@qvac/sdk 0.14.1 → 0.15.0

package/dist/examples/llamacpp-structured-output.js

@@ -1,4 +1,4 @@
-import { loadModel, completion, unloadModel, QWEN3_600M_INST_Q4, } from "@qvac/sdk";
+import { loadModel, completion, unloadModel, QWEN3_600M_INST_Q4 } from '@qvac/sdk';
 // Demonstrates per-request structured output via `responseFormat`.
 // Three modes are exercised against the same model so you can eyeball the
 // difference between free-form text, free-form JSON, and a strict JSON Schema.
@@ -6,24 +6,24 @@ import { loadModel, completion, unloadModel, QWEN3_600M_INST_Q4, } from "@qvac/s
 // Usage:
 //   bun run examples/llamacpp-structured-output.ts
 const PERSON_SCHEMA = {
-    type: "object",
+    type: 'object',
     properties: {
-        name: { type: "string" },
-        age: { type: "integer" },
-        occupation: { type: "string" },
+        name: { type: 'string' },
+        age: { type: 'integer' },
+        occupation: { type: 'string' }
     },
-    required: ["name", "age", "occupation"],
-    additionalProperties: false,
+    required: ['name', 'age', 'occupation'],
+    additionalProperties: false
 };
 const HISTORY = [
     {
-        role: "system",
-        content: "You extract structured information about people from short bios. /no_think",
+        role: 'system',
+        content: 'You extract structured information about people from short bios. /no_think'
     },
     {
-        role: "user",
-        content: "Hi, I'm Alice, I'm 30 years old and I work as a data engineer.",
-    },
+        role: 'user',
+        content: "Hi, I'm Alice, I'm 30 years old and I work as a data engineer."
+    }
 ];
 // `json_object` mode only enforces that the output is *some* valid JSON object
 // — it doesn't pin the keys. Small models (Qwen3-0.6B in this example) will
@@ -34,13 +34,13 @@ const HISTORY = [
 // grammar itself forces the keys.
 const JSON_OBJECT_HISTORY = [
     {
-        role: "system",
-        content: 'You extract structured information about people from short bios and reply ONLY with a JSON object containing "name", "age", and "occupation". /no_think',
+        role: 'system',
+        content: 'You extract structured information about people from short bios and reply ONLY with a JSON object containing "name", "age", and "occupation". /no_think'
     },
     {
-        role: "user",
-        content: "Hi, I'm Alice, I'm 30 years old and I work as a data engineer.",
-    },
+        role: 'user',
+        content: "Hi, I'm Alice, I'm 30 years old and I work as a data engineer."
+    }
 ];
 // Drains the run via the canonical `events` / `final` surface and returns the
 // content text. `tokenStream` / `text` / `stats` still work but are legacy and
@@ -48,11 +48,11 @@ const JSON_OBJECT_HISTORY = [
 // await `final.contentText` for the aggregated result.
 async function runToContent(run) {
     for await (const event of run.events) {
-        if (event.type === "contentDelta")
+        if (event.type === 'contentDelta')
             process.stdout.write(event.text);
     }
     const final = await run.final;
-    process.stdout.write("\n");
+    process.stdout.write('\n');
     return final.contentText;
 }
 try {
@@ -63,49 +63,49 @@ try {
             const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
             process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
             if (p.percentage >= 100)
-                process.stderr.write("\n");
-        },
+                process.stderr.write('\n');
+        }
     });
     console.log(`\n▸ Model loaded: ${modelId}\n`);
-    console.log("▸ --- 1. responseFormat: text (baseline, free-form) ---");
+    console.log('▸ --- 1. responseFormat: text (baseline, free-form) ---');
     await runToContent(completion({
         modelId,
         history: HISTORY,
         stream: true,
-        responseFormat: { type: "text" },
+        responseFormat: { type: 'text' }
     }));
-    console.log("\n▸ --- 2. responseFormat: json_object (any valid JSON) ---");
+    console.log('\n▸ --- 2. responseFormat: json_object (any valid JSON) ---');
     const jsonObjectOut = await runToContent(completion({
         modelId,
         history: JSON_OBJECT_HISTORY,
         stream: true,
-        responseFormat: { type: "json_object" },
+        responseFormat: { type: 'json_object' }
     }));
-    console.log("▸ parsed:", JSON.parse(jsonObjectOut.trim()));
-    console.log("\n▸ --- 3. responseFormat: json_schema (strict shape) ---");
+    console.log('▸ parsed:', JSON.parse(jsonObjectOut.trim()));
+    console.log('\n▸ --- 3. responseFormat: json_schema (strict shape) ---');
     const jsonSchemaOut = await runToContent(completion({
         modelId,
         history: HISTORY,
         stream: true,
         responseFormat: {
-            type: "json_schema",
+            type: 'json_schema',
             json_schema: {
-                name: "person",
-                schema: PERSON_SCHEMA,
-            },
-        },
+                name: 'person',
+                schema: PERSON_SCHEMA
+            }
+        }
     }));
     // Output is guaranteed schema-valid JSON: object with exactly
     // {name: string, age: integer, occupation: string}, no extras.
     const parsed = JSON.parse(jsonSchemaOut.trim());
-    console.log("▸ parsed:", parsed);
-    console.log("▸ schema-valid:", typeof parsed.name === "string" &&
+    console.log('▸ parsed:', parsed);
+    console.log('▸ schema-valid:', typeof parsed.name === 'string' &&
         Number.isInteger(parsed.age) &&
-        typeof parsed.occupation === "string" &&
-        Object.keys(parsed).sort().join(",") === "age,name,occupation");
+        typeof parsed.occupation === 'string' &&
+        Object.keys(parsed).sort().join(',') === 'age,name,occupation');
     await unloadModel({ modelId });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/logging-file-transport.js

@@ -1,13 +1,13 @@
-import { getLogger, loadModel, completion, unloadModel, LLAMA_3_2_1B_INST_Q4_0, VERBOSITY, } from "@qvac/sdk";
-import { writeFileSync, appendFileSync } from "fs";
-import { join } from "path";
+import { getLogger, loadModel, completion, unloadModel, LLAMA_3_2_1B_INST_Q4_0, VERBOSITY } from '@qvac/sdk';
+import { writeFileSync, appendFileSync } from 'fs';
+import { join } from 'path';
 try {
-    const logFile = join(process.cwd(), "model-logs.txt");
+    const logFile = join(process.cwd(), 'model-logs.txt');
     // Clear previous logs
-    writeFileSync(logFile, "");
+    writeFileSync(logFile, '');
     // Create logger with file transport
-    const logger = getLogger("llm-file-demo", {
-        level: "debug",
+    const logger = getLogger('llm-file-demo', {
+        level: 'debug',
         enableConsole: false, // Disable console output (logs go to file only). Set to true for both console + file
         transports: [
             // File transport - append logs to file
@@ -15,42 +15,40 @@ try {
                 const timestamp = new Date().toISOString();
                 const logLine = `[${timestamp}] ${level.toUpperCase()} ${namespace}: ${message}\n`;
                 appendFileSync(logFile, logLine);
-            },
-        ],
+            }
+        ]
     });
-    logger.info("Starting model demo with file logging");
+    logger.info('Starting model demo with file logging');
     // Load model with integrated logging
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
             ctx_size: 2048,
             temp: 0.7,
-            verbosity: VERBOSITY.DEBUG,
+            verbosity: VERBOSITY.DEBUG
         },
-        logger, // All model logs will be captured
+        logger // All model logs will be captured
     });
     logger.info(`Model loaded: ${modelId}`);
     // Run a simple completion
-    logger.info("Starting completion...");
-    const messages = [
-        { role: "user", content: "Tell me a short joke about programming." },
-    ];
-    let responseText = "";
+    logger.info('Starting completion...');
+    const messages = [{ role: 'user', content: 'Tell me a short joke about programming.' }];
+    let responseText = '';
     const result = completion({ modelId, history: messages, stream: true });
     for await (const token of result.tokenStream) {
         responseText += token;
         process.stdout.write(token);
     }
-    console.log("\n");
+    console.log('\n');
     logger.info(`Completion finished, ${responseText.length} characters generated`);
     // Unload model
     await unloadModel({ modelId });
-    logger.info("Model unloaded successfully");
+    logger.info('Model unloaded successfully');
     // Show user where logs were saved
     console.log(`▸ Logs saved to ${logFile}`);
-    console.log("▸ Inspect it to see detailed model operation logs");
+    console.log('▸ Inspect it to see detailed model operation logs');
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/logging-global.js

@@ -2,10 +2,10 @@
 // transport on both the client and the worker, so the only logs printed are
 // the ones our handler below receives.
 const configDir = import.meta.dirname ?? process.cwd();
-process.env["QVAC_CONFIG_PATH"] = `${configDir}/config/logging/logging.config.json`;
-const { loadModel, completion, unloadModel, subscribeServerLogs, LLAMA_3_2_1B_INST_Q4_0 } = await import("@qvac/sdk");
+process.env['QVAC_CONFIG_PATH'] = `${configDir}/config/logging/logging.config.json`;
+const { loadModel, completion, unloadModel, subscribeServerLogs, LLAMA_3_2_1B_INST_Q4_0 } = await import('@qvac/sdk');
 try {
-    console.log("▸ Starting global logging demo...\n");
+    console.log('▸ Starting global logging demo...\n');
     // One subscription captures every server-side log (SDK, models, RAG, …)
     // without having to know any stream IDs ahead of time.
     const unsubscribe = subscribeServerLogs((log) => {
@@ -13,14 +13,14 @@ try {
     });
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
-        modelConfig: { ctx_size: 2048 },
+        modelConfig: { ctx_size: 2048 }
     });
     const result = completion({
         modelId,
-        history: [{ role: "user", content: "Count from 1 to 5." }],
-        stream: true,
+        history: [{ role: 'user', content: 'Count from 1 to 5.' }],
+        stream: true
     });
-    console.log("▸ Response:\n");
+    console.log('▸ Response:\n');
     for await (const token of result.tokenStream) {
         process.stdout.write(token);
     }
@@ -28,7 +28,7 @@ try {
     unsubscribe();
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }
 export {};

package/dist/examples/logging-streaming.js

@@ -1,10 +1,10 @@
-import { loadModel, completion, unloadModel, loggingStream, SDK_LOG_ID, LLAMA_3_2_1B_INST_Q4_0, GTE_LARGE_FP16, VERBOSITY, embed, } from "@qvac/sdk";
+import { loadModel, completion, unloadModel, loggingStream, SDK_LOG_ID, LLAMA_3_2_1B_INST_Q4_0, GTE_LARGE_FP16, VERBOSITY, embed } from '@qvac/sdk';
 try {
-    console.log("▸ Streaming SDK and model logs");
+    console.log('▸ Streaming SDK and model logs');
     // Note: To configure logging (level and console output), use config file:
     // { "loggerLevel": "debug", "loggerConsoleOutput": false } in qvac.config.json/js/ts
     // Subscribe to SDK server logs in background
-    console.log("▸ Subscribing to SDK server logs");
+    console.log('▸ Subscribing to SDK server logs');
     (async () => {
         for await (const log of loggingStream({ id: SDK_LOG_ID })) {
             console.log(`[SDK] [${log.level.toUpperCase()}] [${log.namespace}] ${log.message}`);
@@ -13,19 +13,19 @@ try {
         // Stream terminated - normal on shutdown
     });
     // Load models
-    console.log("▸ Loading models");
+    console.log('▸ Loading models');
     const llmModelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
             ctx_size: 2048,
             temp: 0.7,
-            verbosity: VERBOSITY.ERROR, // Only log errors, remaining logs are captured by loggingStream
-        },
+            verbosity: VERBOSITY.ERROR // Only log errors, remaining logs are captured by loggingStream
+        }
     });
     const embedModelId = await loadModel({
-        modelSrc: GTE_LARGE_FP16,
+        modelSrc: GTE_LARGE_FP16
     });
-    console.log("▸ Subscribing to model logs");
+    console.log('▸ Subscribing to model logs');
     (async () => {
         for await (const log of loggingStream({ id: llmModelId })) {
             const timestamp = new Date(log.timestamp).toISOString();
@@ -42,29 +42,27 @@ try {
     })().catch(() => {
         // Stream terminated - this is normal when model unloads
     });
-    const messages = [
-        { role: "user", content: "Count from 1 to 5 and explain each number." },
-    ];
+    const messages = [{ role: 'user', content: 'Count from 1 to 5 and explain each number.' }];
     const result = completion({
         modelId: llmModelId,
         history: messages,
-        stream: true,
+        stream: true
     });
     const { embedding } = await embed({
         modelId: embedModelId,
-        text: messages[0]?.content ?? "Hello, world!",
+        text: messages[0]?.content ?? 'Hello, world!'
     });
-    console.log("▸ Response");
+    console.log('▸ Response');
     for await (const token of result.tokenStream) {
         process.stdout.write(token);
     }
-    console.log("Embedding (first 20 elements)", embedding.slice(0, 20));
-    console.log("Embeddings length", embedding.length);
-    console.log("▸ Three streams running: [SDK] server, [LLM] inference, [EMBED] embedding");
+    console.log('Embedding (first 20 elements)', embedding.slice(0, 20));
+    console.log('Embeddings length', embedding.length);
+    console.log('▸ Three streams running: [SDK] server, [LLM] inference, [EMBED] embedding');
     await unloadModel({ modelId: llmModelId, clearStorage: false });
     await unloadModel({ modelId: embedModelId, clearStorage: false });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/mcp-websearch.js

@@ -9,18 +9,18 @@
  *
  * Run with: bun run examples/mcp-websearch.ts
  */
-import { completion, loadModel, unloadModel, QWEN3_1_7B_INST_Q4, } from "../index.js";
+import { completion, loadModel, unloadModel, QWEN3_1_7B_INST_Q4 } from '../index.js';
 // MCP SDK is a user-installed optional dependency
 // Install with: bun add @modelcontextprotocol/sdk
 // eslint-disable-next-line import/no-unresolved
-import { Client } from "@modelcontextprotocol/sdk/client/index.js";
+import { Client } from '@modelcontextprotocol/sdk/client/index.js';
 // eslint-disable-next-line import/no-unresolved
-import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";
+import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';
 function parseSearchResults(mcpResult) {
     try {
         const result = mcpResult;
         // Extract text content from MCP response
-        const textContent = result.content?.find((c) => c.type === "text");
+        const textContent = result.content?.find((c) => c.type === 'text');
         if (!textContent?.text) {
             return JSON.stringify(mcpResult);
         }
@@ -28,55 +28,55 @@ function parseSearchResults(mcpResult) {
         const rawResults = JSON.parse(textContent.text);
         // Extract just the useful fields (title, url, snippet)
         const cleanResults = rawResults.slice(0, 5).map((r) => ({
-            title: r.title ?? "Unknown",
-            url: r.url ?? "",
-            snippet: r.snippet ?? "",
+            title: r.title ?? 'Unknown',
+            url: r.url ?? '',
+            snippet: r.snippet ?? ''
         }));
         // Format as concise text for LLM
         return cleanResults
             .map((r, i) => `[${i + 1}] ${r.title}\n    URL: ${r.url}\n    ${r.snippet}`)
-            .join("\n\n");
+            .join('\n\n');
     }
     catch {
         // If parsing fails, return a truncated version
-        const str = typeof mcpResult === "string" ? mcpResult : JSON.stringify(mcpResult);
+        const str = typeof mcpResult === 'string' ? mcpResult : JSON.stringify(mcpResult);
         return str.slice(0, 2000);
     }
 }
 let mcpClient = null;
 try {
-    console.log("▸ MCP DuckDuckGo Search Example\n");
+    console.log('▸ MCP DuckDuckGo Search Example\n');
     // ============================================================
     // STEP 1: Connect to DuckDuckGo MCP server
     // ============================================================
-    console.log("▸ Starting DuckDuckGo MCP server...");
+    console.log('▸ Starting DuckDuckGo MCP server...');
     mcpClient = new Client({
-        name: "qvac-ddg-example",
-        version: "1.0.0",
+        name: 'qvac-ddg-example',
+        version: '1.0.0'
     });
     const transport = new StdioClientTransport({
-        command: "npx",
-        args: ["-y", "@oevortex/ddg_search"],
+        command: 'npx',
+        args: ['-y', '@oevortex/ddg_search']
     });
     await mcpClient.connect(transport);
-    console.log("▸ MCP server connected\n");
+    console.log('▸ MCP server connected\n');
     // ============================================================
     // STEP 2: Load model
     // ============================================================
-    console.log("▸ Loading model...");
+    console.log('▸ Loading model...');
     const modelId = await loadModel({
         modelSrc: QWEN3_1_7B_INST_Q4,
         modelConfig: {
             ctx_size: 4096,
-            tools: true,
+            tools: true
         },
         onProgress: (p) => {
             const mb = (n) => (n / 1e6).toFixed(1);
             const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
             process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
             if (p.percentage >= 100)
-                process.stderr.write("\n");
-        },
+                process.stderr.write('\n');
+        }
     });
     console.log(`\n▸ Model loaded\n`);
     // ============================================================
@@ -84,35 +84,35 @@ try {
     // ============================================================
     const history = [
         {
-            role: "system",
+            role: 'system',
             content: `You are a helpful assistant with access to web search.
 Use the search tool when you need current information.
-Always cite your sources with the URL.`,
+Always cite your sources with the URL.`
         },
         {
-            role: "user",
-            content: "What is the current weather in New York City?",
-        },
+            role: 'user',
+            content: 'What is the current weather in New York City?'
+        }
     ];
-    console.log("▸ Asking AI to search the web...\n");
-    console.log("▸ AI Response:");
+    console.log('▸ Asking AI to search the web...\n');
+    console.log('▸ AI Response:');
     // Pass MCP client directly to completion - tools are adapted internally!
     const result = completion({
         modelId,
         history,
         stream: true,
-        mcp: [{ client: mcpClient, includeResources: false }],
+        mcp: [{ client: mcpClient, includeResources: false }]
     });
     for await (const token of result.tokenStream) {
         process.stdout.write(token);
     }
     const toolCalls = await result.toolCalls;
-    console.log("\n");
+    console.log('\n');
     // ============================================================
     // STEP 4: Execute tool calls using call() - automatic MCP routing!
     // ============================================================
     if (toolCalls.length > 0) {
-        console.log("▸ Executing search...\n");
+        console.log('▸ Executing search...\n');
         const toolResults = [];
         for (const toolCall of toolCalls) {
             console.log(`▸ ${toolCall.name}(${JSON.stringify(toolCall.arguments)})`);
@@ -126,49 +126,49 @@ Always cite your sources with the URL.`,
             const cleanResult = parseSearchResults(mcpResult);
             console.log(`▸ Got search results:`);
             console.log(cleanResult
-                .split("\n")
+                .split('\n')
                 .map((l) => `      ${l}`)
-                .join("\n"));
+                .join('\n'));
             console.log();
             toolResults.push({ id: toolCall.id, result: cleanResult });
         }
         // ============================================================
         // STEP 5: Continue with search results
         // ============================================================
-        console.log("▸ Getting AI response with search results...\n");
+        console.log('▸ Getting AI response with search results...\n');
         history.push({
-            role: "assistant",
-            content: await result.text,
+            role: 'assistant',
+            content: await result.text
         });
         for (const tr of toolResults) {
             history.push({
-                role: "tool",
-                content: tr.result,
+                role: 'tool',
+                content: tr.result
             });
         }
-        console.log("▸ Final Response:");
+        console.log('▸ Final Response:');
         const finalResult = completion({
             modelId,
             history,
             stream: true,
-            mcp: [{ client: mcpClient, includeResources: false }],
+            mcp: [{ client: mcpClient, includeResources: false }]
         });
         for await (const token of finalResult.tokenStream) {
             process.stdout.write(token);
         }
-        console.log("\n");
+        console.log('\n');
     }
     // ============================================================
     // Cleanup
     // ============================================================
-    console.log("▸ Cleaning up...");
+    console.log('▸ Cleaning up...');
     await unloadModel({ modelId, clearStorage: false });
-    console.log("▸ Done\n");
-    console.log("▸ Example completed!");
+    console.log('▸ Done\n');
+    console.log('▸ Example completed!');
     process.exit(0);
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }
 finally {