miii-cli 0.2.1 → 0.2.3

package/README.md

@@ -1,11 +1,11 @@
-# miii-cli
+# miii
 
-> Local AI coding assistant for your terminal. No cloud. No API keys. No latency.
+> A local AI coding assistant that actually works. No cloud. No Python. No API keys required.
 
 ```
-╭─────────────────────────────────────────────────────────╮
-│  miii — Claude Code-level workflows, local models only  │
-╰─────────────────────────────────────────────────────────╯
+╭──────────────────────────────────────────────────────────────╮
+│  miii — Claude Code-level workflows, fully offline           │
+╰──────────────────────────────────────────────────────────────╯
 ```
 
 [![npm version](https://img.shields.io/npm/v/miii-cli)](https://www.npmjs.com/package/miii-cli)
@@ -16,21 +16,27 @@
 
 ---
 
-## Why I built this
+## Why miii exists
 
-I couldn't find a local CLI AI tool that actually worked well. The ones that existed were either too clunky to set up, required cloud APIs, or had terminal output that was genuinely painful to read — weird formatting, broken renders, text that ran together. I wanted something that felt as clean as Claude Code but ran entirely on local models. So I built miii.
+Every local AI coding tool is either too clunky to set up, requires cloud APIs, or has terminal output that's genuinely painful to read. miii is what happens when you build a local coding assistant that takes UX seriously — real Ink-based terminal UI, automatic git context, multi-file refactors with task queues, and a context compactor that keeps local models on-track.
 
----
+Your code never leaves your machine.
 
-## What is miii?
+---
 
-`miii` is a terminal-native AI coding assistant powered by local models via [Ollama](https://ollama.com) or any OpenAI-compatible API (LM Studio, vLLM, Groq, Together, etc.).
+## What makes it different
 
-- **Runs 100% locally** — your code never leaves your machine
-- **File-aware** — type `@filename` to inject any file into context instantly
-- **Tool-enabled** — reads, writes, edits, and runs shell commands autonomously
-- **Session memory** — conversations persist across launches
-- **Extensible** — add custom slash commands via Markdown or TypeScript skill files
+| Feature | miii | aider | shell_gpt | open-interpreter |
+|---|---|---|---|---|
+| Ink terminal UI | ✅ | ❌ | ❌ | ❌ |
+| Zero Python | ✅ | ❌ | ❌ | ❌ |
+| Auto git context | ✅ | ✅ | ❌ | ❌ |
+| Context compaction | ✅ | ✅ | ❌ | ❌ |
+| Multi-file refactor queue | ✅ | partial | ❌ | ❌ |
+| `.miiiignore` | ✅ | ✅ | ❌ | ❌ |
+| Session persistence | ✅ | ❌ | ❌ | ❌ |
+| Planning mode | ✅ | ❌ | ❌ | ❌ |
+| Bundle size | 176K | ~50MB | ~40MB | ~100MB |
 
 ---
 
@@ -47,112 +53,206 @@ npm install -g miii-cli
 ## Quick start
 
 ```bash
-# Make sure Ollama is running
+# Start Ollama
 ollama serve
 
-# Start miii
+# Pull a model
+ollama pull qwen2.5-coder:7b
+
+# Launch miii
 miii
 ```
 
-On launch, miii opens a model picker. Select a model and start coding.
+A model picker opens on launch. Select a model and start coding.
 
 ```bash
 miii                          # default session
-miii --model codellama        # specific model
+miii --model qwen2.5-coder    # specific model
 miii --session myproject      # named session
-miii -s work -m llama3.2      # short flags
+miii -s work -m codellama     # short flags
+```
+
+---
+
+## Auto git context
+
+miii automatically detects changed files via `git status` and injects their contents into the model's context — no `@file` needed for files you're actively working on.
+
+```
+❯ fix the type error in the auth middleware
+
+[auto-loaded 3 changed file(s)]
+```
+
+Only fires for code-related messages. Pure questions ("what is a closure?") skip the git scan entirely. Deduped — same files don't re-inject unless they change.
+
+Disable per-project:
+```json
+{ "gitContext": false }
 ```
 
 ---
 
 ## File context with `@`
 
-Type `@` anywhere in your message to fuzzy-search and inject project files into the model's context:
+Type `@` to fuzzy-search and inject any project file:
 
 ```
 ❯ review the auth logic in @src/auth/middleware.ts
-❯ refactor @src/utils/parser.ts to handle edge cases
+❯ what does @src/utils/parser.ts return when input is empty?
 ```
 
 Automatically excluded: `node_modules`, `dist`, `.git`, lock files, binaries, images.
 
 ---
 
+## `.miiiignore`
+
+Create `.miiiignore` in your project root to exclude files from `@` picker and auto-context:
+
+```
+# .miiiignore
+secrets/
+*.generated.ts
+fixtures/
+*.sql
+```
+
+Supports exact names, relative paths, and `*.ext` glob patterns.
+
+---
+
+## Git integration
+
+Full git workflow from the terminal:
+
+```
+/git status          show working tree
+/git diff            unstaged changes
+/git diff --staged   staged changes
+/git log             recent commits
+/git review          AI reviews current changes for bugs + improvements
+/git branch          list branches
+/git commit <msg>    stage all and commit
+```
+
+The model also has access to git tools directly — it can check status, read diffs, and commit as part of autonomous workflows.
+
+---
+
+## Multi-file refactor
+
+Describe a goal, miii plans and executes across multiple files with isolated context per file:
+
+```
+/refactor extract all database queries into a repository layer
+/refactor rename UserService to AccountService everywhere
+/refactor add input validation to all API route handlers
+```
+
+Uses a priority task queue (P0=blocking, P1=reads parallel, P2=writes sequential, P3=verify). Each file gets its own fresh context — the local model never loses the thread on large codebases.
+
+---
+
+## Planning mode
+
+Structured step-by-step planning before writing any code:
+
+```
+/plan add OAuth2 to this Express app
+/plan refactor the frontend to use React Query
+```
+
+In planning mode:
+```
+/plan:next        suggest next concrete steps
+/plan:breakdown   break topic into subtasks
+/plan:review      critique the plan so far
+/plan:done        exit planning mode
+```
+
+---
+
 ## Built-in commands
 
-Type `/` to open the command palette.
+Type `/` to open the command palette with fuzzy search.
 
 | Command | Description |
 |---|---|
-| `/models` | Switch or pull Ollama models |
+| `/model <name>` | Switch model mid-session — no restart needed |
+| `/models` | Open model picker, pull new Ollama models |
 | `/session <name>` | Switch to or create a named session |
 | `/sessions` | List all sessions with message counts |
+| `/new` | Start a fresh auto-named session |
 | `/clear` | Clear current session history |
-| `/mkdir <path>` | Create a folder (and any missing parents) |
-| `/touch <path>` | Create an empty file |
-| `/mv <from> <to>` | Move or rename a file or folder |
-| `/list` | Show loaded skills |
+| `/plan [topic]` | Enter planning mode |
+| `/refactor <goal>` | Multi-file AI refactor |
+| `/git <subcommand>` | Git commands (see above) |
+| `/list` | Show all loaded skills |
 | `/exit` | Exit miii |
 
 ---
 
 ## Built-in tools
 
-The model can call these tools automatically — no setup needed.
+The model calls these autonomously — reads, writes, edits, runs, tests, and commits on its own.
 
 | Tool | Description |
 |---|---|
 | `read_file` | Read any file |
 | `list_files` | List directory contents |
-| `edit_file` | Create or overwrite a file (auto-creates parent dirs) |
-| `create_folder` | Create a directory and any missing parents |
+| `create_file` | Create a new file (fails if exists) |
+| `edit_file` | Create or fully rewrite a file |
+| `patch_file` | Targeted string replacement — throws if match is ambiguous |
+| `delete_file` | Delete a file |
 | `move_file` | Move or rename a file or directory |
-| `delete_file` | Delete a file (requires confirmation) |
-| `run_command` | Run a shell command in the current directory |
+| `create_folder` | Create directory with parents |
+| `run_command` | Run shell command in cwd (30s timeout) |
+| `run_tests` | Run test suite, auto-detects jest/vitest/mocha from package.json |
+| `git_status` | Working tree status |
+| `git_diff` | Unstaged or staged diff (truncated at 8K) |
+| `git_log` | Recent commit history |
+| `git_commit` | Stage files and commit |
 
-Tool calls chain up to 6 hops deep — the model reads, edits, runs, and verifies on its own.
+Tool calls chain up to 6 hops — model reads, edits, runs tests, and verifies autonomously.
 
 ---
 
-## Thinking indicator
+## Context compaction
 
-While miii processes your request, a rotating sparkle icon and sarcastic phrase appear so you always know something's happening:
+Local models lose the thread after ~15-20 messages. miii auto-compacts context at threshold: keeps system prompt + original goal + tool result summary + last 6 exchanges. The session history is preserved on disk — only the LLM context window gets trimmed.
 
-```
-miii
-  ✦ staring into the abyss (it blinked)…
-```
+---
 
-The phrase rotates every 5 seconds. When a tool call is in flight, the display updates to show exactly which tool is running:
+## Thinking indicator
 
 ```
-miii
-  ⚙ running read_file…
+✦ staring into the abyss (it blinked)…   [0:12]
 
-miii
-  ⚙ running run_command…
+⚙ running patch_file…
+⚙ running run_tests…
 ```
 
-Each tool in a chain updates live as execution moves through them.
+Phrase rotates every 5 seconds. Tool name updates live as each call fires. Elapsed time shown throughout.
 
 ---
 
 ## Sessions
 
-Every conversation is saved and resumed automatically.
+Every conversation persists automatically.
 
 ```bash
 miii                          # resumes "default" session
 miii --session feature-auth   # resumes or creates "feature-auth"
 ```
 
-Sessions stored at `~/.config/miii/sessions/`.
+Sessions stored at `~/.config/miii/sessions/`. History capped at 100 messages in memory, full history on disk.
 
 ---
 
 ## Skills
 
-Skills are custom `/` commands. Create a Markdown file in `~/.config/miii/skills/`:
+Custom `/` commands via Markdown in `~/.config/miii/skills/`:
 
 ```markdown
 ---
@@ -164,32 +264,30 @@ Review the code I'm about to share. Look for bugs, edge cases, and improvements.
 Be direct and specific. No markdown.
 ```
 
-Use it:
-
 ```
 /review
 ```
 
-Skills can also be TypeScript files with an `execute` function for programmatic behavior.
+TypeScript skills with `execute` functions available for programmatic behavior.
 
 ---
 
 ## Configuration
 
-Config is loaded from (in order):
-1. `.miii.json` in the current directory
+Config loaded from (in order):
+1. `.miii.json` in current directory
 2. `~/.config/miii/config.json`
 
-**Ollama (default):**
+**Ollama:**
 ```json
 {
-  "model": "llama3.2",
+  "model": "qwen2.5-coder:7b",
   "provider": "ollama",
   "baseUrl": "http://localhost:11434"
 }
 ```
 
-**OpenAI-compatible API:**
+**OpenAI-compatible (LM Studio, vLLM, Groq, Together, etc.):**
 ```json
 {
   "model": "gpt-4o",
@@ -199,7 +297,17 @@ Config is loaded from (in order):
 }
 ```
 
-Works with LM Studio, vLLM, Groq, Together, and any other OpenAI-compatible server.
+**All options:**
+```json
+{
+  "model": "qwen2.5-coder:7b",
+  "provider": "ollama",
+  "baseUrl": "http://localhost:11434",
+  "apiKey": "",
+  "gitContext": true,
+  "systemPrompt": "optional override"
+}
+```
 
 ---
 
@@ -208,42 +316,25 @@ Works with LM Studio, vLLM, Groq, Together, and any other OpenAI-compatible serv
 | Key | Action |
 |---|---|
 | `enter` | Send message |
-| `ctrl+c` | Abort current request |
-| `ctrl+c` x2 | Exit miii |
-| `esc` | Close overlay or cancel |
 | `↑ / ↓` | Navigate command palette or file picker |
+| `esc` | Close overlay / abort request |
+| `ctrl+c` | Abort current request or exit |
 
 ---
 
 ## Security
 
-miii **0.1.5+** addresses the following OWASP issues:
-
 | Issue | Fix |
 |---|---|
-| Path traversal (A01) | All file tool operations restricted to cwd via `guardPath()` |
-| Path traversal (A01) | `@filename` references validated against `cwd` before reading |
-| Path traversal (A01) | `/mv`, `/mkdir`, `/touch` commands restricted to `cwd` |
-| Path traversal (A01) | Session names sanitized to alphanumeric + hyphens only |
-| Injection (A03) | `run_command` tool enforces a 30-second execution timeout |
-| Insecure deserialization (A08) | Config loading whitelists allowed keys; session data validated as array |
-| XML injection | File paths in context XML attributes are properly escaped |
-| Configurable API key | OpenAI-compatible provider token configurable via `apiKey` in config |
-
-`delete_file` requires explicit user confirmation (`y/n`) before executing.
+| Path traversal (OWASP A01) | All file operations restricted to cwd via `guardPath()` |
+| Path traversal (OWASP A01) | `@file` refs validated against cwd before reading |
+| Path traversal (OWASP A01) | Session names sanitized to alphanumeric + hyphens |
+| Injection (OWASP A03) | `run_command` enforces 30s execution timeout |
+| Insecure deserialization (OWASP A08) | Config whitelists allowed keys; session data validated as array |
 
 ---
 
-## What's new in 0.2.0
-
-- **No streaming** — responses delivered in full, no partial renders or flickering
-- **Thinking indicator** — rotating sarcastic phrases + sparkle icon while model processes
-- **Tool status** — live indicator when a tool call is in flight
-- **Status renamed** — internal status `streaming` → `thinking` throughout
-
----
-
-## Source
+## Build from source
 
 ```bash
 git clone https://github.com/maruakshay/miii-cli
@@ -252,3 +343,19 @@ npm install
 npm run build
 npm link
 ```
+
+---
+
+## What's in 0.2.x
+
+- **Auto git context** — changed files injected automatically, code-gated heuristic
+- **`.miiiignore`** — per-project file exclusions
+- **Multi-file refactor** — macro/micro task queue, isolated context per file
+- **Planning mode** — structured `/plan` workflow
+- **Git integration** — full git toolkit in commands and model tools
+- **`run_tests` tool** — model runs and retries tests autonomously
+- **`/model` live switch** — change models mid-session
+- **Context compaction** — auto-trim at 18 messages, keep goal + summary + recent
+- **Ambiguous patch detection** — `patch_file` throws on multiple matches instead of silently corrupting
+- **176K bundle** — removed dead workers, sourcemaps, unused deps (was 468K)
+- **Debounced session saves** — writes at most once per 2s, no I/O on every keypress

package/dist/config.js

@@ -28,4 +28,3 @@ export function loadConfig() {
     }
     return { ...defaults };
 }
-//# sourceMappingURL=config.js.map

package/dist/files/ops.js

@@ -27,6 +27,14 @@ const SKIP_NAMES = new Set([
     'poetry.lock', 'Gemfile.lock', 'composer.lock',
     '.DS_Store', 'Thumbs.db', '.env.local', 'LICENSE', 'LICENSE.md',
 ]);
+function loadIgnorePatterns(cwd) {
+    const p = join(cwd, '.miiiignore');
+    if (!existsSync(p))
+        return new Set();
+    return new Set(readFileSync(p, 'utf-8').split('\n')
+        .map(l => l.trim())
+        .filter(l => l && !l.startsWith('#')));
+}
 export function readFile(p) {
     if (!existsSync(p))
         return '';
@@ -46,9 +54,10 @@ export function moveFile(from, to) {
     mkdirSync(dirname(to), { recursive: true });
     renameSync(from, to);
 }
-export function listFiles(dir, recursive = false, cwd = process.cwd()) {
+export function listFiles(dir, recursive = false, cwd = process.cwd(), _ignore) {
     if (!existsSync(dir))
         return [];
+    const ignore = _ignore ?? loadIgnorePatterns(cwd);
     const entries = [];
     for (const name of readdirSync(dir)) {
         if (name.startsWith('.'))
@@ -70,12 +79,21 @@ export function listFiles(dir, recursive = false, cwd = process.cwd()) {
             continue;
         }
         const full = join(dir, name);
+        const rel = relative(cwd, full);
         const type = stat.isDirectory() ? 'dir' : 'file';
-        entries.push({ name, path: full, rel: relative(cwd, full), type, size: stat.isFile() ? stat.size : undefined });
+        // check ignore patterns: match by name, rel path, or name/ for dirs
+        if (ignore.size) {
+            const dirSuffix = type === 'dir' ? name + '/' : '';
+            if (ignore.has(name) || ignore.has(rel) || (dirSuffix && ignore.has(dirSuffix)))
+                continue;
+            // *.ext pattern
+            if ([...ignore].some(p => p.startsWith('*.') && name.endsWith(p.slice(1))))
+                continue;
+        }
+        entries.push({ name, path: full, rel, type, size: stat.isFile() ? stat.size : undefined });
         if (recursive && type === 'dir') {
-            entries.push(...listFiles(full, true, cwd));
+            entries.push(...listFiles(full, true, cwd, ignore));
         }
     }
     return entries;
 }
-//# sourceMappingURL=ops.js.map

package/dist/index.js

@@ -9,4 +9,3 @@ main().catch(err => {
     process.stderr.write(`fatal: ${err.message}\n`);
     process.exit(1);
 });
-//# sourceMappingURL=index.js.map

package/dist/init.js

@@ -31,4 +31,3 @@ export async function lazyInit() {
     const { waitUntilExit } = render(React.createElement(InputBar, { config, skills, cwd: process.cwd(), session: sessionName }), { exitOnCtrlC: false });
     await waitUntilExit();
 }
-//# sourceMappingURL=init.js.map

package/dist/llm/ollama.js

@@ -97,4 +97,3 @@ export function fmtSize(bytes) {
         return `${(bytes / 1e6).toFixed(0)}MB`;
     return `${(bytes / 1e3).toFixed(0)}KB`;
 }
-//# sourceMappingURL=ollama.js.map

package/dist/llm/stream.js

@@ -4,7 +4,7 @@ export async function chat(cfg) {
     return chatOllama(cfg);
 }
 async function chatOllama(cfg) {
-    const { model, messages, baseUrl, signal, onDone, onError } = cfg;
+    const { model, messages, baseUrl, signal, onDone, onError, onUsage } = cfg;
     try {
         const res = await fetch(`${baseUrl}/api/chat`, {
             method: 'POST',
@@ -17,6 +17,7 @@ async function chatOllama(cfg) {
             return;
         }
         const obj = await res.json();
+        onUsage?.(obj?.prompt_eval_count ?? 0, obj?.eval_count ?? 0);
         await onDone(obj?.message?.content ?? '');
     }
     catch (err) {
@@ -25,7 +26,7 @@ async function chatOllama(cfg) {
     }
 }
 async function chatOpenAI(cfg) {
-    const { model, messages, baseUrl, apiKey, signal, onDone, onError } = cfg;
+    const { model, messages, baseUrl, apiKey, signal, onDone, onError, onUsage } = cfg;
     try {
         const res = await fetch(`${baseUrl}/v1/chat/completions`, {
             method: 'POST',
@@ -38,6 +39,7 @@ async function chatOpenAI(cfg) {
             return;
         }
         const obj = await res.json();
+        onUsage?.(obj?.usage?.prompt_tokens ?? 0, obj?.usage?.completion_tokens ?? 0);
         await onDone(obj?.choices?.[0]?.message?.content ?? '');
     }
     catch (err) {
@@ -48,4 +50,3 @@ async function chatOpenAI(cfg) {
 function toError(e) {
     return e instanceof Error ? e : new Error(String(e));
 }
-//# sourceMappingURL=stream.js.map

package/dist/parser/stream-parser.js

@@ -1,4 +1,3 @@
-import { appendFileSync } from 'fs';
 const OPEN = '<tool_call>';
 const CLOSE = '</tool_call>';
 const CTAG_OPEN = '<content>';
@@ -7,13 +6,6 @@ const OLD_OPEN = '<old>';
 const OLD_CLOSE = '</old>';
 const NEW_OPEN = '<new>';
 const NEW_CLOSE = '</new>';
-const DEBUG_LOG = '/tmp/miii-debug.log';
-function dbg(msg) {
-    try {
-        appendFileSync(DEBUG_LOG, `[${new Date().toISOString()}] ${msg}\n`);
-    }
-    catch { }
-}
 // Fix literal newlines/tabs inside JSON string values — common LLM output mistake
 function sanitizeJson(s) {
     let result = '';
@@ -157,7 +149,6 @@ export class StreamParser {
                 this.buf = this.buf.slice(end + CLOSE.length);
                 this.inTool = false;
                 try {
-                    dbg(`raw block (${raw.length} chars): ${raw.slice(0, 300)}`);
                     // Extract named content blocks so file content never needs JSON escaping
                     const extraArgs = {};
                     let jsonPart = raw;
@@ -181,11 +172,9 @@ export class StreamParser {
                     extractBlock(NEW_OPEN, NEW_CLOSE, 'new');
                     const obj = parseToolJson(jsonPart);
                     obj.args = { ...(obj.args ?? {}), ...extraArgs };
-                    dbg(`parsed ok: name=${obj.name} args_keys=${Object.keys(obj.args).join(',')}`);
                     out.push({ type: 'tool_call', content: raw, toolName: obj.name, toolArgs: obj.args });
                 }
-                catch (e) {
-                    dbg(`parse FAILED: ${e} | raw: ${raw.slice(0, 300)}`);
+                catch {
                     out.push({ type: 'text', content: `${OPEN}${raw}${CLOSE}` });
                 }
             }
@@ -218,4 +207,3 @@ export class StreamParser {
         return out;
     }
 }
-//# sourceMappingURL=stream-parser.js.map

package/dist/sessions.js

@@ -51,4 +51,3 @@ export function deleteSession(name) {
     if (existsSync(p))
         unlinkSync(p);
 }
-//# sourceMappingURL=sessions.js.map

package/dist/skills/loader.js

@@ -126,4 +126,3 @@ export class SkillLoader {
         return [...new Set(this.map.values())];
     }
 }
-//# sourceMappingURL=loader.js.map

package/dist/tasks/compactor.js

@@ -0,0 +1,68 @@
+const COMPACT_THRESHOLD = 18; // compact when context exceeds this many messages
+const KEEP_RECENT = 6; // always keep last N messages verbatim
+export function shouldCompact(messages) {
+    return messages.length > COMPACT_THRESHOLD;
+}
+/**
+ * Compact context to keep local models on track during long refactors.
+ *
+ * Strategy:
+ *   1. Keep system prompt (index 0)
+ *   2. Keep first user message (original goal)
+ *   3. Summarise completed tool results in the middle into one message
+ *   4. Keep last KEEP_RECENT messages verbatim (model's working memory)
+ */
+export function compactContext(messages, goal) {
+    if (messages.length <= COMPACT_THRESHOLD)
+        return messages;
+    const system = messages[0]?.role === 'system' ? messages[0] : null;
+    const userGoal = messages.find(m => m.role === 'user' && !m.content.startsWith('['));
+    const anchorCount = (system ? 1 : 0) + (userGoal ? 1 : 0);
+    const middle = messages.slice(anchorCount, messages.length - KEEP_RECENT);
+    const recent = messages.slice(messages.length - KEEP_RECENT);
+    const toolResults = middle
+        .filter(m => m.role === 'user' && m.content.startsWith('Tool '))
+        .map(m => {
+        const lines = m.content.split('\n');
+        return `• ${lines[0]}`; // just the "Tool X result:" line
+    });
+    const assistantSummaries = middle
+        .filter(m => m.role === 'assistant' && m.content.trim().length > 0)
+        .map(m => m.content.slice(0, 120).replace(/\n/g, ' '));
+    const parts = [`[context compacted — ${middle.length} messages summarised]`];
+    if (goal)
+        parts.push(`Goal: ${goal}`);
+    if (toolResults.length)
+        parts.push(`Completed:\n${toolResults.join('\n')}`);
+    if (assistantSummaries.length)
+        parts.push(`Last reasoning: ${assistantSummaries.at(-1)}`);
+    const summary = { role: 'user', content: parts.join('\n\n') };
+    return [
+        ...(system ? [system] : []),
+        ...(userGoal ? [userGoal] : []),
+        summary,
+        ...recent,
+    ];
+}
+/**
+ * Build a fresh isolated context for a single-file edit step.
+ * Keeps context tiny — avoids cross-file noise polluting the model.
+ */
+export function fileEditContext(systemPrompt, goal, filePath, fileContent, instruction) {
+    return [
+        { role: 'system', content: systemPrompt },
+        {
+            role: 'user',
+            content: [
+                `Overall goal: ${goal}`,
+                ``,
+                `File to edit: ${filePath}`,
+                `<file>`,
+                fileContent,
+                `</file>`,
+                ``,
+                `Instruction: ${instruction}`,
+            ].join('\n'),
+        },
+    ];
+}