issue') || block.includes('_⚠️ Potential issue_')) {
                severity = 'warning';
            }
            else if (block.includes('🛠️ Refactor suggestion') || block.includes('_🛠️ Refactor suggestion_')) {
                severity = 'suggestion';
            }
            else if (block.includes('🧹 Nitpick')) {
                severity = 'info';
            }
            else if (block.includes('🔒 Security')) {
                severity = 'error';
            }
            // Extract AI prompt
            const aiPromptMatch = block.match(/🤖 Prompt for AI Agents[\s\S]*?```\n([\s\S]*?)\n```/);
            const aiPrompt = aiPromptMatch ? aiPromptMatch[1].trim() : undefined;
            // Extract committable suggestion
            const committableMatch = block.match(/📝 Committable suggestion[\s\S]*?```suggestion\n([\s\S]*?)\n```/);
            const committableSuggestion = committableMatch ? committableMatch[1].trim() : undefined;
            // Extract file path from the block
            const filePathMatch = block.match(/`([^`]+\.(js|ts|tsx|jsx|py|java|cpp|c|h|go|rs|rb|php|cs|kt|swift))`/);
            const filePath = filePathMatch ? filePathMatch[1] : undefined;
            // Extract line range
            const lineRangeMatch = block.match(/lines? (\d+(?:-\d+)?)/i);
            const lineRange = lineRangeMatch ? lineRangeMatch[1] : undefined;
            // Extract description (remove markdown formatting)
            const descLines = block.split('\n').filter(line => line.trim() &&
                !line.includes('⚠️') &&
                !line.includes('🛠️') &&
                !line.includes('🤖 Prompt') &&
                !line.includes('📝 Committable') &&
                !line.trim().startsWith('<') &&
                !line.trim().startsWith('```') &&
                !line.trim().startsWith('---') &&
                line.trim() !== 'suggestion');
            let description = descLines.slice(0, 2).join(' ').trim();
            description = description.replace(/\*\*/g, '').replace(/`([^`]+)`/g, '$1');
            if (description && (aiPrompt || committableSuggestion || filePath)) {
                comments.push({
                    category,
                    severity,
                    description: description || 'CodeRabbit suggestion',
                    ai_prompt: aiPrompt,
                    committable_suggestion: committableSuggestion,
                    file_path: filePath,
                    line_range: lineRange
                });
            }
        }
    }
    return {
        actionable_comments: actionableCount,
        duplicate_comments: duplicateCount,
        nitpick_comments: nitpickCount,
        summary,
        comments,
        files_reviewed: filesReviewed,
        configuration_used: configuration,
        review_profile: reviewProfile
    };
}
/**
 * Get detailed information about a specific CodeRabbit review
 */
export async function getReviewDetails(input, githubClient) {
    // Validate input
    const validatedInput = GetReviewDetailsSchema.parse(input);
    const { owner, repo, pullNumber, reviewId } = validatedInput;
    try {
        // Get all reviews for the PR
        const reviews = await githubClient.getPullRequestReviews(owner, repo, pullNumber);
        // Find the specific CodeRabbit review
        const targetReview = reviews.find(review => review.id === reviewId && review.user.login === 'coderabbitai[bot]');
        if (!targetReview) {
            throw new Error(`CodeRabbit review with ID ${reviewId} not found in PR #${pullNumber}`);
        }
        // Parse the review body for detailed information
        const parsed = parseCoderabbitReviewBodyDetailed(targetReview.body);
        const reviewDetails = {
            id: targetReview.id,
            submitted_at: targetReview.submitted_at,
            html_url: targetReview.html_url,
            state: targetReview.state,
            commit_id: targetReview.commit_id,
            body: targetReview.body,
            parsed_content: {
                actionable_comments: parsed.actionable_comments,
                duplicate_comments: parsed.duplicate_comments,
                nitpick_comments: parsed.nitpick_comments,
                summary: parsed.summary,
                comments: parsed.comments
            },
            files_reviewed: parsed.files_reviewed,
            configuration_used: parsed.configuration_used,
            review_profile: parsed.review_profile
        };
        return reviewDetails;
    }
    catch (error) {
        throw new Error(`Failed to get review details: ${error instanceof Error ? error.message : String(error)}`);
    }
}
//# sourceMappingURL=get-review-details.js.map