evals 2.2.7 → 2.2.8

package/README.md

@@ -1,53 +1,53 @@
-# evals
-
-An npm package for Arize evals functionality.
-
-## About Arize AI
-
-Arize AI is a leading company in AI observability and evaluation, dedicated to ensuring that artificial intelligence systems operate reliably in real-world applications. Founded in 2019, Arize provides tools that help machine learning teams monitor, troubleshoot, and improve model performance across various domains, including structured data, computer vision, and large language models (LLMs).
-
-## Arize Phoenix
-
-**Arize Phoenix** is an open-source library designed for LLM tracing and evaluation. It enables developers to evaluate, experiment, and optimize AI products in real time. Key features include:
-
-- **Application Tracing**: Collects LLM application data with automatic instrumentation, providing comprehensive visibility into model operations
-- **Interactive Prompt Playground**: Offers a flexible environment for prompt and model iteration, allowing users to compare prompts, visualize outputs, and debug failures within their workflow
-- **Streamlined Evaluations and Annotations**: Facilitates efficient assessment and documentation of model performance
-
-Phoenix is built on OpenTelemetry, ensuring seamless setup, full transparency, and no vendor lock-in. It's perfect for teams who want to get started with LLM observability and evaluation in a fully local, open-source environment.
-
-## Arize AX
-
-**Arize AX** is the enterprise AI engineering platform that extends the capabilities of Phoenix, offering a comprehensive suite for development, evaluation, and observability. Key features include:
-
-- **Prompts**: Tools for prompt engineering, including a prompt playground, prompt hub for management and versioning, and a prompt builder powered by AI
-- **Experiments**: Systematic A/B testing of prompts against large datasets to optimize model performance
-- **Tracing**: Detailed tracing capabilities to monitor and debug AI applications effectively
-- **Evaluation**: Comprehensive evaluation metrics and tools to assess model outputs and performance
-- **Alyx**: An AI engineering agent embedded within the platform to assist with various tasks
-
-Arize AX is designed to support teams and organizations with larger data needs, providing robust support, collaboration features, and multiple deployment options including SaaS, Virtual Private Cloud (VPC), and Arize Private Connect.
-
-
-## Installation
-
-```bash
-npm install evals
-```
-
-The post-install script will automatically launch the CLI interface.
-
-## Usage
-
-Run the CLI manually:
-
-```bash
-npm start
-```
-
-Or use the binary:
-
-```bash
-npx evals
-```
-
+# evals
+
+An npm package for Arize evals functionality.
+
+## About Arize AI
+
+Arize AI is a leading company in AI observability and evaluation, dedicated to ensuring that artificial intelligence systems operate reliably in real-world applications. Founded in 2019, Arize provides tools that help machine learning teams monitor, troubleshoot, and improve model performance across various domains, including structured data, computer vision, and large language models (LLMs).
+
+## Arize Phoenix
+
+**Arize Phoenix** is an open-source library designed for LLM tracing and evaluation. It enables developers to evaluate, experiment, and optimize AI products in real time. Key features include:
+
+- **Application Tracing**: Collects LLM application data with automatic instrumentation, providing comprehensive visibility into model operations
+- **Interactive Prompt Playground**: Offers a flexible environment for prompt and model iteration, allowing users to compare prompts, visualize outputs, and debug failures within their workflow
+- **Streamlined Evaluations and Annotations**: Facilitates efficient assessment and documentation of model performance
+
+Phoenix is built on OpenTelemetry, ensuring seamless setup, full transparency, and no vendor lock-in. It's perfect for teams who want to get started with LLM observability and evaluation in a fully local, open-source environment.
+
+## Arize AX
+
+**Arize AX** is the enterprise AI engineering platform that extends the capabilities of Phoenix, offering a comprehensive suite for development, evaluation, and observability. Key features include:
+
+- **Prompts**: Tools for prompt engineering, including a prompt playground, prompt hub for management and versioning, and a prompt builder powered by AI
+- **Experiments**: Systematic A/B testing of prompts against large datasets to optimize model performance
+- **Tracing**: Detailed tracing capabilities to monitor and debug AI applications effectively
+- **Evaluation**: Comprehensive evaluation metrics and tools to assess model outputs and performance
+- **Alyx**: An AI engineering agent embedded within the platform to assist with various tasks
+
+Arize AX is designed to support teams and organizations with larger data needs, providing robust support, collaboration features, and multiple deployment options including SaaS, Virtual Private Cloud (VPC), and Arize Private Connect.
+
+
+## Installation
+
+```bash
+npm install evals
+```
+
+The post-install script will automatically launch the CLI interface.
+
+## Usage
+
+Run the CLI manually:
+
+```bash
+npm start
+```
+
+Or use the binary:
+
+```bash
+npx evals
+```
+

package/cli.js

@@ -1,246 +1,246 @@
-#!/usr/bin/env node
-
-import React, { useState, useEffect } from 'react';
-import { render, Box, Text, useInput, useApp, Static } from 'ink';
-import Gradient from 'ink-gradient';
-import { exec } from 'child_process';
-
-const e = React.createElement;
-
-// Diagnostic: Uncomment to verify script is running
-console.error('CLI script started, platform:', process.platform, 'isTTY:', process.stdin.isTTY);
-
-// "ARIZE EVALS" - EXACTLY matched width (both 44 chars)
-const largeLogo = `
- █████╗   ██████╗    ██╗  ███████╗  ███████╗
-██╔══██╗  ██╔══██╗   ██║  ╚══███╔╝  ██╔════╝
-███████║  ██████╔╝   ██║    ███╔╝   █████╗
-██╔══██║  ██╔══██╗   ██║   ███╔╝    ██╔══╝
-██║  ██║  ██║  ██║   ██║  ███████╗  ███████╗
-╚═╝  ╚═╝  ╚═╝  ╚═╝   ╚═╝  ╚══════╝  ╚══════╝
-███████╗ ██╗   ██╗  █████╗  ██╗     ███████╗
-██╔════╝ ██║   ██║ ██╔══██╗ ██║     ██╔════╝
-█████╗   ██║   ██║ ███████║ ██║     ███████╗
-██╔══╝   ╚██╗ ██╔╝ ██╔══██║ ██║     ╚════██║
-███████╗  ╚████╔╝  ██║  ██║ ███████╗███████║
-╚══════╝   ╚═══╝   ╚═╝  ╚═╝ ╚══════╝╚══════╝`;
-
-const mediumLogo = `
- █████╗   ██████╗    ██╗  ███████╗  ███████╗
-██╔══██╗  ██╔══██╗   ██║  ╚══███╔╝  ██╔════╝
-███████║  ██████╔╝   ██║    ███╔╝   █████╗
-██╔══██║  ██╔══██╗   ██║   ███╔╝    ██╔══╝
-██║  ██║  ██║  ██║   ██║  ███████╗  ███████╗
-╚═╝  ╚═╝  ╚═╝  ╚═╝   ╚═╝  ╚══════╝  ╚══════╝
-█▀▀ █ █ ▄▀█ █   █▀
-██▄ ▀▄▀ █▀█ █▄▄ ▄█`;
-
-const smallLogo = `ARIZE
-EVALS`;
-
-const OPTIONS = [
-  {
-    name: 'Evals Primer',
-    subtext: 'Learn the Craft',
-    url: 'https://arize.com/llm-evaluation/?utm_source=npmevals'
-  },
-  {
-    name: 'Arize AX SaaS',
-    subtext: 'Enterprise Agent Evaluation, Best in Class',
-    url: 'https://arize.com/docs/ax?utm_source=npmevals'
-  },
-  {
-    name: 'Arize Phoenix',
-    subtext: 'OSS Agent Evals & Traces, Fully Local',
-    url: 'https://arize.com/docs/phoenix?utm_source=npmevals'
-  },
-  {
-    name: 'Book an Eval Assessment',
-    subtext: 'Improve Your Agent Today',
-    url: 'https://arize.com/get-an-eval-assessment?utm_source=npmevals'
-  }
-];
-
-function openUrlInBrowser(url) {
-  const command = process.platform === 'win32'
-    ? `start ${url}`
-    : process.platform === 'darwin'
-      ? `open "${url}"`
-      : `xdg-open "${url}"`;
-
-  exec(command, (error) => {
-    if (error) {
-      console.error(`Failed to open browser: ${error.message}`);
-    }
-    process.exit(0);
-  });
-}
-
-// Header with hot pink to dark blue gradient
-function Header() {
-  const terminalWidth = process.stdout.columns || 80;
-
-  let logo;
-  if (terminalWidth >= 45) {
-    logo = largeLogo;
-  } else if (terminalWidth >= 35) {
-    logo = mediumLogo;
-  } else {
-    logo = smallLogo;
-  }
-
-  return e(Box, { flexDirection: 'column', marginBottom: 1 },
-    e(Gradient, { colors: ['#FF008C', '#FF1493', '#8B5CF6', '#1E3A8A'] },
-      e(Text, null, logo)
-    ),
-    e(Text, { color: 'gray', italic: true }, 'Evals and Observability for Agentic AI')
-  );
-}
-
-// Character-by-character shimmer component
-function ShimmerText({ text, shimmerPos }) {
-  const baseColor = '#FF1493';
-  const shimmerColors = ['#FF5AA7', '#FF85C0', '#FFB8D9', '#FFE0EE', '#FFB8D9', '#FF85C0', '#FF5AA7'];
-  const shimmerWidth = shimmerColors.length;
-
-  const chars = text.split('').map((char, i) => {
-    const distanceFromShimmer = i - shimmerPos;
-
-    let color = baseColor;
-    if (distanceFromShimmer >= 0 && distanceFromShimmer < shimmerWidth) {
-      color = shimmerColors[distanceFromShimmer];
-    }
-
-    return e(Text, { key: i, color, bold: true }, char);
-  });
-
-  return e(Box, null, ...chars);
-}
-
-// Menu item component - single line, hot pink selected with character shimmer
-function MenuItem({ name, subtext, isSelected, shimmerPos }) {
-  if (isSelected) {
-    return e(Box, null,
-      e(Text, { bold: true, color: '#FF1493' }, '❯ '),
-      e(ShimmerText, { text: name, shimmerPos }),
-      e(Text, { color: 'gray' }, ' — ' + subtext)
-    );
-  }
-
-  return e(Box, null,
-    e(Text, { color: 'white' }, '  ' + name),
-    e(Text, { color: 'gray' }, ' — ' + subtext)
-  );
-}
-
-// Main App component
-function App() {
-  const [selectedIndex, setSelectedIndex] = useState(0);
-  const [shimmerPos, setShimmerPos] = useState(-7); // Start off-screen
-  const { exit } = useApp();
-
-  const selectedName = OPTIONS[selectedIndex].name;
-
-  // Character-by-character shimmer animation
-  useEffect(() => {
-    const textLength = selectedName.length;
-    const shimmerWidth = 7;
-    const totalPositions = textLength + shimmerWidth + 5; // Extra padding for smooth loop
-
-    const timer = setInterval(() => {
-      setShimmerPos(prev => {
-        const next = prev + 1;
-        return next > totalPositions ? -shimmerWidth : next;
-      });
-    }, 120); // Speed of shimmer movement
-
-    return () => clearInterval(timer);
-  }, [selectedName]);
-
-  // Reset shimmer position when selection changes
-  useEffect(() => {
-    setShimmerPos(-7);
-  }, [selectedIndex]);
-
-  useInput((input, key) => {
-    if (input === 'q' || key.escape) {
-      exit();
-      return;
-    }
-
-    if (key.upArrow || input === 'k') {
-      setSelectedIndex(prev => (prev > 0 ? prev - 1 : OPTIONS.length - 1));
-    }
-
-    if (key.downArrow || input === 'j') {
-      setSelectedIndex(prev => (prev < OPTIONS.length - 1 ? prev + 1 : 0));
-    }
-
-    if (key.return) {
-      const selected = OPTIONS[selectedIndex];
-      exit();
-      setTimeout(() => {
-        console.log(`\nOpening ${selected.url} in your browser...\n`);
-        openUrlInBrowser(selected.url);
-      }, 100);
-    }
-  });
-
-  return e(Box, { flexDirection: 'column', padding: 1 },
-    e(Static, { items: ['header'] }, () => e(Header)),
-    e(Box, { marginBottom: 1 },
-      e(Text, { bold: true, color: 'white' }, 'Select an option:')
-    ),
-    e(Box, { flexDirection: 'column', paddingX: 1 },
-      ...OPTIONS.map((option, index) =>
-        e(MenuItem, {
-          key: index.toString(),
-          name: option.name,
-          subtext: option.subtext,
-          isSelected: index === selectedIndex,
-          shimmerPos: shimmerPos
-        })
-      )
-    ),
-    e(Box, { marginTop: 1 },
-      e(Text, { dimColor: true }, 'Enter to select · ↑↓ to navigate · Esc to cancel')
-    )
-  );
-}
-
-// Run the app
-// Check if we have an interactive terminal (required for Ink)
-if (!process.stdin.isTTY || !process.stdout.isTTY) {
-  console.error('Error: This command requires an interactive terminal.');
-  console.error('Please run `npx evals` from your terminal (not from a script or non-interactive environment).');
-  process.exit(1);
-}
-
-// Handle uncaught errors
-process.on('uncaughtException', (error) => {
-  console.error('Uncaught exception:', error.message);
-  console.error(error.stack);
-  process.exit(1);
-});
-
-process.on('unhandledRejection', (reason, promise) => {
-  console.error('Unhandled rejection at:', promise);
-  console.error('Reason:', reason);
-  process.exit(1);
-});
-
-// Ensure stdout is not buffered (important for Windows)
-if (process.stdout.isTTY) {
-  process.stdout.setEncoding('utf8');
-}
-
-try {
-  render(e(App), { patchConsole: false });
-} catch (error) {
-  console.error('Failed to render interactive CLI:', error.message);
-  console.error(error.stack);
-  if (process.platform === 'win32') {
-    console.error('\nNote: If you\'re using cmd.exe, try running in PowerShell or Git Bash instead.');
-  }
-  process.exit(1);
-}
+#!/usr/bin/env node
+
+import React, { useState, useEffect } from 'react';
+import { render, Box, Text, useInput, useApp, Static } from 'ink';
+import Gradient from 'ink-gradient';
+import { exec } from 'child_process';
+
+const e = React.createElement;
+
+// Diagnostic: Uncomment to verify script is running
+console.error('CLI script started, platform:', process.platform, 'isTTY:', process.stdin.isTTY);
+
+// "ARIZE EVALS" - EXACTLY matched width (both 44 chars)
+const largeLogo = `
+ █████╗   ██████╗    ██╗  ███████╗  ███████╗
+██╔══██╗  ██╔══██╗   ██║  ╚══███╔╝  ██╔════╝
+███████║  ██████╔╝   ██║    ███╔╝   █████╗
+██╔══██║  ██╔══██╗   ██║   ███╔╝    ██╔══╝
+██║  ██║  ██║  ██║   ██║  ███████╗  ███████╗
+╚═╝  ╚═╝  ╚═╝  ╚═╝   ╚═╝  ╚══════╝  ╚══════╝
+███████╗ ██╗   ██╗  █████╗  ██╗     ███████╗
+██╔════╝ ██║   ██║ ██╔══██╗ ██║     ██╔════╝
+█████╗   ██║   ██║ ███████║ ██║     ███████╗
+██╔══╝   ╚██╗ ██╔╝ ██╔══██║ ██║     ╚════██║
+███████╗  ╚████╔╝  ██║  ██║ ███████╗███████║
+╚══════╝   ╚═══╝   ╚═╝  ╚═╝ ╚══════╝╚══════╝`;
+
+const mediumLogo = `
+ █████╗   ██████╗    ██╗  ███████╗  ███████╗
+██╔══██╗  ██╔══██╗   ██║  ╚══███╔╝  ██╔════╝
+███████║  ██████╔╝   ██║    ███╔╝   █████╗
+██╔══██║  ██╔══██╗   ██║   ███╔╝    ██╔══╝
+██║  ██║  ██║  ██║   ██║  ███████╗  ███████╗
+╚═╝  ╚═╝  ╚═╝  ╚═╝   ╚═╝  ╚══════╝  ╚══════╝
+█▀▀ █ █ ▄▀█ █   █▀
+██▄ ▀▄▀ █▀█ █▄▄ ▄█`;
+
+const smallLogo = `ARIZE
+EVALS`;
+
+const OPTIONS = [
+  {
+    name: 'Evals Primer',
+    subtext: 'Learn the Craft',
+    url: 'https://arize.com/llm-evaluation/?utm_source=npmevals'
+  },
+  {
+    name: 'Arize AX SaaS',
+    subtext: 'Enterprise Agent Evaluation, Best in Class',
+    url: 'https://arize.com/docs/ax?utm_source=npmevals'
+  },
+  {
+    name: 'Arize Phoenix',
+    subtext: 'OSS Agent Evals & Traces, Fully Local',
+    url: 'https://arize.com/docs/phoenix?utm_source=npmevals'
+  },
+  {
+    name: 'Book an Eval Assessment',
+    subtext: 'Improve Your Agent Today',
+    url: 'https://arize.com/get-an-eval-assessment?utm_source=npmevals'
+  }
+];
+
+function openUrlInBrowser(url) {
+  const command = process.platform === 'win32'
+    ? `start ${url}`
+    : process.platform === 'darwin'
+      ? `open "${url}"`
+      : `xdg-open "${url}"`;
+
+  exec(command, (error) => {
+    if (error) {
+      console.error(`Failed to open browser: ${error.message}`);
+    }
+    process.exit(0);
+  });
+}
+
+// Header with hot pink to dark blue gradient
+function Header() {
+  const terminalWidth = process.stdout.columns || 80;
+
+  let logo;
+  if (terminalWidth >= 45) {
+    logo = largeLogo;
+  } else if (terminalWidth >= 35) {
+    logo = mediumLogo;
+  } else {
+    logo = smallLogo;
+  }
+
+  return e(Box, { flexDirection: 'column', marginBottom: 1 },
+    e(Gradient, { colors: ['#FF008C', '#FF1493', '#8B5CF6', '#1E3A8A'] },
+      e(Text, null, logo)
+    ),
+    e(Text, { color: 'gray', italic: true }, 'Evals and Observability for Agentic AI')
+  );
+}
+
+// Character-by-character shimmer component
+function ShimmerText({ text, shimmerPos }) {
+  const baseColor = '#FF1493';
+  const shimmerColors = ['#FF5AA7', '#FF85C0', '#FFB8D9', '#FFE0EE', '#FFB8D9', '#FF85C0', '#FF5AA7'];
+  const shimmerWidth = shimmerColors.length;
+
+  const chars = text.split('').map((char, i) => {
+    const distanceFromShimmer = i - shimmerPos;
+
+    let color = baseColor;
+    if (distanceFromShimmer >= 0 && distanceFromShimmer < shimmerWidth) {
+      color = shimmerColors[distanceFromShimmer];
+    }
+
+    return e(Text, { key: i, color, bold: true }, char);
+  });
+
+  return e(Box, null, ...chars);
+}
+
+// Menu item component - single line, hot pink selected with character shimmer
+function MenuItem({ name, subtext, isSelected, shimmerPos }) {
+  if (isSelected) {
+    return e(Box, null,
+      e(Text, { bold: true, color: '#FF1493' }, '❯ '),
+      e(ShimmerText, { text: name, shimmerPos }),
+      e(Text, { color: 'gray' }, ' — ' + subtext)
+    );
+  }
+
+  return e(Box, null,
+    e(Text, { color: 'white' }, '  ' + name),
+    e(Text, { color: 'gray' }, ' — ' + subtext)
+  );
+}
+
+// Main App component
+function App() {
+  const [selectedIndex, setSelectedIndex] = useState(0);
+  const [shimmerPos, setShimmerPos] = useState(-7); // Start off-screen
+  const { exit } = useApp();
+
+  const selectedName = OPTIONS[selectedIndex].name;
+
+  // Character-by-character shimmer animation
+  useEffect(() => {
+    const textLength = selectedName.length;
+    const shimmerWidth = 7;
+    const totalPositions = textLength + shimmerWidth + 5; // Extra padding for smooth loop
+
+    const timer = setInterval(() => {
+      setShimmerPos(prev => {
+        const next = prev + 1;
+        return next > totalPositions ? -shimmerWidth : next;
+      });
+    }, 120); // Speed of shimmer movement
+
+    return () => clearInterval(timer);
+  }, [selectedName]);
+
+  // Reset shimmer position when selection changes
+  useEffect(() => {
+    setShimmerPos(-7);
+  }, [selectedIndex]);
+
+  useInput((input, key) => {
+    if (input === 'q' || key.escape) {
+      exit();
+      return;
+    }
+
+    if (key.upArrow || input === 'k') {
+      setSelectedIndex(prev => (prev > 0 ? prev - 1 : OPTIONS.length - 1));
+    }
+
+    if (key.downArrow || input === 'j') {
+      setSelectedIndex(prev => (prev < OPTIONS.length - 1 ? prev + 1 : 0));
+    }
+
+    if (key.return) {
+      const selected = OPTIONS[selectedIndex];
+      exit();
+      setTimeout(() => {
+        console.log(`\nOpening ${selected.url} in your browser...\n`);
+        openUrlInBrowser(selected.url);
+      }, 100);
+    }
+  });
+
+  return e(Box, { flexDirection: 'column', padding: 1 },
+    e(Static, { items: ['header'] }, () => e(Header)),
+    e(Box, { marginBottom: 1 },
+      e(Text, { bold: true, color: 'white' }, 'Select an option:')
+    ),
+    e(Box, { flexDirection: 'column', paddingX: 1 },
+      ...OPTIONS.map((option, index) =>
+        e(MenuItem, {
+          key: index.toString(),
+          name: option.name,
+          subtext: option.subtext,
+          isSelected: index === selectedIndex,
+          shimmerPos: shimmerPos
+        })
+      )
+    ),
+    e(Box, { marginTop: 1 },
+      e(Text, { dimColor: true }, 'Enter to select · ↑↓ to navigate · Esc to cancel')
+    )
+  );
+}
+
+// Run the app
+// Check if we have an interactive terminal (required for Ink)
+if (!process.stdin.isTTY || !process.stdout.isTTY) {
+  console.error('Error: This command requires an interactive terminal.');
+  console.error('Please run `npx evals` from your terminal (not from a script or non-interactive environment).');
+  process.exit(1);
+}
+
+// Handle uncaught errors
+process.on('uncaughtException', (error) => {
+  console.error('Uncaught exception:', error.message);
+  console.error(error.stack);
+  process.exit(1);
+});
+
+process.on('unhandledRejection', (reason, promise) => {
+  console.error('Unhandled rejection at:', promise);
+  console.error('Reason:', reason);
+  process.exit(1);
+});
+
+// Ensure stdout is not buffered (important for Windows)
+if (process.stdout.isTTY) {
+  process.stdout.setEncoding('utf8');
+}
+
+try {
+  render(e(App), { patchConsole: false });
+} catch (error) {
+  console.error('Failed to render interactive CLI:', error.message);
+  console.error(error.stack);
+  if (process.platform === 'win32') {
+    console.error('\nNote: If you\'re using cmd.exe, try running in PowerShell or Git Bash instead.');
+  }
+  process.exit(1);
+}

package/package.json

@@ -1,27 +1,27 @@
-{
-  "name": "evals",
-  "version": "2.2.7",
-  "description": "Arize evals package",
-  "type": "module",
-  "main": "cli.js",
-  "scripts": {
-    "postinstall": "echo \"run npx evals to complete setup\" && node spawn-interactive.js",
-    "start": "node cli.js"
-  },
-  "dependencies": {
-    "@arizeai/phoenix-evals": "^0.6.5",
-    "@arizeai/phoenix-otel": "*",
-    "ink": "^6.0.0",
-    "ink-gradient": "^3.0.0",
-    "react": "^19.0.0"
-  },
-  "bin": {
-    "evals": "./cli.js"
-  },
-  "keywords": [
-    "arize",
-    "evals"
-  ],
-  "author": "",
-  "license": "ISC"
-}
+{
+  "name": "evals",
+  "version": "2.2.8",
+  "description": "Arize evals package",
+  "type": "module",
+  "main": "cli.js",
+  "scripts": {
+    "postinstall": "echo \"run npx evals to complete setup\" && node spawn-interactive.js",
+    "start": "node cli.js"
+  },
+  "dependencies": {
+    "@arizeai/phoenix-evals": "^0.6.5",
+    "@arizeai/phoenix-otel": "*",
+    "ink": "^6.0.0",
+    "ink-gradient": "^3.0.0",
+    "react": "^19.0.0"
+  },
+  "bin": {
+    "evals": "./cli.js"
+  },
+  "keywords": [
+    "arize",
+    "evals"
+  ],
+  "author": "",
+  "license": "ISC"
+}

package/spawn-interactive.js

@@ -1,51 +1,51 @@
-import { spawn } from 'child_process';
-import { openSync } from 'fs';
-
-function tryOpenTTY() {
-  if (process.platform === 'win32') {
-    // On Windows, avoid CON device (issues in Node.js 22)
-    // Use isTTY check instead, and stdio: 'inherit' for spawning
-    if (process.stdin.isTTY && process.stdout.isTTY && process.stderr.isTTY) {
-      return { available: true, useInherit: true };
-    }
-    return { available: false };
-  } else {
-    // On macOS/Linux, use direct /dev/tty access
-    try {
-      const device = '/dev/tty';
-      openSync(device, 'r');
-      return { device, available: true, useInherit: false };
-    } catch (err) {
-      return { available: false };
-    }
-  }
-}
-
-const tty = tryOpenTTY();
-
-if (tty.available) {
-  const spawnOptions = tty.useInherit
-    ? { stdio: 'inherit', shell: true }
-    : {
-        stdio: [
-          openSync(tty.device, 'r'),
-          openSync(tty.device, 'w'),
-          openSync(tty.device, 'w')
-        ]
-      };
-
-  const child = spawn('node', ['./cli.js'], spawnOptions);
-
-  child.on('close', (code) => {
-    process.exit(code || 0);
-  });
-
-  child.on('error', (err) => {
-    console.error('Failed to start interactive CLI:', err.message);
-    console.log('Run `npx evals` to complete setup.');
-    process.exit(1);
-  });
-} else {
-  console.log('Run `npx evals` to complete setup.');
-  process.exit(1);
+import { spawn } from 'child_process';
+import { openSync } from 'fs';
+
+function tryOpenTTY() {
+  if (process.platform === 'win32') {
+    // On Windows, avoid CON device (issues in Node.js 22)
+    // Use isTTY check instead, and stdio: 'inherit' for spawning
+    if (process.stdin.isTTY && process.stdout.isTTY && process.stderr.isTTY) {
+      return { available: true, useInherit: true };
+    }
+    return { available: false };
+  } else {
+    // On macOS/Linux, use direct /dev/tty access
+    try {
+      const device = '/dev/tty';
+      openSync(device, 'r');
+      return { device, available: true, useInherit: false };
+    } catch (err) {
+      return { available: false };
+    }
+  }
+}
+
+const tty = tryOpenTTY();
+
+if (tty.available) {
+  const spawnOptions = tty.useInherit
+    ? { stdio: 'inherit', shell: true }
+    : {
+        stdio: [
+          openSync(tty.device, 'r'),
+          openSync(tty.device, 'w'),
+          openSync(tty.device, 'w')
+        ]
+      };
+
+  const child = spawn('node', ['./cli.js'], spawnOptions);
+
+  child.on('close', (code) => {
+    process.exit(code || 0);
+  });
+
+  child.on('error', (err) => {
+    console.error('Failed to start interactive CLI:', err.message);
+    console.log('Run `npx evals` to complete setup.');
+    // process.exit(1);
+  });
+} else {
+  console.log('Run `npx evals` to complete setup.');
+  // process.exit(1);
 }