@moikas/code-audit-mcp
Version:
AI-powered code auditing via MCP using local Ollama models for security, performance, and quality analysis
248 lines ā¢ 9.03 kB
JavaScript
/**
* Models command - Manage AI models
*/
import chalk from 'chalk';
import ora from 'ora';
import { Listr } from 'listr2';
import inquirer from 'inquirer';
import { getInstalledModels, pullModel, removeModel, checkOllamaHealth, } from '../utils/ollama.js';
const RECOMMENDED_MODELS = [
{
name: 'codellama:7b',
description: 'Code Llama 7B - Lightweight code generation',
size: '~3.8GB',
},
{
name: 'granite-code:8b',
description: 'IBM Granite Code 8B - Enterprise code analysis',
size: '~4.6GB',
},
{
name: 'codellama:13b',
description: 'Code Llama 13B - Enhanced code generation',
size: '~7.3GB',
},
{
name: 'deepseek-coder:6.7b',
description: 'DeepSeek Coder 6.7B - Advanced code understanding',
size: '~3.8GB',
},
];
function format_size(bytes) {
const size_num = parseInt(bytes);
if (isNaN(size_num))
return bytes;
const gb = size_num / (1024 * 1024 * 1024);
return `${gb.toFixed(1)}GB`;
}
function format_date(date_string) {
try {
const date = new Date(date_string);
return date.toLocaleDateString();
}
catch {
return date_string;
}
}
async function list_models() {
const spinner = ora('Fetching installed models...').start();
try {
const models = await getInstalledModels();
spinner.succeed('Models retrieved successfully');
if (models.length === 0) {
console.log(chalk.yellow('\nš No models installed'));
console.log(chalk.dim('Run "code-audit models --pull <model>" to install a model'));
console.log(chalk.dim('\nRecommended models:'));
RECOMMENDED_MODELS.forEach((model) => {
console.log(chalk.dim(` ā¢ ${model.name} - ${model.description} (${model.size})`));
});
return;
}
console.log(chalk.green(`\nš¦ Installed Models (${models.length})`));
console.log(chalk.dim('ā'.repeat(80)));
models.forEach((model, index) => {
const number = chalk.dim(`${index + 1}.`);
const name = chalk.bold.cyan(model.name);
const size = chalk.yellow(format_size(model.size));
const modified = chalk.dim(format_date(model.modified));
console.log(`${number} ${name}`);
console.log(` Size: ${size} | Modified: ${modified}`);
if (index < models.length - 1)
console.log();
});
console.log(chalk.dim('\nā'.repeat(80)));
console.log(chalk.dim(`Total: ${models.length} models installed`));
}
catch (error) {
spinner.fail('Failed to retrieve models');
console.error(chalk.red('Error:'), error instanceof Error ? error.message : String(error));
process.exit(1);
}
}
async function pull_model_interactive(model_name) {
let target_model = model_name;
if (!target_model) {
const { selected_model } = await inquirer.prompt([
{
type: 'list',
name: 'selected_model',
message: 'Select a model to install:',
choices: [
...RECOMMENDED_MODELS.map((model) => ({
name: `${model.name} - ${model.description} (${model.size})`,
value: model.name,
})),
new inquirer.Separator(),
{
name: 'Enter custom model name',
value: 'custom',
},
],
},
]);
if (selected_model === 'custom') {
const { custom_name } = await inquirer.prompt([
{
type: 'input',
name: 'custom_name',
message: 'Enter model name (e.g., codellama:7b):',
validate: (input) => {
if (!input.trim())
return 'Model name is required';
if (!input.includes(':'))
return 'Model name should include tag (e.g., model:tag)';
return true;
},
},
]);
target_model = custom_name;
}
else {
target_model = selected_model;
}
}
console.log(chalk.blue(`\nš Pulling model: ${chalk.bold(target_model)}`));
console.log(chalk.dim('This may take several minutes depending on model size...'));
const spinner = ora('Downloading model...').start();
try {
await pullModel(target_model);
spinner.succeed(`Model ${chalk.bold(target_model)} installed successfully`);
console.log(chalk.green('\nā
Model is ready for use!'));
}
catch (error) {
spinner.fail(`Failed to install model ${target_model}`);
console.error(chalk.red('Error:'), error instanceof Error ? error.message : String(error));
console.log(chalk.yellow('\nš” Troubleshooting tips:'));
console.log(chalk.dim(' ā¢ Check your internet connection'));
console.log(chalk.dim(' ā¢ Verify Ollama is running: ollama list'));
console.log(chalk.dim(' ā¢ Ensure model name is correct'));
process.exit(1);
}
}
async function remove_model_interactive(model_name) {
let target_model = model_name;
if (!target_model) {
const models = await getInstalledModels();
if (models.length === 0) {
console.log(chalk.yellow('š No models installed to remove'));
return;
}
const { selected_model } = await inquirer.prompt([
{
type: 'list',
name: 'selected_model',
message: 'Select a model to remove:',
choices: models.map((model) => ({
name: `${model.name} (${format_size(model.size)})`,
value: model.name,
})),
},
]);
target_model = selected_model;
}
const { confirm } = await inquirer.prompt([
{
type: 'confirm',
name: 'confirm',
message: `Are you sure you want to remove ${chalk.bold(target_model)}?`,
default: false,
},
]);
if (!confirm) {
console.log(chalk.yellow('ā Removal cancelled'));
return;
}
const spinner = ora(`Removing model ${target_model}...`).start();
try {
await removeModel(target_model);
spinner.succeed(`Model ${chalk.bold(target_model)} removed successfully`);
}
catch (error) {
spinner.fail(`Failed to remove model ${target_model}`);
console.error(chalk.red('Error:'), error instanceof Error ? error.message : String(error));
process.exit(1);
}
}
async function update_all_models() {
console.log(chalk.blue('š Updating all installed models...'));
const models = await getInstalledModels();
if (models.length === 0) {
console.log(chalk.yellow('š No models installed to update'));
return;
}
const tasks = new Listr(models.map((model) => ({
title: `Updating ${model.name}`,
task: async () => {
await pullModel(model.name);
},
})), {
concurrent: false,
exitOnError: false,
});
try {
await tasks.run();
console.log(chalk.green('\nā
All models updated successfully!'));
}
catch {
console.error(chalk.red('\nā Some models failed to update'));
console.log(chalk.yellow('š” Try updating individual models manually'));
}
}
export async function modelsCommand(options) {
const { list = false, pull, remove, update = false } = options;
try {
const spinner = ora('Checking Ollama connection...').start();
await checkOllamaHealth();
spinner.succeed('Connected to Ollama');
}
catch (error) {
console.error(chalk.red('ā Cannot connect to Ollama'));
console.error(chalk.red('Error:'), error instanceof Error ? error.message : String(error));
console.log(chalk.yellow('\nš” Make sure Ollama is installed and running:'));
console.log(chalk.dim(' ā¢ Install: https://ollama.ai/download'));
console.log(chalk.dim(' ā¢ Start: ollama serve'));
process.exit(1);
}
try {
if (list) {
await list_models();
}
else if (pull) {
await pull_model_interactive(pull);
}
else if (remove) {
await remove_model_interactive(remove);
}
else if (update) {
await update_all_models();
}
else {
await list_models();
}
}
catch (error) {
console.error(chalk.red('Models command failed:'), error instanceof Error ? error.message : String(error));
process.exit(1);
}
}
//# sourceMappingURL=models.js.map