langtrain 0.1.14 → 0.1.16

package/src/cli.ts

@@ -1,622 +0,0 @@
-#!/usr/bin/env node
-import {
-    intro, outro, text, select, confirm, spinner, isCancel, note, cancel, password
-} from '@clack/prompts';
-import { bgCyan, black, red, green, yellow, gray } from 'kleur/colors';
-import { Command } from 'commander';
-import { AgentClient, AgentCreate, FileClient, TrainingClient, SubscriptionClient, Langvision, Langtune } from './index';
-import fs from 'fs';
-import path from 'path';
-import os from 'os';
-import gradient from 'gradient-string';
-
-// Configuration
-const CONFIG_DIR = path.join(os.homedir(), '.langtrain');
-const CONFIG_FILE = path.join(CONFIG_DIR, 'config.json');
-
-function getConfig() {
-    if (!fs.existsSync(CONFIG_FILE)) return {};
-    try {
-        return JSON.parse(fs.readFileSync(CONFIG_FILE, 'utf-8'));
-    } catch {
-        return {};
-    }
-}
-
-function saveConfig(config: any) {
-    if (!fs.existsSync(CONFIG_DIR)) {
-        fs.mkdirSync(CONFIG_DIR, { recursive: true });
-    }
-    fs.writeFileSync(CONFIG_FILE, JSON.stringify(config, null, 2));
-}
-
-const packageJson = require(path.join(__dirname, '../package.json'));
-
-async function main() {
-    const program = new Command();
-    const version = packageJson.version;
-
-    program
-        .name('langtrain')
-        .description(packageJson.description || 'Langtrain CLI for AI Model Fine-tuning and Generation')
-        .version(version);
-
-    program.action(async () => {
-        console.clear();
-
-        // Gradient Banner
-        const banner = `
-    ██╗      █████╗ ███╗   ██╗ ██████╗████████╗██████╗  █████╗ ██╗███╗   ██╗
-    ██║     ██╔══██╗████╗  ██║██╔════╝╚══██╔══╝██╔══██╗██╔══██╗██║████╗  ██║
-    ██║     ███████║██╔██╗ ██║██║  ███╗  ██║   ██████╔╝███████║██║██╔██╗ ██║
-    ██║     ██╔══██║██║╚██╗██║██║   ██║  ██║   ██╔══██╗██╔══██║██║██║╚██╗██║
-    ███████╗██║  ██║██║ ╚████║╚██████╔╝  ██║   ██║  ██║██║  ██║██║██║ ╚████║
-    ╚══════╝╚═╝  ╚═╝╚═╝  ╚═══╝ ╚═════╝   ╚═╝   ╚═╝  ╚═╝╚═╝  ╚═╝╚═╝╚═╝  ╚═══╝
-    `;
-        console.log(gradient(['#00DC82', '#36E4DA', '#0047E1'])(banner)); // Custom Langtrain Green-Cyan-Blue gradient
-        intro(`${bgCyan(black(` Langtrain SDK v${version} `))}`);
-
-        // Check auth (only show login if missing)
-        const config = getConfig();
-        if (!config.apiKey) {
-            intro(yellow('Authentication required'));
-            await handleLogin();
-        }
-
-        // Operation Handlers Map (O(1) lookup)
-        const handlers: Record<string, (clients?: any) => Promise<void>> = {
-            'login': handleLogin,
-            'status': handleSubscriptionStatus,
-            'tune-finetune': (c) => handleTuneFinetune(c.tune),
-            'tune-generate': (c) => handleTuneGenerate(c.tune),
-            'vision-finetune': (c) => handleVisionFinetune(c.vision),
-            'vision-generate': (c) => handleVisionGenerate(c.vision),
-            'agent-list': (c) => handleAgentList(c.agent),
-            'agent-create': (c) => handleAgentCreate(c.agent),
-            'agent-delete': (c) => handleAgentDelete(c.agent),
-            'exit': async () => { outro('Goodbye!'); process.exit(0); },
-        };
-
-        while (true) {
-            const operation = await select({
-                message: 'Select an operation:',
-                options: [
-                    { value: 'group-agents', label: '🤖 Agents (Server)', hint: 'Chat with custom agents' },
-                    { value: 'agent-list', label: '  ↳ List & Run Agents' },
-                    { value: 'agent-create', label: '  ↳ Create New Agent' },
-                    { value: 'agent-delete', label: '  ↳ Delete Agent' },
-
-                    { value: 'group-tune', label: '🧠 Langtune (LLM)', hint: 'Fine-tuning & Text Generation' },
-                    { value: 'tune-finetune', label: '  ↳ Fine-tune Text Model' },
-                    { value: 'tune-generate', label: '  ↳ Generate Text' },
-
-                    { value: 'group-vision', label: '👁️ Langvision (Vision)', hint: 'Vision Analysis & Tuning' },
-                    { value: 'vision-finetune', label: '  ↳ Fine-tune Vision Model' },
-                    { value: 'vision-generate', label: '  ↳ Generate Vision Response' },
-
-                    { value: 'group-settings', label: '⚙️ Settings' },
-                    { value: 'login', label: '  ↳ Update API Key' },
-                    { value: 'exit', label: '  ↳ Exit' }
-                ],
-            });
-
-            if (isCancel(operation)) {
-                outro('Goodbye!');
-                process.exit(0);
-            }
-
-            if (typeof operation === 'string') {
-                if (operation.startsWith('group-')) continue;
-
-                // Execute handler via map lookup
-                const handler = handlers[operation];
-                if (handler) {
-                    try {
-                        // Re-read config & re-init clients freshly for each operation
-                        const currentConfig = getConfig();
-                        const clients = {
-                            vision: new Langvision({ apiKey: currentConfig.apiKey }),
-                            tune: new Langtune({ apiKey: currentConfig.apiKey }),
-                            agent: new AgentClient({ apiKey: currentConfig.apiKey, baseUrl: currentConfig.baseUrl })
-                        };
-
-                        await handler(clients);
-                    } catch (error: any) {
-                        outro(red(`Error: ${error.message}`));
-                    }
-                }
-            }
-        }
-    });
-
-    program.parse(process.argv);
-}
-
-async function handleLogin() {
-    const apiKey = await password({
-        message: 'Enter your new Langtrain API Key:',
-        validate(value) {
-            if (!value || value.length === 0) return 'API Key is required';
-        },
-    });
-
-    if (isCancel(apiKey)) cancel('Operation cancelled');
-
-    const config = getConfig();
-    saveConfig({ ...config, apiKey });
-    intro(green('API Key updated successfully!'));
-}
-
-async function handleSubscriptionStatus() {
-    const config = getConfig();
-    if (!config.apiKey) {
-        intro(red('Not logged in. Run "login" first.'));
-        return;
-    }
-    const client = new SubscriptionClient({ apiKey: config.apiKey });
-    const s = spinner();
-    s.start('Fetching subscription status...');
-    try {
-        const info = await client.getStatus();
-        s.stop(green('Subscription Status:'));
-
-        console.log(gray('Plan: ') + (info.plan === 'pro' ? bgCyan(' PRO ') : info.plan.toUpperCase()));
-        console.log(gray('Active: ') + (info.is_active ? green('Yes') : red('No')));
-        if (info.expires_at) console.log(gray('Expires: ') + new Date(info.expires_at).toLocaleDateString());
-
-        console.log(gray('\nLimits:'));
-        console.log(`  Models: ${info.limits.max_models === -1 ? 'Unlimited' : info.limits.max_models}`);
-        console.log(`  Training Jobs: ${info.limits.max_training_jobs}`);
-
-    } catch (e: any) {
-        s.stop(red('Failed to fetch status.'));
-        console.error(e.message);
-    }
-}
-
-async function handleAgentCreate(client: AgentClient) {
-    const name = await text({
-        message: 'Agent Name:',
-        placeholder: 'e.g. Support Bot',
-        validate(value) {
-            if (!value || value.length === 0) return 'API Key is required';
-        },
-    });
-    if (isCancel(name)) return;
-
-    const description = await text({
-        message: 'Description:',
-        placeholder: 'e.g. A helpful support assistant',
-    });
-    if (isCancel(description)) return;
-
-    const systemPrompt = await text({
-        message: 'System Prompt:',
-        placeholder: 'e.g. You are a helpful assistant.',
-        initialValue: 'You are a helpful assistant.'
-    });
-    if (isCancel(systemPrompt)) return;
-
-    const s = spinner();
-    s.start('Creating agent...');
-
-    try {
-        // We need a workspace ID. server usually infers it from API key context if not provided?
-        // But the schema says workspace_id is required in AgentCreate.
-        // The server implementation of create_agent takes AgentCreate which has workspace_id.
-        // However, standard users might not know their exact workspace UUID. 
-        // We might need to fetch it or rely on server to fill it if we made it optional in schema (which we didn't).
-        // EDIT: Let's fetch one agent to get the workspace_id or assume one?
-        // Better: List agents, get workspace_id from first one. Hacky but works for single-workspace users.
-        // Use 'default' or similar if server supports it?
-        // Checking agents.py: verify_api_key returns workspace_id.
-        // But create_agent payload requires it.
-        // I'll try to fetch list first to get workspace ID. If list empty, we are stuck?
-        // Wait, list_agents returns `AgentListResponse` which doesn't explicitly return workspace_id at top level, but agents have it.
-        // If no agents, we can't guess it.
-        // Maybe I should fetch user profile? No endpoint for that in CLI yet.
-        // I'll try to pass a placeholder and hope server ignores it if it uses context?
-        // Server code: `workspace_id=agent_in.workspace_id`. It uses payload.
-        // I might need to ask user for workspace ID or update server to be smarter.
-        // For now, I'll attempt to LIST agents to get a workspace ID.
-
-        const agents = await client.list();
-        let workspaceId = "";
-        if (agents.length > 0) {
-            workspaceId = agents[0].workspace_id;
-        } else {
-            // Fallback: Ask user or fail?
-            // Or maybe decoding JWT/API key client side? No.
-            // I'll prompt for it if not found, with a hint.
-            s.stop(yellow('Workspace ID needed (no existing agents found).'));
-            const wid = await text({
-                message: 'Enter Workspace ID (UUID):',
-                validate(value) {
-                    if (!value || value.length === 0) return 'Required';
-                },
-            });
-            if (isCancel(wid)) return;
-            workspaceId = wid as string;
-            s.start('Creating agent...');
-        }
-
-        const agent = await client.create({
-            workspace_id: workspaceId,
-            name: name as string,
-            description: description as string,
-            config: {
-                system_prompt: systemPrompt as string,
-                model: 'gpt-4o' // Default or prompt? simplified
-            }
-        });
-        s.stop(green(`Agent "${agent.name}" created successfully! ID: ${agent.id}`));
-    } catch (e: any) {
-        s.stop(red('Failed to create agent.'));
-        throw e;
-    }
-}
-
-async function handleAgentDelete(client: AgentClient) {
-    const s = spinner();
-    s.start('Fetching agents...');
-    const agents = await client.list();
-    s.stop(`Found ${agents.length} agents`);
-
-    if (agents.length === 0) {
-        intro(yellow('No agents to delete.'));
-        return;
-    }
-
-    const agentId = await select({
-        message: 'Select an agent to DELETE:',
-        options: agents.map(a => ({ value: a.id, label: a.name, hint: a.description || 'No description' }))
-    });
-
-    if (isCancel(agentId)) return;
-
-    const confirm = await select({
-        message: `Are you sure you want to delete this agent?`,
-        options: [
-            { value: 'yes', label: 'Yes, delete it', hint: 'Cannot be undone' },
-            { value: 'no', label: 'No, keep it' }
-        ]
-    });
-
-    if (confirm !== 'yes') {
-        intro(gray('Deletion cancelled.'));
-        return;
-    }
-
-    const d = spinner();
-    d.start('Deleting agent...');
-    try {
-        await client.delete(agentId as string);
-        d.stop(green('Agent deleted successfully.'));
-    } catch (e: any) {
-        d.stop(red('Failed to delete agent.'));
-        throw e;
-    }
-}
-
-async function handleAgentList(client: AgentClient) {
-    const s = spinner();
-    s.start('Fetching agents...');
-    const agents = await client.list();
-    s.stop(`Found ${agents.length} agents`);
-
-    if (agents.length === 0) {
-        intro(yellow('No agents found in your workspace.'));
-        return;
-    }
-
-    const agentId = await select({
-        message: 'Select an agent to run:',
-        options: agents.map(a => ({ value: a.id, label: a.name, hint: a.description || 'No description' }))
-    });
-
-    if (isCancel(agentId)) return;
-
-    await handleAgentRun(client, agentId as string, agents.find(a => a.id === agentId)?.name || 'Agent');
-}
-
-async function handleAgentRun(client: AgentClient, agentId: string, agentName: string) {
-    intro(bgCyan(black(` Chatting with ${agentName} `)));
-    console.log(gray('Type "exit" to quit conversation.'));
-
-    let conversationId: string | undefined = undefined;
-
-    while (true) {
-        const input = await text({
-            message: 'You:',
-            placeholder: 'Type a message...',
-        });
-
-        if (isCancel(input) || input === 'exit') {
-            break;
-        }
-
-        const s = spinner();
-        s.start('Agent is thinking...');
-        try {
-            const result = await client.execute(agentId, { prompt: input }, [], conversationId);
-            s.stop();
-
-            if (result.output && result.output.response) {
-                console.log(gradient.pastel(`Agent: ${result.output.response}`));
-            } else {
-                console.log(gradient.pastel(`Agent: ${JSON.stringify(result.output)}`));
-            }
-
-            conversationId = result.conversation_id;
-        } catch (e: any) {
-            s.stop(red('Error running agent.'));
-            console.error(e);
-        }
-    }
-}
-
-
-// Handler for Langtune Fine-tuning
-async function handleTuneFinetune(tune: Langtune) {
-    const model = await text({
-        message: 'Enter base model (e.g., gpt-3.5-turbo):',
-        placeholder: 'gpt-3.5-turbo',
-        validate(value) {
-            if (!value || value.length === 0) return 'Value is required!';
-        },
-    });
-    if (isCancel(model)) cancel('Operation cancelled.');
-
-    const trainFile = await text({
-        message: 'Enter path to training file:',
-        placeholder: './data.jsonl',
-        validate(value) {
-            if (!value || value.length === 0) return 'Value is required!';
-        },
-    });
-    if (isCancel(trainFile)) cancel('Operation cancelled.');
-
-    const epochs = await text({
-        message: 'Num Epochs:',
-        placeholder: '3',
-        initialValue: '3'
-    });
-    if (isCancel(epochs)) cancel('Operation cancelled.');
-
-    const track = await select({
-        message: 'Track this job on Langtrain Cloud?',
-        options: [
-            { value: 'yes', label: 'Yes', hint: 'Upload dataset and log job' },
-            { value: 'no', label: 'No', hint: 'Local only' }
-        ]
-    });
-    if (isCancel(track)) cancel('Operation cancelled.');
-
-    if (track === 'yes') {
-        const s = spinner();
-        s.start('Connecting to Cloud...');
-        try {
-            const config = getConfig();
-            if (!config.apiKey) throw new Error('API Key required. Run "login" first.');
-
-            // Check Subscription
-            const subClient = new SubscriptionClient({ apiKey: config.apiKey });
-            const sub = await subClient.getStatus();
-            if (!sub.features.includes('cloud_finetuning')) {
-                s.stop(red('Feature "cloud_finetuning" is not available on your plan.'));
-                const upgrade = await confirm({ message: 'Upgrade to Pro for cloud tracking?' });
-                if (upgrade && !isCancel(upgrade)) {
-                    console.log(bgCyan(black(' Visit https://langtrain.ai/dashboard/billing to upgrade. ')));
-                }
-                return;
-            }
-
-            const fileClient = new FileClient({ apiKey: config.apiKey });
-            const trainingClient = new TrainingClient({ apiKey: config.apiKey });
-
-            s.message('Uploading dataset...');
-            const fileResp = await fileClient.upload(trainFile as string);
-
-            s.message('Creating Job...');
-            const job = await trainingClient.createJob({
-                name: `cli-sft-${Date.now()}`,
-                base_model: model as string,
-                dataset_id: fileResp.id,
-                task: 'text',
-                hyperparameters: {
-                    n_epochs: parseInt(epochs as string)
-                }
-            });
-            s.stop(green(`Job tracked: ${job.id}`));
-        } catch (e: any) {
-            s.stop(red(`Tracking failed: ${e.message}`));
-            const cont = await confirm({ message: 'Continue with local training anyway?' });
-            if (!cont || isCancel(cont)) return;
-        }
-    }
-
-    const s = spinner();
-    s.start('Starting local fine-tuning...');
-
-    try {
-        // Check if FinetuneConfig types match what's needed. 
-        // Casting to any to bypass strict type checking for this demo or ensure types are imported correctly.
-        // In a real scenario, we'd construct the full config object.
-        const config: any = {
-            model: model as string,
-            trainFile: trainFile as string,
-            preset: 'default', // simplified
-            epochs: parseInt(epochs as string),
-            batchSize: 1,
-            learningRate: 2e-5,
-            loraRank: 16,
-            outputDir: './output'
-        };
-
-        await tune.finetune(config);
-        s.stop(green('Fine-tuning job started successfully! 🚀'));
-    } catch (e: any) {
-        s.stop(red('Failed to start job.'));
-        throw e;
-    }
-}
-
-// Handler for Langtune Generation
-async function handleTuneGenerate(tune: Langtune) {
-    const model = await text({
-        message: 'Enter model path:',
-        placeholder: './output/model',
-        initialValue: './output/model'
-    });
-    if (isCancel(model)) cancel('Operation cancelled');
-
-    const prompt = await text({
-        message: 'Enter prompt:',
-        placeholder: 'Hello world',
-    });
-    if (isCancel(prompt)) cancel('Operation cancelled');
-
-    const s = spinner();
-    s.start('Connecting to Langtrain Inference API...');
-
-    try {
-        const response = await tune.generate(model as string, { prompt: prompt as string });
-        s.stop('Generation complete');
-        intro('Response:');
-        console.log(gradient.pastel(response));
-    } catch (e: any) {
-        s.stop(red('Generation failed.'));
-        throw e;
-    }
-}
-
-// Handler for Langvision Fine-tuning
-async function handleVisionFinetune(vision: Langvision) {
-    const model = await text({
-        message: 'Enter base vision model:',
-        placeholder: 'llava-v1.5-7b',
-        initialValue: 'llava-v1.5-7b'
-    });
-    if (isCancel(model)) cancel('Operation cancelled');
-
-    const dataset = await text({
-        message: 'Enter dataset path:',
-        placeholder: './dataset',
-    });
-    if (isCancel(dataset)) cancel('Operation cancelled');
-
-    const epochs = await text({
-        message: 'Num Epochs:',
-        placeholder: '3',
-        initialValue: '3'
-    });
-    if (isCancel(epochs)) cancel('Operation cancelled');
-
-    const track = await select({
-        message: 'Track this job on Langtrain Cloud?',
-        options: [
-            { value: 'yes', label: 'Yes', hint: 'Upload dataset and log job' },
-            { value: 'no', label: 'No', hint: 'Local only' }
-        ]
-    });
-    if (isCancel(track)) cancel('Operation cancelled');
-
-    if (track === 'yes') {
-        const s = spinner();
-        s.start('Connecting to Cloud...');
-        try {
-            const config = getConfig();
-            if (!config.apiKey) throw new Error('API Key required. Run "login" first.');
-
-            // Check Subscription
-            const subClient = new SubscriptionClient({ apiKey: config.apiKey });
-            const sub = await subClient.getStatus();
-            if (!sub.features.includes('cloud_finetuning')) {
-                s.stop(red('Feature "cloud_finetuning" is not available on your plan.'));
-                const upgrade = await confirm({ message: 'Upgrade to Pro for cloud tracking?' });
-                if (upgrade && !isCancel(upgrade)) {
-                    console.log(bgCyan(black(' Visit https://langtrain.ai/dashboard/billing to upgrade. ')));
-                }
-                return;
-            }
-
-            const fileClient = new FileClient({ apiKey: config.apiKey });
-            const trainingClient = new TrainingClient({ apiKey: config.apiKey });
-
-            s.message('Uploading dataset...');
-            const fileResp = await fileClient.upload(dataset as string, undefined, 'fine-tune-vision');
-
-            s.message('Creating Job...');
-            const job = await trainingClient.createJob({
-                name: `cli-vision-${Date.now()}`,
-                base_model: model as string,
-                dataset_id: fileResp.id,
-                task: 'vision',
-                training_method: 'lora',
-                hyperparameters: {
-                    n_epochs: parseInt(epochs as string)
-                }
-            });
-            s.stop(green(`Job tracked: ${job.id}`));
-        } catch (e: any) {
-            s.stop(red(`Tracking failed: ${e.message}`));
-            const cont = await confirm({ message: 'Continue with local training anyway?' });
-            if (!cont || isCancel(cont)) return;
-        }
-    }
-
-    const s = spinner();
-    s.start('Analyzing dataset structure...');
-    await new Promise(r => setTimeout(r, 800));
-    s.message('Starting vision fine-tuning on Langtrain Cloud...');
-
-    try {
-        const config: any = {
-            model: model as string,
-            dataset: dataset as string,
-            epochs: parseInt(epochs as string),
-            batchSize: 1,
-            learningRate: 2e-5,
-            loraRank: 16,
-            outputDir: './vision-output'
-        };
-        await vision.finetune(config);
-        s.stop(green('Vision fine-tuning started successfully! 👁️'));
-    } catch (e: any) {
-        s.stop(red('Failed to start vision job.'));
-        throw e;
-    }
-}
-
-// Handler for Langvision Generation
-async function handleVisionGenerate(vision: Langvision) {
-    const model = await text({
-        message: 'Enter model path:',
-        placeholder: './vision-output/model',
-        initialValue: './vision-output/model'
-    });
-    if (isCancel(model)) cancel('Operation cancelled');
-
-    const prompt = await text({
-        message: 'Enter prompt/image path:', // Simplified for CLI
-        placeholder: 'Describe this image...',
-    });
-    if (isCancel(prompt)) cancel('Operation cancelled');
-
-    const s = spinner();
-    s.start('Uploading image and context...');
-    await new Promise(r => setTimeout(r, 600));
-    s.message('Generating vision response...');
-
-    try {
-        const response = await vision.generate(model as string, { prompt: prompt as string });
-        s.stop('Generation complete');
-        intro('Response:');
-        console.log(gradient.pastel(response));
-    } catch (e: any) {
-        s.stop(red('Generation failed.'));
-        throw e;
-    }
-}
-
-main().catch(console.error);