@codeyam/codeyam-cli 0.1.0-staging.323686 → 0.1.0-staging.483fdc2

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/BatchSchemaProcessor.ts

@@ -41,6 +41,7 @@ export interface BatchProcessorStats {
 
 export class BatchSchemaProcessor {
   private workQueue: WorkItem[] = [];
+  private queueIndex = 0;
   private stats: BatchProcessorStats = {
     itemsProcessed: 0,
     itemsSkipped: 0, // Kept for API compatibility, always 0
@@ -71,17 +72,28 @@ export class BatchSchemaProcessor {
    * Check if there are more items to process.
    */
   hasWork(): boolean {
-    return this.workQueue.length > 0;
+    return this.queueIndex < this.workQueue.length;
   }
 
   /**
    * Get the next work item from the queue.
    */
   getNextWork(): WorkItem | undefined {
-    const item = this.workQueue.shift();
+    if (this.queueIndex >= this.workQueue.length) {
+      return undefined;
+    }
+
+    const item = this.workQueue[this.queueIndex++];
     if (item) {
       this.stats.itemsProcessed++;
     }
+
+    // Once we've consumed the queue, reset to avoid unbounded growth
+    if (this.queueIndex >= this.workQueue.length) {
+      this.workQueue = [];
+      this.queueIndex = 0;
+    }
+
     return item;
   }
 
@@ -122,6 +134,7 @@ export class BatchSchemaProcessor {
    */
   reset(): void {
     this.workQueue = [];
+    this.queueIndex = 0;
     this.stats = {
       itemsProcessed: 0,
       itemsSkipped: 0,
@@ -134,6 +147,6 @@ export class BatchSchemaProcessor {
    * Get current queue size (for debugging/monitoring).
    */
   getQueueSize(): number {
-    return this.workQueue.length;
+    return this.workQueue.length - this.queueIndex;
   }
 }

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/ScopeTreeManager.ts

@@ -117,9 +117,10 @@ export class ScopeTreeManager {
    */
   findNode(name: string): ScopeTreeNode | undefined {
     const queue: ScopeTreeNode[] = [this.tree];
+    let index = 0;
 
-    while (queue.length > 0) {
-      const node = queue.shift()!;
+    while (index < queue.length) {
+      const node = queue[index++]!;
       if (node.name === name) {
         return node;
       }
@@ -135,9 +136,10 @@ export class ScopeTreeManager {
   getAllNames(): string[] {
     const names: string[] = [];
     const queue: ScopeTreeNode[] = [this.tree];
+    let index = 0;
 
-    while (queue.length > 0) {
-      const node = queue.shift()!;
+    while (index < queue.length) {
+      const node = queue[index++]!;
       names.push(node.name);
       queue.push(...node.children);
     }

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/cleanKnownObjectFunctions.ts

@@ -35,6 +35,24 @@ const getParts = (path: string): string[] => {
   return parts;
 };
 
+function buildPrefixParts(parts: string[]): string[] {
+  if (parts.length === 0) return [];
+  const prefixes = new Array(parts.length);
+  let current = '';
+  for (let i = 0; i < parts.length; i++) {
+    const part = parts[i];
+    if (i === 0) {
+      current = part;
+    } else if (part.startsWith('[') || part.startsWith('(')) {
+      current += part;
+    } else {
+      current += `.${part}`;
+    }
+    prefixes[i] = current;
+  }
+  return prefixes;
+}
+
 /**
  * Clear the module-level cache.
  * Call this between entity analyses to prevent unbounded memory growth.
@@ -133,6 +151,7 @@ function scrub(
 
   for (const keyPath of sortedKeys) {
     const keyParts = getParts(keyPath);
+    const prefixParts = buildPrefixParts(keyParts);
 
     if (
       IGNORE_CLASSES_AND_OBJECTS.find(
@@ -145,7 +164,7 @@ function scrub(
 
     // walk back through the prefixes: foo.bar.baz.qux → foo.bar.baz → foo.bar …
     for (let cut = keyParts.length - 1; cut > 0; --cut) {
-      const prefix = joinParenthesesAndArrays(keyParts.slice(0, cut));
+      const prefix = prefixParts[cut - 1];
       const rootKind = rootTable.get(prefix);
       if (!rootKind) continue;
 

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/cleanNonObjectFunctions.ts

@@ -178,17 +178,36 @@ export function clearCleanNonObjectFunctionsCache(): {
 // Maps parent prefix -> array of last parts (method names without parens)
 type PrefixIndex = Map<string, string[]>;
 
+function buildPrefixPartsWithRoot(parts: string[]): string[] {
+  const prefixes = new Array(parts.length + 1);
+  prefixes[0] = '';
+  let current = '';
+  for (let i = 0; i < parts.length; i++) {
+    const part = parts[i];
+    if (i === 0) {
+      current = part;
+    } else if (part.startsWith('[') || part.startsWith('(')) {
+      current += part;
+    } else {
+      current += `.${part}`;
+    }
+    prefixes[i + 1] = current;
+  }
+  return prefixes;
+}
+
 function buildPrefixIndex(allPaths: string[]): PrefixIndex {
   const index: PrefixIndex = new Map();
 
   for (const path of allPaths) {
     const parts = cachedSplit(path);
     if (parts.length < 1) continue;
+    const prefixes = buildPrefixPartsWithRoot(parts);
 
     // For each possible prefix length, add the last part to that prefix's entry
     // This allows lookup at any depth
     for (let prefixLen = 0; prefixLen < parts.length; prefixLen++) {
-      const prefix = joinParenthesesAndArrays(parts.slice(0, prefixLen));
+      const prefix = prefixes[prefixLen];
       const lastPart = parts[prefixLen]?.split('(')?.[0];
       if (lastPart) {
         let existing = index.get(prefix);
@@ -1085,29 +1104,32 @@ function clearAttributes(
     knownBasePaths.add(path);
   }
 
+  // OPTIMIZATION: Instead of O(keys × basePaths), use O(keys × keyDepth) where keyDepth is small
+  // For each key, find all its prefixes and check if any is a known basePath (O(1) Set lookup)
   for (const key in mapping) {
     const keyParts = cachedSplit(key);
-    for (const basePath of knownBasePaths) {
-      const basePathParts = cachedSplit(basePath);
-      if (keyParts.length <= basePathParts.length) continue;
-      if (
-        joinParenthesesAndArrays(keyParts.slice(0, basePathParts.length)) !==
-        basePath
-      )
-        continue;
+    if (keyParts.length <= 1) continue; // No possible parent basePath
+
+    // Check each prefix of the key to see if it's a known basePath
+    // Start from shortest prefix (most likely to match) for early exit
+    for (let prefixLen = 1; prefixLen < keyParts.length; prefixLen++) {
+      const prefix = joinParenthesesAndArrays(keyParts.slice(0, prefixLen));
+      if (!knownBasePaths.has(prefix)) continue;
+
+      // Found a matching basePath - now check the conditions
+      const basePath = prefix;
+
       // Preserve functionCallReturnValue for non-string types, UNLESS
       // the method is a primitive-returning method (like some, every, includes, indexOf)
       // whose return values are not mockable
-      const methodPart = keyParts[basePathParts.length];
+      const methodPart = keyParts[prefixLen];
       const methodName = methodPart?.split('(')[0];
       const isPrimitiveReturning = primitiveReturningMethodsSet.has(
         methodName ?? '',
       );
       if (
         mapping[basePath] !== 'string' &&
-        keyParts[basePathParts.length + 1]?.startsWith(
-          'functionCallReturnValue',
-        ) &&
+        keyParts[prefixLen + 1]?.startsWith('functionCallReturnValue') &&
         !isPrimitiveReturning
       )
         continue;

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/convertTypeAnnotationsToValues.ts

@@ -0,0 +1,160 @@
+/**
+ * Converts type annotation strings that appear as values in mockData to actual values.
+ *
+ * The LLM sometimes echoes the schema type annotation as the actual value.
+ * For example, if the schema says { filePath: "string | undefined" },
+ * the LLM might return { filePath: "string | undefined" } instead of
+ * generating an actual value like { filePath: "/path/to/file.tsx" }.
+ *
+ * This causes runtime errors when the code tries to use these "values"
+ * as actual data (e.g., calling .length on the string "string | undefined").
+ *
+ * This function detects type annotation strings and converts them to
+ * appropriate default values:
+ * - "string" → ""
+ * - "number" → 0
+ * - "boolean" → false
+ * - "function" → () => {}
+ * - "Date" → new Date()
+ * - "T | undefined" → undefined
+ * - "T | null" → null
+ * - "string[]" → []
+ */
+
+// Pattern to detect if a string looks like a TypeScript type annotation
+// This matches common type patterns:
+// - Primitive types: string, number, boolean, function, Date, object, any, unknown
+// - Union types: string | undefined, number | null, T | null | undefined
+// - Array types: string[], number[], T[]
+// - Optional types: T?
+const TYPE_ANNOTATION_PATTERN =
+  /^(string|number|boolean|function|Date|object|any|unknown|void|never|null|undefined|array)(\s*\|\s*(string|number|boolean|function|Date|object|any|unknown|void|never|null|undefined|array))*(\[\])?$/i;
+
+/**
+ * Check if a string value looks like a TypeScript type annotation.
+ */
+function isTypeAnnotation(value: string): boolean {
+  if (typeof value !== 'string') return false;
+
+  const trimmed = value.trim();
+
+  // Quick check: type annotations are typically short
+  if (trimmed.length > 100) return false;
+
+  // Check against the pattern
+  if (TYPE_ANNOTATION_PATTERN.test(trimmed)) return true;
+
+  // Also check for common patterns that the regex might miss
+  const commonTypePatterns = [
+    'string',
+    'number',
+    'boolean',
+    'function',
+    'object',
+    'array',
+    'Date',
+    'any',
+    'unknown',
+    'void',
+    'never',
+    'null',
+    'undefined',
+    'string | undefined',
+    'string | null',
+    'string | null | undefined',
+    'number | undefined',
+    'number | null',
+    'number | null | undefined',
+    'boolean | undefined',
+    'boolean | null',
+    'boolean | null | undefined',
+    'Date | undefined',
+    'Date | null',
+    'function | undefined',
+    'object | undefined',
+    'object | null',
+    'string[]',
+    'number[]',
+    'boolean[]',
+    'any[]',
+    'string | number',
+    'string | number | undefined',
+  ];
+
+  return commonTypePatterns.includes(trimmed);
+}
+
+/**
+ * Convert a type annotation string to an appropriate default value.
+ */
+function typeAnnotationToValue(typeAnnotation: string): unknown {
+  const trimmed = typeAnnotation.trim().toLowerCase();
+
+  // Handle undefined types first (they take precedence)
+  if (trimmed.includes('| undefined') || trimmed.includes('undefined |')) {
+    return undefined;
+  }
+
+  // Handle null types
+  if (trimmed.includes('| null') || trimmed.includes('null |')) {
+    return null;
+  }
+
+  // Handle array types
+  if (trimmed.endsWith('[]') || trimmed === 'array') {
+    return [];
+  }
+
+  // Handle primitive types
+  if (trimmed === 'string') return '';
+  if (trimmed === 'number') return 0;
+  if (trimmed === 'boolean') return false;
+  if (trimmed === 'function') return () => {};
+  if (trimmed === 'date') return new Date();
+  if (trimmed === 'object') return {};
+  if (trimmed === 'null') return null;
+  if (trimmed === 'undefined') return undefined;
+  if (trimmed === 'any' || trimmed === 'unknown') return undefined;
+  if (trimmed === 'void' || trimmed === 'never') return undefined;
+
+  // Handle union types - prefer the first concrete type
+  if (trimmed.includes(' | ')) {
+    const parts = trimmed.split(' | ').map((p) => p.trim());
+    for (const part of parts) {
+      if (part === 'string') return '';
+      if (part === 'number') return 0;
+      if (part === 'boolean') return false;
+    }
+  }
+
+  // Fallback: return undefined for unknown type patterns
+  return undefined;
+}
+
+/**
+ * Recursively convert type annotation strings to values in a mockData object.
+ * Mutates the object in place.
+ */
+export default function convertTypeAnnotationsToValues(
+  mockData: Record<string, unknown>,
+): void {
+  for (const [key, value] of Object.entries(mockData)) {
+    if (typeof value === 'string' && isTypeAnnotation(value)) {
+      mockData[key] = typeAnnotationToValue(value);
+    } else if (Array.isArray(value)) {
+      // Recursively process array elements
+      for (const item of value) {
+        if (item !== null && typeof item === 'object' && !Array.isArray(item)) {
+          convertTypeAnnotationsToValues(item as Record<string, unknown>);
+        }
+      }
+    } else if (
+      value !== null &&
+      typeof value === 'object' &&
+      !Array.isArray(value)
+    ) {
+      // Recursively process nested objects
+      convertTypeAnnotationsToValues(value as Record<string, unknown>);
+    }
+  }
+}

package/analyzer-template/packages/ai/src/lib/dataStructure/helpers/deduplicateFunctionSchemas.ts

@@ -77,42 +77,52 @@ export default function deduplicateFunctionSchemas(
 
       // If multiple variants with arguments exist, find the one with the most robust structure
       if (withArgs.length > 1) {
-        // Count the number of related schema keys for each variant
-        const variantDepths = withArgs.map((variant) => {
-          // Find the end of the function call (the closing parenthesis)
+        // Pre-compute function call prefixes for all variants
+        const variantPrefixes = withArgs.map((variant) => {
           const closingParenIndex = variant.lastIndexOf(')');
-          const functionCallPrefix =
-            closingParenIndex > -1
-              ? variant.slice(0, closingParenIndex + 1)
-              : variant;
-
-          const relatedKeys = Object.keys(schema).filter((k) => {
-            // Check if this key is a child of the variant (starts with the function call)
-            return (
-              k.startsWith(functionCallPrefix + '.') ||
-              k.startsWith(functionCallPrefix + '[')
-            );
-          });
-
           return {
             key: variant,
-            depth: relatedKeys.length,
-            maxDepth: Math.max(
-              0,
-              ...relatedKeys.map((k) => {
-                // Count the nesting depth by counting dots and brackets after the variant
-                const remaining = k.slice(functionCallPrefix.length);
-                return (
-                  (remaining.match(/\./g) || []).length +
-                  (remaining.match(/\[/g) || []).length
-                );
-              }),
-            ),
+            prefix:
+              closingParenIndex > -1
+                ? variant.slice(0, closingParenIndex + 1)
+                : variant,
+            depth: 0,
+            maxDepth: 0,
           };
         });
 
+        // Single pass through schema keys to count related keys for ALL variants at once.
+        // This is O(schemaSize × variants) but done in one pass instead of
+        // O(variants × schemaSize) with separate passes per variant.
+        const schemaKeys = Object.keys(schema);
+        for (const k of schemaKeys) {
+          for (const variantInfo of variantPrefixes) {
+            // Check if key starts with prefix followed by . or [
+            const prefixLen = variantInfo.prefix.length;
+            if (k.length > prefixLen) {
+              const nextChar = k[prefixLen];
+              if (
+                (nextChar === '.' || nextChar === '[') &&
+                k.startsWith(variantInfo.prefix)
+              ) {
+                variantInfo.depth++;
+                // Count nesting depth by counting . and [ characters (faster than regex)
+                const remaining = k.slice(prefixLen);
+                let nestingDepth = 0;
+                for (let i = 0; i < remaining.length; i++) {
+                  const c = remaining[i];
+                  if (c === '.' || c === '[') nestingDepth++;
+                }
+                if (nestingDepth > variantInfo.maxDepth) {
+                  variantInfo.maxDepth = nestingDepth;
+                }
+              }
+            }
+          }
+        }
+
         // Sort by depth (number of related keys) and then by max depth (nesting level)
-        variantDepths.sort((a, b) => {
+        variantPrefixes.sort((a, b) => {
           if (b.depth !== a.depth) {
             return b.depth - a.depth;
           }
@@ -120,7 +130,7 @@ export default function deduplicateFunctionSchemas(
         });
 
         // Keep the one with the most robust structure, remove others
-        const keepVariant = variantDepths[0].key;
+        const keepVariant = variantPrefixes[0].key;
         withArgs.forEach((variant) => {
           if (variant !== keepVariant) {
             keysToRemove.add(variant);