cursor2api/test/perf-diag-results.json

[
  {
    "name": "① 简短问答",
    "direct": {
      "totalMs": 3746,
      "headerMs": 911,
      "ttfbMs": 2711,
      "streamMs": 1035,
      "textLength": 194,
      "chunkCount": 29,
      "text": "Quicksort has an average-case time complexity of O(n log n) and a worst-case time complexity of O(n²), which occurs when the pivot selection consistently results in highly unbalanced partitions."
    },
    "proxy": {
      "totalMs": 2784,
      "headerMs": 19,
      "ttfbMs": 2784,
      "firstContentMs": 2784,
      "streamMs": 0,
      "textLength": 242,
      "chunkCount": 1,
      "text": "Quicksort has an average-case time complexity of O(n log n) and a worst-case time complexity of O(n²), which occurs when the pivot selection consistently results in unbalanced partitions (e.g., already sorted input with a naive pivot choice)."
    }
  },
  {
    "name": "② 中等代码",
    "direct": {
      "totalMs": 18302,
      "headerMs": 1058,
      "ttfbMs": 2076,
      "streamMs": 16226,
      "textLength": 4471,
      "chunkCount": 515,
      "text": "Here is a clean, well-documented Python function to validate an IPv4 address:\n\n```python\ndef is_valid_ipv4(address: str) -> bool:\n    \"\"\"\n    Check if a given string is a valid IPv4 address.\n\n    An IPv4 address is considered valid if it meets all of the following criteria:\n        - Consists of exactly 4 octets separated by periods ('.')\n        - Each octet is a decimal integer in the range 0–255 (inclusive)\n        - No octet contains leading zeros (e.g., '01', '00' are invalid)\n        - No extra whitespace or empty octets are present\n\n    Args:\n        address (str): The string to validate.\n\n    Returns:\n        bool: True if the string is a valid IPv4 address, False otherwise.\n\n    Examples:\n        >>> is_valid_ipv4(\"192.168.1.1\")\n        True\n        >>> is_valid_ipv4(\"255.255.255.255\")\n        True\n        >>> is_valid_ipv4(\"0.0.0.0\")\n        True\n        >>> is_valid_ipv4(\"256.100.50.25\")\n        False\n        >>> is_valid_ipv4(\"192.168.1\")\n        False\n        >>> is_valid_ipv4(\"192.168.01.1\")\n        False\n        >>> is_valid_ipv4(\"192.168.1.1.1\")\n        False\n        >>> is_valid_ipv4(\"\")\n        False\n        >>> is_valid_ipv4(\"abc.def.ghi.jkl\")\n        False\n    \"\"\"\n    if not isinstance(address, str) or not address:\n        return False\n\n    octets = address.split(\".\")\n\n    if len(octets) != 4:\n        return False\n\n    for octet in octets:\n        # Reject empty octets or those with leading zeros (except \"0\" itself)\n        if not octet or (octet.startswith(\"0\") and len(octet) > 1):\n            return False\n\n        # Reject non-numeric octets\n        if not octet.isdigit():\n            return False\n\n        # Reject octets outside the valid range\n        if not (0 <= int(octet) <= 255):\n            return False\n\n    return True\n\n\n# --- Quick self-test ---\nif __name__ == \"__main__\":\n    test_cases = [\n        (\"192.168.1.1\",     True),\n        (\"255.255.255.255\", True),\n        (\"0.0.0.0\",         True),\n        (\"10.0.0.1\",        True),\n        (\"256.100.50.25\",   False),  # octet > 255\n        (\"192.168.1\",       False),  # only 3 octets\n        (\"192.168.01.1\",    False),  # leading zero\n        (\"192.168.1.1.1\",   False),  # 5 octets\n        (\"\",                False),  # empty string\n        (\"abc.def.ghi.jkl\", False),  # non-numeric\n        (\"192.168.1.-1\",    False),  # negative number\n        (\"192.168. 1.1\",    False),  # whitespace\n        (\" 192.168.1.1\",    False),  # leading whitespace\n        (\"192.168.1.1 \",    False),  # trailing whitespace\n    ]\n\n    all_passed = True\n    for ip, expected in test_cases:\n        result = is_valid_ipv4(ip)\n        status = \"PASS\" if result == expected else \"FAIL\"\n        if status == \"FAIL\":\n            all_passed = False\n        print(f\"[{status}] is_valid_ipv4({ip!r:25}) => {result} (expected {expected})\")\n\n    print(\"\\nAll tests passed!\" if all_passed else \"\\nSome tests FAILED.\")\n```\n\n**Sample output:**\n\n```\n[PASS] is_valid_ipv4('192.168.1.1'           ) => True (expected True)\n[PASS] is_valid_ipv4('255.255.255.255'        ) => True (expected True)\n[PASS] is_valid_ipv4('0.0.0.0'               ) => True (expected True)\n[PASS] is_valid_ipv4('10.0.0.1'              ) => True (expected True)\n[PASS] is_valid_ipv4('256.100.50.25'         ) => False (expected False)\n[PASS] is_valid_ipv4('192.168.1'             ) => False (expected False)\n[PASS] is_valid_ipv4('192.168.01.1'          ) => False (expected False)\n[PASS] is_valid_ipv4('192.168.1.1.1'         ) => False (expected False)\n[PASS] is_valid_ipv4(''                      ) => False (expected False)\n[PASS] is_valid_ipv4('abc.def.ghi.jkl'       ) => False (expected False)\n[PASS] is_valid_ipv4('192.168.1.-1'          ) => False (expected False)\n[PASS] is_valid_ipv4('192.168. 1.1'          ) => False (expected False)\n[PASS] is_valid_ipv4(' 192.168.1.1'          ) => False (expected False)\n[PASS] is_valid_ipv4('192.168.1.1 '          ) => False (expected False)\n\nAll tests passed!\n```\n\n---\n\n**Key design decisions explained:**\n\n| Check | Reason |\n|---|---|\n| `isinstance(address, str)` | Guards against non-string inputs like `None` or integers |\n| `len(octets) != 4` | Ensures exactly 4 dot-separated parts |\n| Leading zero check | `\"01\"` is ambiguous (octal notation) and not valid in IPv4 |\n| `isdigit()` before `int()` | Rejects negatives (`\"-1\"`), spaces, and letters without risking a `ValueError` |\n| `0 <= int(octet) <= 255` | Enforces the valid octet range per the IPv4 spec |"
    },
    "proxy": {
      "totalMs": 16276,
      "headerMs": 5,
      "ttfbMs": 2737,
      "firstContentMs": 2737,
      "streamMs": 13539,
      "textLength": 3777,
      "chunkCount": 431,
      "text": "Here is a clean, well-documented Python function to validate an IPv4 address:\n\n```python\ndef is_valid_ipv4(address: str) -> bool:\n    \"\"\"\n    Check if a given string is a valid IPv4 address.\n\n    An IPv4 address is considered valid if it meets ALL of the following criteria:\n      - Consists of exactly 4 parts separated by dots ('.')\n      - Each part is a valid integer (no leading zeros, no letters or symbols)\n      - Each integer is in the range 0 to 255 (inclusive)\n      - The string is not empty and contains no extra whitespace or characters\n\n    Args:\n        address (str): The string to validate as an IPv4 address.\n\n    Returns:\n        bool: True if the string is a valid IPv4 address, False otherwise.\n\n    Examples:\n        >>> is_valid_ipv4(\"192.168.1.1\")\n        True\n        >>> is_valid_ipv4(\"255.255.255.255\")\n        True\n        >>> is_valid_ipv4(\"0.0.0.0\")\n        True\n        >>> is_valid_ipv4(\"256.100.50.25\")\n        False\n        >>> is_valid_ipv4(\"192.168.1\")\n        False\n        >>> is_valid_ipv4(\"192.168.01.1\")   # Leading zero\n        False\n        >>> is_valid_ipv4(\"192.168.1.1.1\")  # Too many octets\n        False\n        >>> is_valid_ipv4(\"abc.def.ghi.jkl\")\n        False\n        >>> is_valid_ipv4(\"\")\n        False\n    \"\"\"\n    if not isinstance(address, str) or not address:\n        return False\n\n    parts = address.split(\".\")\n\n    if len(parts) != 4:\n        return False\n\n    for part in parts:\n        # Reject empty parts, parts with leading zeros, and non-numeric parts\n        if not part or not part.isdigit():\n            return False\n\n        # Reject leading zeros (e.g. \"01\", \"001\"), but allow \"0\" itself\n        if len(part) > 1 and part[0] == \"0\":\n            return False\n\n        # Reject values outside the valid octet range\n        if not (0 <= int(part) <= 255):\n            return False\n\n    return True\n```\n\n---\n\n**Key design decisions explained:**\n\n| Check | Reason |\n|---|---|\n| `isinstance(address, str)` | Guards against non-string inputs like `None` or integers |\n| `split(\".\")` produces exactly 4 parts | Enforces the 4-octet structure |\n| `part.isdigit()` | Rejects letters, symbols, spaces, and negative signs (e.g. `\"-1\"`) |\n| Leading zero check | `\"01\"` or `\"001\"` are ambiguous/invalid in standard IPv4 notation |\n| `0 <= int(part) <= 255` | Each octet must be within the valid byte range |\n\n---\n\n**Quick test run:**\n\n```python\ntest_cases = [\n    (\"192.168.1.1\",     True),\n    (\"0.0.0.0\",         True),\n    (\"255.255.255.255\", True),\n    (\"256.0.0.1\",       False),  # Out of range\n    (\"192.168.01.1\",    False),  # Leading zero\n    (\"192.168.1\",       False),  # Too few octets\n    (\"192.168.1.1.1\",   False),  # Too many octets\n    (\"192.168.1.-1\",    False),  # Negative number\n    (\"abc.def.ghi.jkl\", False),  # Non-numeric\n    (\"\",                False),  # Empty string\n    (\"1.2.3. 4\",        False),  # Whitespace\n]\n\nfor ip, expected in test_cases:\n    result = is_valid_ipv4(ip)\n    status = \"PASS\" if result == expected else \"FAIL\"\n    print(f\"[{status}] is_valid_ipv4({ip!r:20}) => {result}\")\n```\n\n**Output:**\n```\n[PASS] is_valid_ipv4('192.168.1.1'      ) => True\n[PASS] is_valid_ipv4('0.0.0.0'          ) => True\n[PASS] is_valid_ipv4('255.255.255.255'  ) => True\n[PASS] is_valid_ipv4('256.0.0.1'        ) => False\n[PASS] is_valid_ipv4('192.168.01.1'     ) => False\n[PASS] is_valid_ipv4('192.168.1'        ) => False\n[PASS] is_valid_ipv4('192.168.1.1.1'    ) => False\n[PASS] is_valid_ipv4('192.168.1.-1'     ) => False\n[PASS] is_valid_ipv4('abc.def.ghi.jkl'  ) => False\n[PASS] is_valid_ipv4(''                 ) => False\n[PASS] is_valid_ipv4('1.2.3. 4'         ) => False\n```\n\nAll 11 cases pass. The function handles all common edge cases without relying on any external libraries or regex."
    }
  },
  {
    "name": "③ 长代码生成",
    "direct": {
      "totalMs": 38394,
      "headerMs": 861,
      "ttfbMs": 2590,
      "streamMs": 35804,
      "textLength": 11113,
      "chunkCount": 1371,
      "text": "Here is a complete, well-typed Binary Search Tree implementation in TypeScript:\n\n```typescript\n// ============================================================\n// Type Definitions\n// ============================================================\n\ninterface BSTNode<T> {\n  value: T;\n  left: BSTNode<T> | null;\n  right: BSTNode<T> | null;\n}\n\ntype Comparator<T> = (a: T, b: T) => number;\n\n// ============================================================\n// Node Factory\n// ============================================================\n\nfunction createNode<T>(value: T): BSTNode<T> {\n  return { value, left: null, right: null };\n}\n\n// ============================================================\n// Binary Search Tree Class\n// ============================================================\n\nclass BinarySearchTree<T> {\n  private root: BSTNode<T> | null = null;\n  private comparator: Comparator<T>;\n\n  /**\n   * @param comparator - Optional custom comparator function.\n   * Defaults to natural (ascending) ordering for numbers and strings.\n   * Return negative if a < b, 0 if a === b, positive if a > b.\n   */\n  constructor(comparator?: Comparator<T>) {\n    this.comparator =\n      comparator ??\n      ((a, b) => {\n        if (a < b) return -1;\n        if (a > b) return 1;\n        return 0;\n      });\n  }\n\n  // ──────────────────────────────────────────────────────────\n  // INSERT\n  // ──────────────────────────────────────────────────────────\n\n  /**\n   * Inserts a value into the BST. Duplicate values are ignored.\n   * Time complexity: O(h), where h is the height of the tree.\n   */\n  insert(value: T): this {\n    this.root = this._insertNode(this.root, value);\n    return this; // enables chaining: bst.insert(1).insert(2)\n  }\n\n  private _insertNode(node: BSTNode<T> | null, value: T): BSTNode<T> {\n    if (node === null) return createNode(value);\n\n    const cmp = this.comparator(value, node.value);\n\n    if (cmp < 0) {\n      node.left = this._insertNode(node.left, value);\n    } else if (cmp > 0) {\n      node.right = this._insertNode(node.right, value);\n    }\n    // cmp === 0: duplicate — do nothing\n\n    return node;\n  }\n\n  // ──────────────────────────────────────────────────────────\n  // SEARCH\n  // ──────────────────────────────────────────────────────────\n\n  /**\n   * Searches for a value in the BST.\n   * Returns the matching node, or null if not found.\n   * Time complexity: O(h)\n   */\n  search(value: T): BSTNode<T> | null {\n    return this._searchNode(this.root, value);\n  }\n\n  /** Convenience method — returns true/false instead of the node. */\n  contains(value: T): boolean {\n    return this.search(value) !== null;\n  }\n\n  private _searchNode(\n    node: BSTNode<T> | null,\n    value: T\n  ): BSTNode<T> | null {\n    if (node === null) return null;\n\n    const cmp = this.comparator(value, node.value);\n\n    if (cmp < 0) return this._searchNode(node.left, value);\n    if (cmp > 0) return this._searchNode(node.right, value);\n    return node; // found\n  }\n\n  // ──────────────────────────────────────────────────────────\n  // DELETE\n  // ──────────────────────────────────────────────────────────\n\n  /**\n   * Removes a value from the BST (if it exists).\n   * Uses the in-order successor strategy for nodes with two children.\n   * Time complexity: O(h)\n   */\n  delete(value: T): this {\n    this.root = this._deleteNode(this.root, value);\n    return this;\n  }\n\n  private _deleteNode(\n    node: BSTNode<T> | null,\n    value: T\n  ): BSTNode<T> | null {\n    if (node === null) return null; // value not found — no-op\n\n    const cmp = this.comparator(value, node.value);\n\n    if (cmp < 0) {\n      node.left = this._deleteNode(node.left, value);\n    } else if (cmp > 0) {\n      node.right = this._deleteNode(node.right, value);\n    } else {\n      // Node to delete found — three cases:\n\n      // Case 1: Leaf node (no children)\n      if (node.left === null && node.right === null) {\n        return null;\n      }\n\n      // Case 2: One child — replace node with its child\n      if (node.left === null) return node.right;\n      if (node.right === null) return node.left;\n\n      // Case 3: Two children — replace value with in-order successor\n      // (smallest value in the right subtree), then delete the successor.\n      const successor = this._findMin(node.right);\n      node.value = successor.value;\n      node.right = this._deleteNode(node.right, successor.value);\n    }\n\n    return node;\n  }\n\n  // ──────────────────────────────────────────────────────────\n  // TRAVERSALS\n  // ──────────────────────────────────────────────────────────\n\n  /**\n   * In-order traversal (Left → Root → Right).\n   * Visits nodes in ascending sorted order.\n   * Time complexity: O(n)\n   */\n  inorder(): T[] {\n    const result: T[] = [];\n    this._inorderTraverse(this.root, result);\n    return result;\n  }\n\n  private _inorderTraverse(node: BSTNode<T> | null, result: T[]): void {\n    if (node === null) return;\n    this._inorderTraverse(node.left, result);\n    result.push(node.value);\n    this._inorderTraverse(node.right, result);\n  }\n\n  /**\n   * Pre-order traversal (Root → Left → Right).\n   * Useful for serializing/copying the tree structure.\n   * Time complexity: O(n)\n   */\n  preorder(): T[] {\n    const result: T[] = [];\n    this._preorderTraverse(this.root, result);\n    return result;\n  }\n\n  private _preorderTraverse(node: BSTNode<T> | null, result: T[]): void {\n    if (node === null) return;\n    result.push(node.value);\n    this._preorderTraverse(node.left, result);\n    this._preorderTraverse(node.right, result);\n  }\n\n  /**\n   * Post-order traversal (Left → Right → Root).\n   * Useful for deleting or evaluating the entire tree.\n   * Time complexity: O(n)\n   */\n  postorder(): T[] {\n    const result: T[] = [];\n    this._postorderTraverse(this.root, result);\n    return result;\n  }\n\n  private _postorderTraverse(node: BSTNode<T> | null, result: T[]): void {\n    if (node === null) return;\n    this._postorderTraverse(node.left, result);\n    this._postorderTraverse(node.right, result);\n    result.push(node.value);\n  }\n\n  // ──────────────────────────────────────────────────────────\n  // UTILITY METHODS\n  // ──────────────────────────────────────────────────────────\n\n  /** Returns the node with the minimum value in the tree (or subtree). */\n  findMin(): BSTNode<T> | null {\n    if (this.root === null) return null;\n    return this._findMin(this.root);\n  }\n\n  private _findMin(node: BSTNode<T>): BSTNode<T> {\n    return node.left === null ? node : this._findMin(node.left);\n  }\n\n  /** Returns the node with the maximum value in the tree (or subtree). */\n  findMax(): BSTNode<T> | null {\n    if (this.root === null) return null;\n    return this._findMax(this.root);\n  }\n\n  private _findMax(node: BSTNode<T>): BSTNode<T> {\n    return node.right === null ? node : this._findMax(node.right);\n  }\n\n  /** Returns the height of the tree. An empty tree has height -1. */\n  height(): number {\n    return this._calcHeight(this.root);\n  }\n\n  private _calcHeight(node: BSTNode<T> | null): number {\n    if (node === null) return -1;\n    return 1 + Math.max(this._calcHeight(node.left), this._calcHeight(node.right));\n  }\n\n  /** Returns the total number of nodes. */\n  size(): number {\n    return this._countNodes(this.root);\n  }\n\n  private _countNodes(node: BSTNode<T> | null): number {\n    if (node === null) return 0;\n    return 1 + this._countNodes(node.left) + this._countNodes(node.right);\n  }\n\n  /** Removes all nodes from the tree. */\n  clear(): void {\n    this.root = null;\n  }\n\n  /** Returns true if the tree is empty. */\n  isEmpty(): boolean {\n    return this.root === null;\n  }\n}\n\n// ============================================================\n// Usage Examples\n// ============================================================\n\n// --- Numeric BST ---\nconst numTree = new BinarySearchTree<number>();\n\nnumTree.insert(10).insert(5).insert(15).insert(3).insert(7).insert(12).insert(18);\n//         10\n//        /  \\\n//       5    15\n//      / \\  /  \\\n//     3   7 12  18\n\nconsole.log(\"Inorder:  \", numTree.inorder());   // [3, 5, 7, 10, 12, 15, 18]\nconsole.log(\"Preorder: \", numTree.preorder());  // [10, 5, 3, 7, 15, 12, 18]\nconsole.log(\"Postorder:\", numTree.postorder()); // [3, 7, 5, 12, 18, 15, 10]\n\nconsole.log(\"Contains 7:\",  numTree.contains(7));   // true\nconsole.log(\"Contains 99:\", numTree.contains(99));  // false\nconsole.log(\"Search 15:\",   numTree.search(15));    // { value: 15, left: {...}, right: {...} }\n\nconsole.log(\"Min:\", numTree.findMin()?.value); // 3\nconsole.log(\"Max:\", numTree.findMax()?.value); // 18\nconsole.log(\"Height:\", numTree.height());      // 2\nconsole.log(\"Size:\",   numTree.size());        // 7\n\n// Delete a leaf\nnumTree.delete(3);\nconsole.log(\"After deleting 3:\", numTree.inorder()); // [5, 7, 10, 12, 15, 18]\n\n// Delete a node with one child\nnumTree.delete(5);\nconsole.log(\"After deleting 5:\", numTree.inorder()); // [7, 10, 12, 15, 18]\n\n// Delete a node with two children\nnumTree.delete(15);\nconsole.log(\"After deleting 15:\", numTree.inorder()); // [7, 10, 12, 18]\n\n// --- String BST ---\nconst strTree = new BinarySearchTree<string>();\nstrTree.insert(\"banana\").insert(\"apple\").insert(\"cherry\").insert(\"date\");\nconsole.log(\"String BST inorder:\", strTree.inorder()); // [\"apple\", \"banana\", \"cherry\", \"date\"]\n\n// --- Custom Object BST (with comparator) ---\ninterface Person {\n  name: string;\n  age: number;\n}\n\nconst personTree = new BinarySearchTree<Person>(\n  (a, b) => a.age - b.age // sort by age\n);\n\npersonTree.insert({ name: \"Alice\", age: 30 });\npersonTree.insert({ name: \"Bob\",   age: 25 });\npersonTree.insert({ name: \"Carol\", age: 35 });\n\nconsole.log(\n  \"People by age:\",\n  personTree.inorder().map((p) => `${p.name}(${p.age})`)\n); // [\"Bob(25)\", \"Alice(30)\", \"Carol(35)\"]\n```\n\n---\n\nHere is a breakdown of the key design decisions:\n\n**Generic with a custom comparator**\nThe class is `BinarySearchTree<T>`, so it works with numbers, strings, or any custom object. A default comparator handles primitives automatically; pass a custom one for complex types.\n\n**Delete uses the in-order successor strategy**\nWhen deleting a node with two children, it finds the smallest node in the right subtree (in-order successor), copies its value up, then deletes the successor — this preserves BST ordering with minimal restructuring.\n\n**Method chaining**\n`insert` and `delete` return `this`, so you can chain calls like `bst.insert(1).insert(2).delete(1)`.\n\n**All three traversals included**\n- `inorder` (L → Root → R): produces sorted output, the most commonly needed\n- `preorder` (Root → L → R): useful for serialization/cloning\n- `postorder` (L → R → Root): useful for safe deletion or expression evaluation\n\n**Time complexities**\n\n| Method | Average | Worst (skewed tree) |\n|---|---|---|\n| `insert` | O(log n) | O(n) |\n| `search` / `contains` | O(log n) | O(n) |\n| `delete` | O(log n) | O(n) |\n| `inorder` / `preorder` / `postorder` | O(n) | O(n) |\n| `findMin` / `findMax` | O(log n) | O(n) |\n| `height` / `size` | O(n) | O(n) |\n\nIf you need guaranteed O(log n) performance regardless of insertion order, consider extending this into a self-balancing tree like an AVL or Red-Black tree."
    },
    "proxy": {
      "totalMs": 27861,
      "headerMs": 4,
      "ttfbMs": 3324,
      "firstContentMs": 3324,
      "streamMs": 24537,
      "textLength": 7426,
      "chunkCount": 940,
      "text": "Here is a complete, well-typed Binary Search Tree implementation in TypeScript:\n\n```typescript\n// ─── Type Definitions ────────────────────────────────────────────────────────\n\ntype Nullable<T> = T | null;\n\ninterface BSTNode<T> {\n  value: T;\n  left: Nullable<BSTNode<T>>;\n  right: Nullable<BSTNode<T>>;\n}\n\ninterface IBinarySearchTree<T> {\n  insert(value: T): void;\n  delete(value: T): void;\n  search(value: T): Nullable<BSTNode<T>>;\n  inorder(): T[];\n  isEmpty(): boolean;\n  getRoot(): Nullable<BSTNode<T>>;\n}\n\n// ─── Node Factory ─────────────────────────────────────────────────────────────\n\nfunction createNode<T>(value: T): BSTNode<T> {\n  return { value, left: null, right: null };\n}\n\n// ─── Comparator Type ─────────────────────────────────────────────────────────\n\n/**\n * Returns:\n *  -1 (or any negative) if a < b\n *   0                   if a === b\n *   1 (or any positive) if a > b\n */\ntype Comparator<T> = (a: T, b: T) => number;\n\nconst defaultComparator = <T>(a: T, b: T): number => {\n  if (a < b) return -1;\n  if (a > b) return 1;\n  return 0;\n};\n\n// ─── BinarySearchTree Class ───────────────────────────────────────────────────\n\nclass BinarySearchTree<T> implements IBinarySearchTree<T> {\n  private root: Nullable<BSTNode<T>> = null;\n  private readonly compare: Comparator<T>;\n\n  constructor(comparator: Comparator<T> = defaultComparator) {\n    this.compare = comparator;\n  }\n\n  // ── Public API ─────────────────────────────────────────────────────────────\n\n  isEmpty(): boolean {\n    return this.root === null;\n  }\n\n  getRoot(): Nullable<BSTNode<T>> {\n    return this.root;\n  }\n\n  /**\n   * Inserts a value into the BST.\n   * Duplicate values are ignored.\n   * Time complexity: O(h) where h is the height of the tree.\n   */\n  insert(value: T): void {\n    this.root = this.insertNode(this.root, value);\n  }\n\n  /**\n   * Removes a value from the BST if it exists.\n   * Time complexity: O(h) where h is the height of the tree.\n   */\n  delete(value: T): void {\n    this.root = this.deleteNode(this.root, value);\n  }\n\n  /**\n   * Searches for a value and returns its node, or null if not found.\n   * Time complexity: O(h) where h is the height of the tree.\n   */\n  search(value: T): Nullable<BSTNode<T>> {\n    return this.searchNode(this.root, value);\n  }\n\n  /**\n   * Returns all values in sorted (ascending) order via in-order traversal.\n   * Time complexity: O(n).\n   */\n  inorder(): T[] {\n    const result: T[] = [];\n    this.inorderTraversal(this.root, result);\n    return result;\n  }\n\n  // ── Private Helpers ────────────────────────────────────────────────────────\n\n  private insertNode(node: Nullable<BSTNode<T>>, value: T): BSTNode<T> {\n    // Base case: found an empty slot — create and return a new node\n    if (node === null) {\n      return createNode(value);\n    }\n\n    const cmp = this.compare(value, node.value);\n\n    if (cmp < 0) {\n      // Value is smaller — go left\n      node.left = this.insertNode(node.left, value);\n    } else if (cmp > 0) {\n      // Value is larger — go right\n      node.right = this.insertNode(node.right, value);\n    }\n    // cmp === 0: duplicate, do nothing\n\n    return node;\n  }\n\n  private deleteNode(node: Nullable<BSTNode<T>>, value: T): Nullable<BSTNode<T>> {\n    // Base case: value not found\n    if (node === null) return null;\n\n    const cmp = this.compare(value, node.value);\n\n    if (cmp < 0) {\n      // Target is in the left subtree\n      node.left = this.deleteNode(node.left, value);\n    } else if (cmp > 0) {\n      // Target is in the right subtree\n      node.right = this.deleteNode(node.right, value);\n    } else {\n      // Found the node to delete — handle 3 cases:\n\n      // Case 1: Leaf node (no children)\n      if (node.left === null && node.right === null) {\n        return null;\n      }\n\n      // Case 2a: Only right child\n      if (node.left === null) {\n        return node.right;\n      }\n\n      // Case 2b: Only left child\n      if (node.right === null) {\n        return node.left;\n      }\n\n      // Case 3: Two children\n      // Replace with in-order successor (smallest value in right subtree),\n      // then delete the successor from the right subtree.\n      const successor = this.findMin(node.right);\n      node.value = successor.value;\n      node.right = this.deleteNode(node.right, successor.value);\n    }\n\n    return node;\n  }\n\n  private searchNode(node: Nullable<BSTNode<T>>, value: T): Nullable<BSTNode<T>> {\n    if (node === null) return null;\n\n    const cmp = this.compare(value, node.value);\n\n    if (cmp < 0) return this.searchNode(node.left, value);\n    if (cmp > 0) return this.searchNode(node.right, value);\n    return node; // cmp === 0: found\n  }\n\n  private inorderTraversal(node: Nullable<BSTNode<T>>, result: T[]): void {\n    if (node === null) return;\n    this.inorderTraversal(node.left, result);\n    result.push(node.value);\n    this.inorderTraversal(node.right, result);\n  }\n\n  /**\n   * Returns the node with the minimum value in a given subtree.\n   * Assumes the subtree is non-null.\n   */\n  private findMin(node: BSTNode<T>): BSTNode<T> {\n    let current = node;\n    while (current.left !== null) {\n      current = current.left;\n    }\n    return current;\n  }\n}\n\n// ─── Usage Examples ───────────────────────────────────────────────────────────\n\n// --- Numbers ---\nconst numTree = new BinarySearchTree<number>();\n\nnumTree.insert(10);\nnumTree.insert(5);\nnumTree.insert(15);\nnumTree.insert(3);\nnumTree.insert(7);\nnumTree.insert(12);\nnumTree.insert(20);\n\nconsole.log(\"Inorder:\", numTree.inorder());\n// → [3, 5, 7, 10, 12, 15, 20]\n\nconsole.log(\"Search 7:\", numTree.search(7));\n// → { value: 7, left: null, right: null }\n\nconsole.log(\"Search 99:\", numTree.search(99));\n// → null\n\nnumTree.delete(5);  // Node with two children\nconsole.log(\"After deleting 5:\", numTree.inorder());\n// → [3, 7, 10, 12, 15, 20]\n\nnumTree.delete(15); // Node with two children\nconsole.log(\"After deleting 15:\", numTree.inorder());\n// → [3, 7, 10, 12, 20]\n\nnumTree.delete(3);  // Leaf node\nconsole.log(\"After deleting 3:\", numTree.inorder());\n// → [7, 10, 12, 20]\n\n// --- Strings (custom comparator) ---\nconst strTree = new BinarySearchTree<string>((a, b) => a.localeCompare(b));\n\nstrTree.insert(\"banana\");\nstrTree.insert(\"apple\");\nstrTree.insert(\"cherry\");\nstrTree.insert(\"date\");\n\nconsole.log(\"String tree inorder:\", strTree.inorder());\n// → [\"apple\", \"banana\", \"cherry\", \"date\"]\n```\n\n---\n\n## Key Design Decisions\n\n| Decision | Rationale |\n|---|---|\n| **Generic `<T>`** | Works with numbers, strings, or any custom type |\n| **Custom `Comparator<T>`** | Makes the tree flexible for non-primitive types (e.g., objects) |\n| **Interface `IBinarySearchTree<T>`** | Enforces the public API contract and is easy to extend or mock |\n| **`Nullable<T>` alias** | Cleaner than repeating `T \\| null` everywhere |\n| **In-order successor for delete** | Standard, correct approach for two-child deletion |\n| **Duplicate values ignored** | Common BST convention; easily changed to allow duplicates |\n| **Recursive implementation** | Clear and readable; for very deep trees, an iterative version avoids stack overflow |\n\n## Complexity Summary\n\n| Operation | Average Case | Worst Case (skewed tree) |\n|---|---|---|\n| `insert` | O(log n) | O(n) |\n| `delete` | O(log n) | O(n) |\n| `search` | O(log n) | O(n) |\n| `inorder` | O(n) | O(n) |\n\nFor guaranteed O(log n) worst-case performance, consider extending this into a self-balancing tree such as an AVL tree or Red-Black tree."
    }
  }
]