specweave 0.24.8 → 0.24.9

package/plugins/specweave/lib/hooks/reflection-parser.ts.bak

@@ -0,0 +1,484 @@
+/**
+ * Reflection Parser
+ *
+ * Parses markdown output from reflective-reviewer agent
+ * Extracts structured data (issues, metrics, lessons learned, etc.)
+ *
+ * @module reflection-parser
+ */
+
+import {
+  ReflectionResult,
+  ReflectionIssue,
+  IssueSeverity,
+  IssueCategory,
+  LessonLearned,
+  ReflectionMetrics,
+  ReflectionModel
+} from './types/reflection-types';
+
+/**
+ * Extract section content by heading
+ * @param markdown Markdown content
+ * @param heading Section heading to find (e.g., "## What Was Accomplished")
+ * @returns Section content (text after heading until next heading or end)
+ */
+function extractSection(markdown: string, heading: string): string {
+  const headingRegex = new RegExp(`^${heading}\\s*$`, 'mi');
+  const match = markdown.match(headingRegex);
+
+  if (!match || match.index === undefined) {
+    return '';
+  }
+
+  const startIndex = match.index + match[0].length;
+  const afterHeading = markdown.slice(startIndex);
+
+  // Find next heading at same or higher level
+  const nextHeadingMatch = afterHeading.match(/^#{1,3}\s+/m);
+  const endIndex = nextHeadingMatch?.index ?? afterHeading.length;
+
+  return afterHeading.slice(0, endIndex).trim();
+}
+
+/**
+ * Parse accomplishments from "What Was Accomplished" section
+ * @param markdown Reflection markdown
+ * @returns Array of accomplishment strings
+ */
+function parseAccomplishments(markdown: string): string[] {
+  const section = extractSection(markdown, '## ✅ What Was Accomplished');
+  if (!section) return [];
+
+  // Extract bullet points and paragraphs
+  const lines = section.split('\n').filter(line => line.trim());
+  const accomplishments: string[] = [];
+
+  for (const line of lines) {
+    const trimmed = line.trim();
+    // Match bullet points (-, *, +) or numbered lists (1., 2.)
+    if (/^[-*+]\s+/.test(trimmed) || /^\d+\.\s+/.test(trimmed)) {
+      accomplishments.push(trimmed.replace(/^[-*+\d.]\s+/, '').trim());
+    } else if (trimmed.length > 10 && !trimmed.startsWith('#')) {
+      // Include non-heading paragraphs
+      accomplishments.push(trimmed);
+    }
+  }
+
+  return accomplishments;
+}
+
+/**
+ * Parse strengths from "Quality Assessment" section
+ * @param markdown Reflection markdown
+ * @returns Array of strength strings
+ */
+function parseStrengths(markdown: string): string[] {
+  const section = extractSection(markdown, '### ✅ Strengths');
+  if (!section) return [];
+
+  const lines = section.split('\n').filter(line => line.trim());
+  const strengths: string[] = [];
+
+  for (const line of lines) {
+    const trimmed = line.trim();
+    if (trimmed.startsWith('- ✅') || trimmed.startsWith('✅')) {
+      strengths.push(trimmed.replace(/^[-*+]?\s*✅\s*/, '').trim());
+    }
+  }
+
+  return strengths;
+}
+
+/**
+ * Parse severity from issue description
+ * @param text Issue text (e.g., "**CRITICAL (SECURITY)**")
+ * @returns Issue severity enum or undefined
+ */
+function parseSeverity(text: string): IssueSeverity | undefined {
+  if (/CRITICAL/i.test(text)) return IssueSeverity.CRITICAL;
+  if (/HIGH/i.test(text)) return IssueSeverity.HIGH;
+  if (/MEDIUM/i.test(text)) return IssueSeverity.MEDIUM;
+  if (/LOW/i.test(text)) return IssueSeverity.LOW;
+  return undefined;
+}
+
+/**
+ * Parse category from issue description
+ * @param text Issue text (e.g., "**CRITICAL (SECURITY)**")
+ * @returns Issue category enum or undefined
+ */
+function parseCategory(text: string): IssueCategory | undefined {
+  if (/SECURITY/i.test(text)) return IssueCategory.SECURITY;
+  if (/QUALITY/i.test(text)) return IssueCategory.QUALITY;
+  if (/TESTING/i.test(text)) return IssueCategory.TESTING;
+  if (/PERFORMANCE/i.test(text)) return IssueCategory.PERFORMANCE;
+  if (/TECHNICAL[_\s]DEBT/i.test(text)) return IssueCategory.TECHNICAL_DEBT;
+  return undefined;
+}
+
+/**
+ * Parse location from issue text
+ * @param text Issue text (may contain **Location**: `path/to/file.ts:123`)
+ * @returns Location object or undefined
+ */
+function parseLocation(text: string): ReflectionIssue['location'] | undefined {
+  // Match: **Location**: `path/to/file.ts:123`
+  const locationMatch = text.match(/\*\*Location\*\*:\s*`([^`]+)`/i);
+  if (!locationMatch) return undefined;
+
+  const locationStr = locationMatch[1];
+  const [file, lineStr] = locationStr.split(':');
+  const line = lineStr ? parseInt(lineStr, 10) : undefined;
+
+  return {
+    file: file.trim(),
+    line
+  };
+}
+
+/**
+ * Parse issues from "Issues Identified" section
+ * @param markdown Reflection markdown
+ * @returns Array of ReflectionIssue objects
+ */
+function parseIssues(markdown: string): ReflectionIssue[] {
+  const section = extractSection(markdown, '### ⚠️ Issues Identified');
+  if (!section) return [];
+
+  const issues: ReflectionIssue[] = [];
+  const lines = section.split('\n');
+
+  let currentIssue: Partial<ReflectionIssue> | null = null;
+  let currentField: 'description' | 'impact' | 'recommendation' | null = null;
+
+  for (const line of lines) {
+    const trimmed = line.trim();
+
+    // Check for severity/category header (e.g., "**CRITICAL (SECURITY)**")
+    const headerMatch = trimmed.match(/^\*\*(CRITICAL|HIGH|MEDIUM|LOW)\s*\(([^)]+)\)\*\*/i);
+    if (headerMatch) {
+      // Save previous issue
+      if (currentIssue && currentIssue.severity && currentIssue.category && currentIssue.description) {
+        issues.push(currentIssue as ReflectionIssue);
+      }
+
+      // Start new issue
+      currentIssue = {
+        severity: parseSeverity(headerMatch[1])!,
+        category: parseCategory(headerMatch[2])!,
+        description: '',
+        impact: '',
+        recommendation: ''
+      };
+      currentField = null;
+      continue;
+    }
+
+    // Check for issue bullet point (❌ or ⚠️)
+    if ((trimmed.startsWith('- ❌') || trimmed.startsWith('❌') || trimmed.startsWith('- ⚠️')) && currentIssue) {
+      currentField = 'description';
+      const descText = trimmed.replace(/^[-*+]?\s*[❌⚠️]\s*/, '').trim();
+      currentIssue.description = descText;
+      continue;
+    }
+
+    // Check for Impact field
+    if (trimmed.startsWith('- **Impact**:') || trimmed.startsWith('**Impact**:')) {
+      currentField = 'impact';
+      const impactText = trimmed.replace(/^[-*+]?\s*\*\*Impact\*\*:\s*/, '').trim();
+      if (currentIssue) currentIssue.impact = impactText;
+      continue;
+    }
+
+    // Check for Recommendation field
+    if (trimmed.startsWith('- **Recommendation**:') || trimmed.startsWith('**Recommendation**:')) {
+      currentField = 'recommendation';
+      const recText = trimmed.replace(/^[-*+]?\s*\*\*Recommendation\*\*:\s*/, '').trim();
+      if (currentIssue) currentIssue.recommendation = recText;
+      continue;
+    }
+
+    // Check for Location field
+    if (trimmed.startsWith('- **Location**:') || trimmed.startsWith('**Location**:')) {
+      if (currentIssue) {
+        const location = parseLocation(trimmed);
+        if (location) currentIssue.location = location;
+      }
+      currentField = null;
+      continue;
+    }
+
+    // Append continuation lines to current field
+    if (currentField && currentIssue && trimmed && !trimmed.startsWith('#')) {
+      const fieldValue = currentIssue[currentField] || '';
+      currentIssue[currentField] = fieldValue + ' ' + trimmed;
+    }
+  }
+
+  // Save last issue
+  if (currentIssue && currentIssue.severity && currentIssue.category && currentIssue.description) {
+    issues.push(currentIssue as ReflectionIssue);
+  }
+
+  return issues;
+}
+
+/**
+ * Parse recommended actions from "Recommended Follow-Up Actions" section
+ * @param markdown Reflection markdown
+ * @returns Recommended actions object with priority levels
+ */
+function parseRecommendedActions(markdown: string): ReflectionResult['recommendedActions'] {
+  const section = extractSection(markdown, '## 🔧 Recommended Follow-Up Actions');
+
+  const actions = {
+    priority1: [] as string[],
+    priority2: [] as string[],
+    priority3: [] as string[]
+  };
+
+  if (!section) return actions;
+
+  const lines = section.split('\n');
+  let currentPriority: 'priority1' | 'priority2' | 'priority3' | null = null;
+
+  for (const line of lines) {
+    const trimmed = line.trim();
+
+    // Check for priority headers
+    if (/Priority 1|MUST FIX/i.test(trimmed)) {
+      currentPriority = 'priority1';
+      continue;
+    }
+    if (/Priority 2|SHOULD FIX/i.test(trimmed)) {
+      currentPriority = 'priority2';
+      continue;
+    }
+    if (/Priority 3|NICE TO HAVE/i.test(trimmed)) {
+      currentPriority = 'priority3';
+      continue;
+    }
+
+    // Parse action items (numbered or bulleted)
+    if (currentPriority && (/^\d+\.\s/.test(trimmed) || /^[-*+]\s/.test(trimmed))) {
+      const actionText = trimmed.replace(/^(\d+\.|-|\*|\+)\s+/, '').trim();
+      if (actionText.length > 0) {
+        actions[currentPriority].push(actionText);
+      }
+    }
+  }
+
+  return actions;
+}
+
+/**
+ * Parse lessons learned from "Lessons Learned" section
+ * @param markdown Reflection markdown
+ * @returns LessonLearned object
+ */
+function parseLessonsLearned(markdown: string): LessonLearned {
+  const section = extractSection(markdown, '## 📚 Lessons Learned');
+
+  const lessons: LessonLearned = {
+    whatWentWell: [],
+    whatCouldImprove: [],
+    forNextTime: []
+  };
+
+  if (!section) return lessons;
+
+  const lines = section.split('\n');
+  let currentCategory: 'whatWentWell' | 'whatCouldImprove' | 'forNextTime' | null = null;
+
+  for (const line of lines) {
+    const trimmed = line.trim();
+
+    // Check for category headers
+    if (/What went well/i.test(trimmed)) {
+      currentCategory = 'whatWentWell';
+      continue;
+    }
+    if (/What could improve/i.test(trimmed)) {
+      currentCategory = 'whatCouldImprove';
+      continue;
+    }
+    if (/For next time/i.test(trimmed)) {
+      currentCategory = 'forNextTime';
+      continue;
+    }
+
+    // Parse bullet points
+    if (currentCategory && /^[-*+]\s/.test(trimmed)) {
+      const text = trimmed.replace(/^[-*+]\s+/, '').trim();
+      if (text.length > 0) {
+        lessons[currentCategory].push(text);
+      }
+    }
+  }
+
+  return lessons;
+}
+
+/**
+ * Parse metrics from "Metrics" section
+ * @param markdown Reflection markdown
+ * @returns ReflectionMetrics object
+ */
+function parseMetrics(markdown: string): ReflectionMetrics {
+  const section = extractSection(markdown, '## 📊 Metrics');
+
+  const metrics: ReflectionMetrics = {
+    codeQuality: 5,
+    security: 5,
+    testCoverage: undefined,
+    technicalDebt: 'MEDIUM',
+    performance: 'ACCEPTABLE'
+  };
+
+  if (!section) return metrics;
+
+  // Parse Code Quality (1-10)
+  const qualityMatch = section.match(/Code Quality.*?(\d+)/i);
+  if (qualityMatch) {
+    metrics.codeQuality = parseInt(qualityMatch[1], 10);
+  }
+
+  // Parse Security (1-10)
+  const securityMatch = section.match(/Security.*?(\d+)/i);
+  if (securityMatch) {
+    metrics.security = parseInt(securityMatch[1], 10);
+  }
+
+  // Parse Test Coverage (percentage)
+  const coverageMatch = section.match(/Test Coverage.*?(\d+)%/i);
+  if (coverageMatch) {
+    metrics.testCoverage = parseInt(coverageMatch[1], 10);
+  }
+
+  // Parse Technical Debt (LOW/MEDIUM/HIGH)
+  const debtMatch = section.match(/Technical Debt.*?(LOW|MEDIUM|HIGH)/i);
+  if (debtMatch) {
+    metrics.technicalDebt = debtMatch[1].toUpperCase() as 'LOW' | 'MEDIUM' | 'HIGH';
+  }
+
+  // Parse Performance (GOOD/ACCEPTABLE/NEEDS WORK)
+  const perfMatch = section.match(/Performance.*?(GOOD|ACCEPTABLE|NEEDS[\s_]WORK)/i);
+  if (perfMatch) {
+    metrics.performance = perfMatch[1].toUpperCase().replace(/[\s_]/g, '_') as 'GOOD' | 'ACCEPTABLE' | 'NEEDS_WORK';
+  }
+
+  return metrics;
+}
+
+/**
+ * Parse complete reflection result from markdown
+ * Main parsing function
+ *
+ * @param markdown Reflection markdown from reflective-reviewer agent
+ * @param taskName Task name for result
+ * @param model Model used for reflection
+ * @param reflectionTime Time taken in seconds
+ * @param estimatedCost Estimated cost in USD
+ * @returns Complete ReflectionResult object
+ */
+export function parseReflectionMarkdown(
+  markdown: string,
+  taskName: string,
+  model: ReflectionModel = ReflectionModel.HAIKU,
+  reflectionTime: number = 0,
+  estimatedCost: number = 0
+): ReflectionResult {
+  const result: ReflectionResult = {
+    taskName,
+    completed: new Date().toISOString(),
+    duration: reflectionTime > 0 ? `${reflectionTime}s` : undefined,
+    filesModified: {
+      count: 0,
+      linesAdded: 0,
+      linesRemoved: 0
+    },
+    accomplishments: parseAccomplishments(markdown),
+    strengths: parseStrengths(markdown),
+    issues: parseIssues(markdown),
+    recommendedActions: parseRecommendedActions(markdown),
+    lessonsLearned: parseLessonsLearned(markdown),
+    metrics: parseMetrics(markdown),
+    metadata: {
+      model,
+      reflectionTime,
+      estimatedCost
+    }
+  };
+
+  // Extract file modification stats from markdown if present
+  const filesChangedMatch = markdown.match(/\*\*Files Modified\*\*:\s*(\d+)\s*files?,\s*\+(\d+)\s*-(\d+)/i);
+  if (filesChangedMatch) {
+    result.filesModified = {
+      count: parseInt(filesChangedMatch[1], 10),
+      linesAdded: parseInt(filesChangedMatch[2], 10),
+      linesRemoved: parseInt(filesChangedMatch[3], 10)
+    };
+  }
+
+  return result;
+}
+
+/**
+ * Validate parsed reflection result
+ * Checks for required fields and data quality
+ *
+ * @param result Parsed reflection result
+ * @returns Validation result with errors
+ */
+export function validateReflectionResult(result: ReflectionResult): {
+  valid: boolean;
+  errors: string[];
+} {
+  const errors: string[] = [];
+
+  // Check required fields
+  if (!result.taskName) {
+    errors.push('Missing task name');
+  }
+
+  if (!result.completed) {
+    errors.push('Missing completion timestamp');
+  }
+
+  // Check for meaningful content
+  if (result.accomplishments.length === 0 && result.strengths.length === 0) {
+    errors.push('No accomplishments or strengths identified (reflection too sparse)');
+  }
+
+  // Validate metrics
+  if (result.metrics.codeQuality < 1 || result.metrics.codeQuality > 10) {
+    errors.push('Code quality must be 1-10');
+  }
+
+  if (result.metrics.security < 1 || result.metrics.security > 10) {
+    errors.push('Security must be 1-10');
+  }
+
+  if (result.metrics.testCoverage !== undefined && (result.metrics.testCoverage < 0 || result.metrics.testCoverage > 100)) {
+    errors.push('Test coverage must be 0-100%');
+  }
+
+  // Validate issues have required fields
+  for (const issue of result.issues) {
+    if (!issue.description) {
+      errors.push('Issue missing description');
+    }
+    if (!issue.impact) {
+      errors.push('Issue missing impact explanation');
+    }
+    if (!issue.recommendation) {
+      errors.push('Issue missing recommendation');
+    }
+  }
+
+  return {
+    valid: errors.length === 0,
+    errors
+  };
+}

package/plugins/specweave/lib/hooks/reflection-prompt-builder.d.ts.bak

@@ -0,0 +1,56 @@
+/**
+ * Reflection Prompt Builder
+ *
+ * Builds comprehensive prompts for the reflective-reviewer agent
+ * Includes modified files context, configuration, task information
+ *
+ * @module reflection-prompt-builder
+ */
+import { ReflectionConfig, GitDiffInfo } from './types/reflection-types';
+/**
+ * Build the complete reflection prompt
+ * Main function for creating agent prompts
+ *
+ * @param options Prompt options
+ * @returns Complete markdown-formatted prompt for reflective-reviewer agent
+ */
+export declare function buildReflectionPrompt(options: {
+    taskId: string;
+    taskName?: string;
+    modifiedFiles: GitDiffInfo[];
+    config: ReflectionConfig;
+    incrementId?: string;
+    includeFullDiff?: boolean;
+}): string;
+/**
+ * Estimate token count for prompt (rough approximation)
+ * Used for cost estimation
+ *
+ * @param prompt Reflection prompt
+ * @returns Estimated token count (1 token ≈ 4 characters)
+ */
+export declare function estimatePromptTokens(prompt: string): number;
+/**
+ * Truncate large diffs to fit within token budget
+ * Preserves most recent changes (bottom of diff)
+ *
+ * @param modifiedFiles Array of modified files
+ * @param maxTokens Maximum tokens to use (default: 8000)
+ * @returns Truncated array of modified files
+ */
+export declare function truncateModifiedFiles(modifiedFiles: GitDiffInfo[], maxTokens?: number): GitDiffInfo[];
+/**
+ * Build a simplified prompt for very large changes
+ * Used when full diff exceeds reasonable token budget
+ *
+ * @param options Prompt options (same as buildReflectionPrompt)
+ * @returns Simplified prompt with file list only (no diffs)
+ */
+export declare function buildSimplifiedPrompt(options: {
+    taskId: string;
+    taskName?: string;
+    modifiedFiles: GitDiffInfo[];
+    config: ReflectionConfig;
+    incrementId?: string;
+}): string;
+//# sourceMappingURL=reflection-prompt-builder.d.ts.map