@merlean/analyzer 1.1.1 → 1.2.0

package/lib/analyzer.js

@@ -67,12 +67,19 @@ async function scanCodebase(codebasePath) {
       const content = fs.readFileSync(file, 'utf-8');
       const relativePath = path.relative(codebasePath, file);
       
-      // Limit content per file
-      const truncatedContent = content.slice(0, 3000);
+      // Smart extraction: if file is large, extract route-like patterns
+      let extractedContent;
+      if (content.length > 8000) {
+        // For large files, extract route definitions and API patterns
+        extractedContent = extractRoutePatterns(content, relativePath);
+      } else {
+        // For smaller files, include more content
+        extractedContent = content.slice(0, 8000);
+      }
       
       fileContents.push({
         path: relativePath,
-        content: truncatedContent
+        content: extractedContent
       });
     } catch (error) {
       // Skip files that can't be read
@@ -82,6 +89,73 @@ async function scanCodebase(codebasePath) {
   return fileContents;
 }
 
+/**
+ * Extract route patterns from large files
+ */
+function extractRoutePatterns(content, filePath) {
+  const lines = content.split('\n');
+  const relevantLines = [];
+  
+  // Patterns that indicate API routes/endpoints
+  const routePatterns = [
+    /app\.(get|post|put|patch|delete|use)\s*\(/i,
+    /router\.(get|post|put|patch|delete|use)\s*\(/i,
+    /Route::(get|post|put|patch|delete)\s*\(/i,
+    /@(Get|Post|Put|Patch|Delete|RequestMapping)/i,
+    /def\s+(get|post|put|patch|delete|index|create|update|destroy)/i,
+    /function\s+\w+\s*\(\s*(req|request)/i,
+    /fetch\s*\(/i,
+    /axios\./i,
+    /api['"]\s*:/i,
+    /endpoint/i,
+    /\/api\//i
+  ];
+  
+  let inRouteBlock = false;
+  let braceCount = 0;
+  
+  for (let i = 0; i < lines.length; i++) {
+    const line = lines[i];
+    
+    // Check if line matches any route pattern
+    const isRouteLine = routePatterns.some(pattern => pattern.test(line));
+    
+    if (isRouteLine) {
+      // Include context: 2 lines before
+      for (let j = Math.max(0, i - 2); j < i; j++) {
+        if (!relevantLines.includes(lines[j])) {
+          relevantLines.push(`// Line ${j + 1}: ${lines[j]}`);
+        }
+      }
+      relevantLines.push(`// Line ${i + 1}: ${line}`);
+      inRouteBlock = true;
+      braceCount = (line.match(/{/g) || []).length - (line.match(/}/g) || []).length;
+    } else if (inRouteBlock) {
+      // Continue capturing the route handler
+      relevantLines.push(line);
+      braceCount += (line.match(/{/g) || []).length - (line.match(/}/g) || []).length;
+      
+      // End of route block
+      if (braceCount <= 0) {
+        inRouteBlock = false;
+        relevantLines.push('// ---');
+      }
+      
+      // Safety limit per block
+      if (relevantLines.length > 200) {
+        inRouteBlock = false;
+      }
+    }
+  }
+  
+  // If we found relevant lines, return them; otherwise return truncated content
+  if (relevantLines.length > 10) {
+    return `// Extracted route patterns from ${filePath} (${lines.length} lines total)\n\n${relevantLines.join('\n')}`;
+  }
+  
+  return content.slice(0, 8000);
+}
+
 /**
  * Prioritize files based on keywords in path/name
  */

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@merlean/analyzer",
-  "version": "1.1.1",
+  "version": "1.2.0",
   "description": "AI Bot codebase analyzer - generates site maps for AI assistant integration",
   "keywords": ["ai", "bot", "analyzer", "claude", "anthropic", "widget"],
   "author": "zmaren",