agentv 5.3.0-next.1 → 5.3.2-next.1

package/dist/interactive-RUY3OCBI.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["../src/commands/eval/interactive.ts","../src/commands/eval/last-config.ts"],"sourcesContent":["import { existsSync } from 'node:fs';\nimport path from 'node:path';\nimport { listTargetNames, readTargetDefinitions } from '@agentv/core';\nimport { checkbox, confirm, number, search, select } from '@inquirer/prompts';\n\nimport { TARGET_FILE_CANDIDATES, fileExists } from '../../utils/targets.js';\nimport {\n  type DiscoveredEvalFile,\n  discoverEvalFiles,\n  filterByCategory,\n  getCategories,\n} from './discover.js';\nimport { type LastConfig, loadLastConfig, saveLastConfig } from './last-config.js';\nimport { resolveExistingRunPrimaryPath } from './result-layout.js';\nimport { runEvalCommand } from './run-eval.js';\nimport { findRepoRoot } from './shared.js';\n\nconst ANSI_BOLD = '\\x1b[1m';\nconst ANSI_DIM = '\\x1b[2m';\nconst ANSI_CYAN = '\\x1b[36m';\nconst ANSI_GREEN = '\\x1b[32m';\nconst ANSI_RESET = '\\x1b[0m';\n\nexport interface InteractiveConfig {\n  readonly evalPaths: readonly string[];\n  readonly target: string;\n  readonly workers: number;\n  readonly cache: boolean;\n}\n\n/**\n * Launch the interactive wizard when `agentv eval` is called with no arguments.\n */\nexport async function launchInteractiveWizard(): Promise<void> {\n  const cwd = process.cwd();\n\n  console.log(`\\n${ANSI_BOLD}${ANSI_CYAN}AgentV Interactive Mode${ANSI_RESET}\\n`);\n\n  const lastConfig = await loadLastConfig();\n  const action = await promptMainMenu(lastConfig);\n\n  if (action === 'exit') {\n    return;\n  }\n\n  if (action === 'resume' && lastConfig?.outputDir) {\n    const relativeDir = path.relative(cwd, lastConfig.outputDir) || lastConfig.outputDir;\n    console.log(`\\n${ANSI_DIM}Resuming run at ${relativeDir}...${ANSI_RESET}\\n`);\n    await executeConfig(\n      {\n        evalPaths: lastConfig.evalPaths,\n        target: lastConfig.target,\n        workers: lastConfig.workers,\n        cache: lastConfig.cache,\n      },\n      { resumeOutputDir: lastConfig.outputDir },\n    );\n    return;\n  }\n\n  if (action === 'rerun' && lastConfig) {\n    console.log(`\\n${ANSI_DIM}Rerunning last configuration...${ANSI_RESET}\\n`);\n    await executeConfig({\n      evalPaths: lastConfig.evalPaths,\n      target: lastConfig.target,\n      workers: lastConfig.workers,\n      cache: lastConfig.cache,\n    });\n    return;\n  }\n\n  // Run new evaluation flow\n  const config = await promptNewEvaluation(cwd);\n  if (!config) {\n    return;\n  }\n\n  // Review & confirm\n  const confirmed = await promptReviewAndConfirm(config, cwd);\n  if (!confirmed) {\n    return;\n  }\n\n  await executeConfig(config);\n}\n\nasync function promptMainMenu(\n  lastConfig: LastConfig | undefined,\n): Promise<'new' | 'rerun' | 'resume' | 'exit'> {\n  type MenuChoice = 'new' | 'rerun' | 'resume' | 'exit';\n  const choices: Array<{ name: string; value: MenuChoice; description?: string }> = [];\n\n  // Resume entry: only when the prior run has a known artifact dir with a manifest.\n  if (lastConfig?.outputDir) {\n    const indexPath = resolveExistingRunPrimaryPath(lastConfig.outputDir);\n    if (indexPath && existsSync(indexPath)) {\n      const dirLabel = path.basename(lastConfig.outputDir);\n      choices.push({\n        name: '⏯  Resume last run',\n        value: 'resume',\n        description: `${dirLabel} (target: ${lastConfig.target})`,\n      });\n    }\n  }\n\n  if (lastConfig) {\n    const evalCount = lastConfig.evalPaths.length;\n    choices.push({\n      name: '🔄 Rerun last config',\n      value: 'rerun',\n      description: `${evalCount} eval file(s), target: ${lastConfig.target}`,\n    });\n  }\n\n  choices.push({ name: '🚀 Run new evaluation', value: 'new' }, { name: '✕ Exit', value: 'exit' });\n\n  return select<MenuChoice>({\n    message: 'What would you like to do?',\n    choices,\n  });\n}\n\nasync function promptNewEvaluation(cwd: string): Promise<InteractiveConfig | undefined> {\n  // Step 1: Discover eval files\n  console.log(`\\n${ANSI_DIM}Scanning for eval files...${ANSI_RESET}`);\n  const allFiles = await discoverEvalFiles(cwd);\n\n  if (allFiles.length === 0) {\n    console.log(\n      '\\n⚠  No eval files found in the current directory.\\n' +\n        '   Place .yaml or .jsonl eval files in your project, or use:\\n' +\n        '   agentv eval <path-to-eval.yaml>\\n',\n    );\n    return undefined;\n  }\n\n  console.log(`${ANSI_DIM}Found ${allFiles.length} eval file(s)${ANSI_RESET}\\n`);\n\n  // Step 2: Select eval files (optionally filter by category first)\n  const selectedFiles = await promptEvalSelection(allFiles);\n  if (selectedFiles.length === 0) {\n    console.log('\\nNo eval files selected.');\n    return undefined;\n  }\n\n  // Step 3: Select target\n  const target = await promptTargetSelection(cwd, selectedFiles[0].path);\n\n  // Step 4: Advanced options\n  const advanced = await promptAdvancedOptions();\n\n  return {\n    evalPaths: selectedFiles.map((f) => f.path),\n    target,\n    ...advanced,\n  };\n}\n\nasync function promptEvalSelection(\n  allFiles: readonly DiscoveredEvalFile[],\n): Promise<DiscoveredEvalFile[]> {\n  const categories = getCategories(allFiles);\n\n  // If only one category or few files, skip category selection\n  let filesToSelect: readonly DiscoveredEvalFile[];\n\n  if (categories.length > 1) {\n    const selectedCategory = await search<string>({\n      message: 'Select a category (type to search)',\n      source: async (term) => {\n        const filtered = term\n          ? categories.filter((c) => c.toLowerCase().includes(term.toLowerCase()))\n          : categories;\n        return [\n          { name: '(all categories)', value: '__all__' },\n          ...filtered.map((c) => {\n            const count = filterByCategory(allFiles, c).length;\n            return { name: `${c} (${count} file${count > 1 ? 's' : ''})`, value: c };\n          }),\n        ];\n      },\n    });\n\n    filesToSelect =\n      selectedCategory === '__all__' ? allFiles : filterByCategory(allFiles, selectedCategory);\n  } else {\n    filesToSelect = allFiles;\n  }\n\n  return checkbox<DiscoveredEvalFile>({\n    message: 'Select eval files to run (space to toggle, enter to confirm)',\n    choices: filesToSelect.map((f) => ({\n      name: f.relativePath,\n      value: f,\n      checked: filesToSelect.length <= 5, // auto-select if few files\n    })),\n    required: true,\n  });\n}\n\nasync function promptTargetSelection(cwd: string, firstEvalPath: string): Promise<string> {\n  const repoRoot = await findRepoRoot(cwd);\n\n  // Try to find targets.yaml — search near the eval file first, then cwd/repoRoot\n  const targetsPath = await findTargetsFile(cwd, repoRoot, firstEvalPath);\n\n  if (!targetsPath) {\n    console.log(`${ANSI_DIM}No targets.yaml found. Using default target.${ANSI_RESET}`);\n    return 'default';\n  }\n\n  const definitions = await readTargetDefinitions(targetsPath);\n  const targetNames = listTargetNames(definitions);\n\n  if (targetNames.length === 0) {\n    return 'default';\n  }\n\n  if (targetNames.length === 1) {\n    console.log(`${ANSI_DIM}Using target: ${targetNames[0]}${ANSI_RESET}`);\n    return targetNames[0];\n  }\n\n  return search<string>({\n    message: 'Select a target (type to search)',\n    source: async (term) => {\n      const filtered = term\n        ? targetNames.filter((t) => t.toLowerCase().includes(term.toLowerCase()))\n        : targetNames;\n      return filtered.map((t) => {\n        const def = definitions.find((d) => d.name === t);\n        return {\n          name: t,\n          value: t,\n          description: def ? `provider: ${def.provider}` : undefined,\n        };\n      });\n    },\n  });\n}\n\nasync function findTargetsFile(\n  cwd: string,\n  repoRoot: string,\n  evalFilePath?: string,\n): Promise<string | undefined> {\n  // Build directory chain: eval file dir → cwd → repoRoot (mirrors discoverTargetsFile)\n  const dirsToSearch: string[] = [];\n\n  if (evalFilePath) {\n    const evalDir = path.dirname(evalFilePath);\n    if (!dirsToSearch.includes(evalDir)) {\n      dirsToSearch.push(evalDir);\n    }\n  }\n\n  if (!dirsToSearch.includes(cwd)) {\n    dirsToSearch.push(cwd);\n  }\n\n  if (repoRoot !== cwd && !dirsToSearch.includes(repoRoot)) {\n    dirsToSearch.push(repoRoot);\n  }\n\n  for (const dir of dirsToSearch) {\n    for (const candidate of TARGET_FILE_CANDIDATES) {\n      const fullPath = `${dir}/${candidate}`;\n      if (await fileExists(fullPath)) {\n        return fullPath;\n      }\n    }\n  }\n\n  return undefined;\n}\n\nasync function promptAdvancedOptions(): Promise<{\n  workers: number;\n  cache: boolean;\n}> {\n  const customize = await confirm({\n    message: 'Configure advanced options?',\n    default: false,\n  });\n\n  if (!customize) {\n    return { workers: 3, cache: false };\n  }\n\n  const workers =\n    (await number({\n      message: 'Number of parallel workers (1-50)',\n      default: 3,\n      min: 1,\n      max: 50,\n    })) ?? 3;\n\n  const cache = await confirm({\n    message: 'Enable response cache?',\n    default: false,\n  });\n\n  return { workers, cache };\n}\n\nasync function promptReviewAndConfirm(config: InteractiveConfig, cwd: string): Promise<boolean> {\n  const evalDisplay = config.evalPaths\n    .map((p) => {\n      const rel = p.startsWith(cwd) ? p.slice(cwd.length + 1) : p;\n      return `  ${rel}`;\n    })\n    .join('\\n');\n\n  console.log(`\\n${ANSI_BOLD}Review Configuration${ANSI_RESET}`);\n  console.log(`${ANSI_DIM}${'─'.repeat(40)}${ANSI_RESET}`);\n  console.log(`${ANSI_GREEN}Eval files:${ANSI_RESET}\\n${evalDisplay}`);\n  console.log(`${ANSI_GREEN}Target:${ANSI_RESET}    ${config.target}`);\n  console.log(`${ANSI_GREEN}Workers:${ANSI_RESET}   ${config.workers}`);\n  console.log(`${ANSI_GREEN}Cache:${ANSI_RESET}     ${config.cache ? 'yes' : 'no'}`);\n  console.log(`${ANSI_DIM}${'─'.repeat(40)}${ANSI_RESET}`);\n\n  return confirm({\n    message: 'Run evaluation with this configuration?',\n    default: true,\n  });\n}\n\nasync function executeConfig(\n  config: InteractiveConfig,\n  opts?: { resumeOutputDir?: string },\n): Promise<void> {\n  const cwd = process.cwd();\n  const rawOptions: Record<string, unknown> = {\n    target: config.target,\n    workers: config.workers,\n    cache: config.cache,\n    ...(opts?.resumeOutputDir ? { output: opts.resumeOutputDir, resume: true } : {}),\n    agentTimeout: 120,\n    maxRetries: 2,\n    verbose: false,\n    keepWorkspaces: false,\n    cleanupWorkspaces: false,\n    trace: false,\n  };\n\n  const result = await runEvalCommand({\n    testFiles: [...config.evalPaths],\n    rawOptions,\n  });\n\n  // Persist config with the resolved artifact dir so the wizard can offer\n  // \"Resume last run\" on the next invocation. Done after a successful run so\n  // the saved outputDir always points at a real run manifest.\n  if (result) {\n    await saveLastConfig({\n      timestamp: new Date().toISOString(),\n      cwd,\n      evalPaths: config.evalPaths,\n      target: config.target,\n      workers: config.workers,\n      cache: config.cache,\n      outputDir: path.dirname(result.outputPath),\n    });\n  }\n\n  // Prompt to retry errors when execution errors were detected in a TTY\n  if (result && result.executionErrorCount > 0 && process.stdin.isTTY) {\n    await promptRetryErrors(config, result.outputPath);\n  }\n}\n\nasync function promptRetryErrors(config: InteractiveConfig, outputPath: string): Promise<void> {\n  const shouldRetry = await confirm({\n    message: `Retry ${ANSI_BOLD}execution errors${ANSI_RESET} from this run?`,\n    default: true,\n  });\n\n  if (!shouldRetry) {\n    return;\n  }\n\n  console.log(`\\n${ANSI_DIM}Retrying execution errors...${ANSI_RESET}\\n`);\n\n  const rawOptions: Record<string, unknown> = {\n    target: config.target,\n    workers: config.workers,\n    cache: config.cache,\n    retryErrors: outputPath,\n    out: outputPath,\n    agentTimeout: 120,\n    maxRetries: 2,\n    verbose: false,\n    keepWorkspaces: false,\n    cleanupWorkspaces: false,\n    trace: false,\n  };\n\n  const retryResult = await runEvalCommand({\n    testFiles: [...config.evalPaths],\n    rawOptions,\n  });\n\n  // Allow chained retries if there are still errors\n  if (retryResult && retryResult.executionErrorCount > 0 && process.stdin.isTTY) {\n    await promptRetryErrors(config, retryResult.outputPath);\n  }\n}\n","import { mkdir, readFile, writeFile } from 'node:fs/promises';\nimport path from 'node:path';\nimport { getAgentvConfigDir } from '@agentv/core';\n\nconst CONFIG_DIR = getAgentvConfigDir();\nconst LAST_CONFIG_PATH = path.join(CONFIG_DIR, 'last-config.json');\n\nexport interface LastConfig {\n  readonly timestamp: string;\n  readonly cwd: string;\n  readonly evalPaths: readonly string[];\n  readonly target: string;\n  readonly workers: number;\n  readonly cache: boolean;\n  /**\n   * Resolved artifact directory of the last completed wizard run. Used to\n   * power the wizard's \"Resume last run\" entry. Optional for backward\n   * compatibility with configs saved before this field existed.\n   */\n  readonly outputDir?: string;\n}\n\nexport async function loadLastConfig(): Promise<LastConfig | undefined> {\n  try {\n    const content = await readFile(LAST_CONFIG_PATH, 'utf-8');\n    return JSON.parse(content) as LastConfig;\n  } catch {\n    return undefined;\n  }\n}\n\nexport async function saveLastConfig(config: LastConfig): Promise<void> {\n  await mkdir(CONFIG_DIR, { recursive: true });\n  await writeFile(LAST_CONFIG_PATH, JSON.stringify(config, null, 2), 'utf-8');\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;AAAA,SAAS,kBAAkB;AAC3B,OAAOA,WAAU;AAEjB,SAAS,UAAU,SAAS,QAAQ,QAAQ,cAAc;;;ACH1D,SAAS,OAAO,UAAU,iBAAiB;AAC3C,OAAO,UAAU;AAGjB,IAAM,aAAa,mBAAmB;AACtC,IAAM,mBAAmB,KAAK,KAAK,YAAY,kBAAkB;AAiBjE,eAAsB,iBAAkD;AACtE,MAAI;AACF,UAAM,UAAU,MAAM,SAAS,kBAAkB,OAAO;AACxD,WAAO,KAAK,MAAM,OAAO;AAAA,EAC3B,QAAQ;AACN,WAAO;AAAA,EACT;AACF;AAEA,eAAsB,eAAe,QAAmC;AACtE,QAAM,MAAM,YAAY,EAAE,WAAW,KAAK,CAAC;AAC3C,QAAM,UAAU,kBAAkB,KAAK,UAAU,QAAQ,MAAM,CAAC,GAAG,OAAO;AAC5E;;;ADjBA,IAAM,YAAY;AAClB,IAAM,WAAW;AACjB,IAAM,YAAY;AAClB,IAAM,aAAa;AACnB,IAAM,aAAa;AAYnB,eAAsB,0BAAyC;AAC7D,QAAM,MAAM,QAAQ,IAAI;AAExB,UAAQ,IAAI;AAAA,EAAK,SAAS,GAAG,SAAS,0BAA0B,UAAU;AAAA,CAAI;AAE9E,QAAM,aAAa,MAAM,eAAe;AACxC,QAAM,SAAS,MAAM,eAAe,UAAU;AAE9C,MAAI,WAAW,QAAQ;AACrB;AAAA,EACF;AAEA,MAAI,WAAW,YAAY,YAAY,WAAW;AAChD,UAAM,cAAcC,MAAK,SAAS,KAAK,WAAW,SAAS,KAAK,WAAW;AAC3E,YAAQ,IAAI;AAAA,EAAK,QAAQ,mBAAmB,WAAW,MAAM,UAAU;AAAA,CAAI;AAC3E,UAAM;AAAA,MACJ;AAAA,QACE,WAAW,WAAW;AAAA,QACtB,QAAQ,WAAW;AAAA,QACnB,SAAS,WAAW;AAAA,QACpB,OAAO,WAAW;AAAA,MACpB;AAAA,MACA,EAAE,iBAAiB,WAAW,UAAU;AAAA,IAC1C;AACA;AAAA,EACF;AAEA,MAAI,WAAW,WAAW,YAAY;AACpC,YAAQ,IAAI;AAAA,EAAK,QAAQ,kCAAkC,UAAU;AAAA,CAAI;AACzE,UAAM,cAAc;AAAA,MAClB,WAAW,WAAW;AAAA,MACtB,QAAQ,WAAW;AAAA,MACnB,SAAS,WAAW;AAAA,MACpB,OAAO,WAAW;AAAA,IACpB,CAAC;AACD;AAAA,EACF;AAGA,QAAM,SAAS,MAAM,oBAAoB,GAAG;AAC5C,MAAI,CAAC,QAAQ;AACX;AAAA,EACF;AAGA,QAAM,YAAY,MAAM,uBAAuB,QAAQ,GAAG;AAC1D,MAAI,CAAC,WAAW;AACd;AAAA,EACF;AAEA,QAAM,cAAc,MAAM;AAC5B;AAEA,eAAe,eACb,YAC8C;AAE9C,QAAM,UAA4E,CAAC;AAGnF,MAAI,YAAY,WAAW;AACzB,UAAM,YAAY,8BAA8B,WAAW,SAAS;AACpE,QAAI,aAAa,WAAW,SAAS,GAAG;AACtC,YAAM,WAAWA,MAAK,SAAS,WAAW,SAAS;AACnD,cAAQ,KAAK;AAAA,QACX,MAAM;AAAA,QACN,OAAO;AAAA,QACP,aAAa,GAAG,QAAQ,aAAa,WAAW,MAAM;AAAA,MACxD,CAAC;AAAA,IACH;AAAA,EACF;AAEA,MAAI,YAAY;AACd,UAAM,YAAY,WAAW,UAAU;AACvC,YAAQ,KAAK;AAAA,MACX,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aAAa,GAAG,SAAS,0BAA0B,WAAW,MAAM;AAAA,IACtE,CAAC;AAAA,EACH;AAEA,UAAQ,KAAK,EAAE,MAAM,gCAAyB,OAAO,MAAM,GAAG,EAAE,MAAM,eAAU,OAAO,OAAO,CAAC;AAE/F,SAAO,OAAmB;AAAA,IACxB,SAAS;AAAA,IACT;AAAA,EACF,CAAC;AACH;AAEA,eAAe,oBAAoB,KAAqD;AAEtF,UAAQ,IAAI;AAAA,EAAK,QAAQ,6BAA6B,UAAU,EAAE;AAClE,QAAM,WAAW,MAAM,kBAAkB,GAAG;AAE5C,MAAI,SAAS,WAAW,GAAG;AACzB,YAAQ;AAAA,MACN;AAAA,IAGF;AACA,WAAO;AAAA,EACT;AAEA,UAAQ,IAAI,GAAG,QAAQ,SAAS,SAAS,MAAM,gBAAgB,UAAU;AAAA,CAAI;AAG7E,QAAM,gBAAgB,MAAM,oBAAoB,QAAQ;AACxD,MAAI,cAAc,WAAW,GAAG;AAC9B,YAAQ,IAAI,2BAA2B;AACvC,WAAO;AAAA,EACT;AAGA,QAAM,SAAS,MAAM,sBAAsB,KAAK,cAAc,CAAC,EAAE,IAAI;AAGrE,QAAM,WAAW,MAAM,sBAAsB;AAE7C,SAAO;AAAA,IACL,WAAW,cAAc,IAAI,CAAC,MAAM,EAAE,IAAI;AAAA,IAC1C;AAAA,IACA,GAAG;AAAA,EACL;AACF;AAEA,eAAe,oBACb,UAC+B;AAC/B,QAAM,aAAa,cAAc,QAAQ;AAGzC,MAAI;AAEJ,MAAI,WAAW,SAAS,GAAG;AACzB,UAAM,mBAAmB,MAAM,OAAe;AAAA,MAC5C,SAAS;AAAA,MACT,QAAQ,OAAO,SAAS;AACtB,cAAM,WAAW,OACb,WAAW,OAAO,CAAC,MAAM,EAAE,YAAY,EAAE,SAAS,KAAK,YAAY,CAAC,CAAC,IACrE;AACJ,eAAO;AAAA,UACL,EAAE,MAAM,oBAAoB,OAAO,UAAU;AAAA,UAC7C,GAAG,SAAS,IAAI,CAAC,MAAM;AACrB,kBAAM,QAAQ,iBAAiB,UAAU,CAAC,EAAE;AAC5C,mBAAO,EAAE,MAAM,GAAG,CAAC,KAAK,KAAK,QAAQ,QAAQ,IAAI,MAAM,EAAE,KAAK,OAAO,EAAE;AAAA,UACzE,CAAC;AAAA,QACH;AAAA,MACF;AAAA,IACF,CAAC;AAED,oBACE,qBAAqB,YAAY,WAAW,iBAAiB,UAAU,gBAAgB;AAAA,EAC3F,OAAO;AACL,oBAAgB;AAAA,EAClB;AAEA,SAAO,SAA6B;AAAA,IAClC,SAAS;AAAA,IACT,SAAS,cAAc,IAAI,CAAC,OAAO;AAAA,MACjC,MAAM,EAAE;AAAA,MACR,OAAO;AAAA,MACP,SAAS,cAAc,UAAU;AAAA;AAAA,IACnC,EAAE;AAAA,IACF,UAAU;AAAA,EACZ,CAAC;AACH;AAEA,eAAe,sBAAsB,KAAa,eAAwC;AACxF,QAAM,WAAW,MAAM,aAAa,GAAG;AAGvC,QAAM,cAAc,MAAM,gBAAgB,KAAK,UAAU,aAAa;AAEtE,MAAI,CAAC,aAAa;AAChB,YAAQ,IAAI,GAAG,QAAQ,+CAA+C,UAAU,EAAE;AAClF,WAAO;AAAA,EACT;AAEA,QAAM,cAAc,MAAM,sBAAsB,WAAW;AAC3D,QAAM,cAAc,gBAAgB,WAAW;AAE/C,MAAI,YAAY,WAAW,GAAG;AAC5B,WAAO;AAAA,EACT;AAEA,MAAI,YAAY,WAAW,GAAG;AAC5B,YAAQ,IAAI,GAAG,QAAQ,iBAAiB,YAAY,CAAC,CAAC,GAAG,UAAU,EAAE;AACrE,WAAO,YAAY,CAAC;AAAA,EACtB;AAEA,SAAO,OAAe;AAAA,IACpB,SAAS;AAAA,IACT,QAAQ,OAAO,SAAS;AACtB,YAAM,WAAW,OACb,YAAY,OAAO,CAAC,MAAM,EAAE,YAAY,EAAE,SAAS,KAAK,YAAY,CAAC,CAAC,IACtE;AACJ,aAAO,SAAS,IAAI,CAAC,MAAM;AACzB,cAAM,MAAM,YAAY,KAAK,CAAC,MAAM,EAAE,SAAS,CAAC;AAChD,eAAO;AAAA,UACL,MAAM;AAAA,UACN,OAAO;AAAA,UACP,aAAa,MAAM,aAAa,IAAI,QAAQ,KAAK;AAAA,QACnD;AAAA,MACF,CAAC;AAAA,IACH;AAAA,EACF,CAAC;AACH;AAEA,eAAe,gBACb,KACA,UACA,cAC6B;AAE7B,QAAM,eAAyB,CAAC;AAEhC,MAAI,cAAc;AAChB,UAAM,UAAUA,MAAK,QAAQ,YAAY;AACzC,QAAI,CAAC,aAAa,SAAS,OAAO,GAAG;AACnC,mBAAa,KAAK,OAAO;AAAA,IAC3B;AAAA,EACF;AAEA,MAAI,CAAC,aAAa,SAAS,GAAG,GAAG;AAC/B,iBAAa,KAAK,GAAG;AAAA,EACvB;AAEA,MAAI,aAAa,OAAO,CAAC,aAAa,SAAS,QAAQ,GAAG;AACxD,iBAAa,KAAK,QAAQ;AAAA,EAC5B;AAEA,aAAW,OAAO,cAAc;AAC9B,eAAW,aAAa,wBAAwB;AAC9C,YAAM,WAAW,GAAG,GAAG,IAAI,SAAS;AACpC,UAAI,MAAM,WAAW,QAAQ,GAAG;AAC9B,eAAO;AAAA,MACT;AAAA,IACF;AAAA,EACF;AAEA,SAAO;AACT;AAEA,eAAe,wBAGZ;AACD,QAAM,YAAY,MAAM,QAAQ;AAAA,IAC9B,SAAS;AAAA,IACT,SAAS;AAAA,EACX,CAAC;AAED,MAAI,CAAC,WAAW;AACd,WAAO,EAAE,SAAS,GAAG,OAAO,MAAM;AAAA,EACpC;AAEA,QAAM,UACH,MAAM,OAAO;AAAA,IACZ,SAAS;AAAA,IACT,SAAS;AAAA,IACT,KAAK;AAAA,IACL,KAAK;AAAA,EACP,CAAC,KAAM;AAET,QAAM,QAAQ,MAAM,QAAQ;AAAA,IAC1B,SAAS;AAAA,IACT,SAAS;AAAA,EACX,CAAC;AAED,SAAO,EAAE,SAAS,MAAM;AAC1B;AAEA,eAAe,uBAAuB,QAA2B,KAA+B;AAC9F,QAAM,cAAc,OAAO,UACxB,IAAI,CAAC,MAAM;AACV,UAAM,MAAM,EAAE,WAAW,GAAG,IAAI,EAAE,MAAM,IAAI,SAAS,CAAC,IAAI;AAC1D,WAAO,KAAK,GAAG;AAAA,EACjB,CAAC,EACA,KAAK,IAAI;AAEZ,UAAQ,IAAI;AAAA,EAAK,SAAS,uBAAuB,UAAU,EAAE;AAC7D,UAAQ,IAAI,GAAG,QAAQ,GAAG,SAAI,OAAO,EAAE,CAAC,GAAG,UAAU,EAAE;AACvD,UAAQ,IAAI,GAAG,UAAU,cAAc,UAAU;AAAA,EAAK,WAAW,EAAE;AACnE,UAAQ,IAAI,GAAG,UAAU,UAAU,UAAU,OAAO,OAAO,MAAM,EAAE;AACnE,UAAQ,IAAI,GAAG,UAAU,WAAW,UAAU,MAAM,OAAO,OAAO,EAAE;AACpE,UAAQ,IAAI,GAAG,UAAU,SAAS,UAAU,QAAQ,OAAO,QAAQ,QAAQ,IAAI,EAAE;AACjF,UAAQ,IAAI,GAAG,QAAQ,GAAG,SAAI,OAAO,EAAE,CAAC,GAAG,UAAU,EAAE;AAEvD,SAAO,QAAQ;AAAA,IACb,SAAS;AAAA,IACT,SAAS;AAAA,EACX,CAAC;AACH;AAEA,eAAe,cACb,QACA,MACe;AACf,QAAM,MAAM,QAAQ,IAAI;AACxB,QAAM,aAAsC;AAAA,IAC1C,QAAQ,OAAO;AAAA,IACf,SAAS,OAAO;AAAA,IAChB,OAAO,OAAO;AAAA,IACd,GAAI,MAAM,kBAAkB,EAAE,QAAQ,KAAK,iBAAiB,QAAQ,KAAK,IAAI,CAAC;AAAA,IAC9E,cAAc;AAAA,IACd,YAAY;AAAA,IACZ,SAAS;AAAA,IACT,gBAAgB;AAAA,IAChB,mBAAmB;AAAA,IACnB,OAAO;AAAA,EACT;AAEA,QAAM,SAAS,MAAM,eAAe;AAAA,IAClC,WAAW,CAAC,GAAG,OAAO,SAAS;AAAA,IAC/B;AAAA,EACF,CAAC;AAKD,MAAI,QAAQ;AACV,UAAM,eAAe;AAAA,MACnB,YAAW,oBAAI,KAAK,GAAE,YAAY;AAAA,MAClC;AAAA,MACA,WAAW,OAAO;AAAA,MAClB,QAAQ,OAAO;AAAA,MACf,SAAS,OAAO;AAAA,MAChB,OAAO,OAAO;AAAA,MACd,WAAWA,MAAK,QAAQ,OAAO,UAAU;AAAA,IAC3C,CAAC;AAAA,EACH;AAGA,MAAI,UAAU,OAAO,sBAAsB,KAAK,QAAQ,MAAM,OAAO;AACnE,UAAM,kBAAkB,QAAQ,OAAO,UAAU;AAAA,EACnD;AACF;AAEA,eAAe,kBAAkB,QAA2B,YAAmC;AAC7F,QAAM,cAAc,MAAM,QAAQ;AAAA,IAChC,SAAS,SAAS,SAAS,mBAAmB,UAAU;AAAA,IACxD,SAAS;AAAA,EACX,CAAC;AAED,MAAI,CAAC,aAAa;AAChB;AAAA,EACF;AAEA,UAAQ,IAAI;AAAA,EAAK,QAAQ,+BAA+B,UAAU;AAAA,CAAI;AAEtE,QAAM,aAAsC;AAAA,IAC1C,QAAQ,OAAO;AAAA,IACf,SAAS,OAAO;AAAA,IAChB,OAAO,OAAO;AAAA,IACd,aAAa;AAAA,IACb,KAAK;AAAA,IACL,cAAc;AAAA,IACd,YAAY;AAAA,IACZ,SAAS;AAAA,IACT,gBAAgB;AAAA,IAChB,mBAAmB;AAAA,IACnB,OAAO;AAAA,EACT;AAEA,QAAM,cAAc,MAAM,eAAe;AAAA,IACvC,WAAW,CAAC,GAAG,OAAO,SAAS;AAAA,IAC/B;AAAA,EACF,CAAC;AAGD,MAAI,eAAe,YAAY,sBAAsB,KAAK,QAAQ,MAAM,OAAO;AAC7E,UAAM,kBAAkB,QAAQ,YAAY,UAAU;AAAA,EACxD;AACF;","names":["path","path"]}

package/dist/templates/.agentv/targets.yaml

@@ -1,97 +0,0 @@
-# A list of all supported evaluation targets for the project.
-# Each target defines a provider and its specific configuration.
-# Actual values for paths/keys are stored in the local .env file.
-# Agent and CLI targets use grader_target to reference an LLM target for scoring.
-
-targets:
-  - id: default
-    provider: azure
-    endpoint: "{{ env.AZURE_OPENAI_ENDPOINT }}"
-    api_key: "{{ env.AZURE_OPENAI_API_KEY }}"
-    model: "{{ env.AZURE_DEPLOYMENT_NAME }}"
-    # version: "{{ env.AZURE_OPENAI_API_VERSION }}"  # Optional: uncomment to override default (v1)
-
-  - id: codex
-    provider: codex
-    grader_target: azure-llm
-    # Uses the Codex CLI (defaults to `codex` on PATH)
-    # executable: "{{ env.CODEX_CLI_PATH }}"        # Optional: override executable path
-    # args:                             # Optional additional CLI arguments
-    #   - --profile
-    #   - "{{ env.CODEX_PROFILE }}"
-    #   - --model
-    #   - "{{ env.CODEX_MODEL }}"
-    #   - --ask-for-approval
-    #   - "{{ env.CODEX_APPROVAL_PRESET }}"
-    cwd: "{{ env.CODEX_WORKSPACE_DIR }}"            # Where scratch workspaces are created
-    log_dir: "{{ env.CODEX_LOG_DIR }}"              # Optional: where Codex CLI stream logs are stored (defaults to ./.agentv/logs/codex)
-    stream_log: raw                    # Optional: 'summary' for consolidated logs or 'raw' for per-event logs
-
-  # Claude - Anthropic's Claude Agent SDK
-  - id: claude
-    provider: claude
-    grader_target: azure-llm
-    # Uses the @anthropic-ai/claude-agent-sdk
-    # model: claude-sonnet-4-20250514          # Optional: override model
-    # cwd: "{{ env.CLAUDE_WORKSPACE_DIR }}"         # Optional: working directory (defaults to process.cwd())
-    # max_turns: 50                            # Optional: max conversation turns
-    # max_budget_usd: 5.0                      # Optional: max cost budget in USD
-    # log_dir: "{{ env.CLAUDE_LOG_DIR }}"           # Optional: where stream logs are stored (defaults to ./.agentv/logs/claude)
-    stream_log: raw                           # Optional: 'summary' for consolidated logs or 'raw' for per-event logs
-    # system_prompt: optional override (default instructs agent to include code in response)
-
-  - id: azure-llm
-    provider: azure
-    endpoint: "{{ env.AZURE_OPENAI_ENDPOINT }}"
-    api_key: "{{ env.AZURE_OPENAI_API_KEY }}"
-    model: "{{ env.AZURE_DEPLOYMENT_NAME }}"
-    version: "{{ env.AZURE_OPENAI_API_VERSION }}"
-
-  - id: gemini-llm
-    provider: gemini
-    api_key: "{{ env.GOOGLE_GENERATIVE_AI_API_KEY }}"
-    model: "{{ env.GEMINI_MODEL_NAME }}"
-
-  - id: local_cli
-    provider: cli
-    grader_target: azure-llm
-    # Passes the fully rendered prompt and any attached files to a local Python script
-    # NOTE: Do not add quotes around {PROMPT} or {FILES} - they are already shell-escaped
-    command: uv run ./mock_cli.py --prompt {PROMPT} {FILES} --output {OUTPUT_FILE}
-    # Format for each file in {FILES}. {path} and {basename} are automatically shell-escaped, so no quotes needed
-    files_format: --file {path}
-    # Optional working directory resolved from .env
-    cwd: "{{ env.CLI_EVALS_DIR }}"
-    healthcheck:
-      command: uv run ./mock_cli.py --healthcheck
-
-  # -- MiMo (Xiaomi) via OpenRouter -----------------------------------
-  # All MiMo models are available through OpenRouter with OpenAI-compatible API.
-  # See https://openrouter.ai/xiaomi/mimo-v2.5-pro for pricing and limits.
-  #
-  # Models:
-  #   mimo-v2.5-pro    - 1M context, 131K output, flagship
-  #   mimo-v2-pro      - 1M context, ~131K output
-  #   mimo-v2.5        - 1M context, ~131K output, multimodal
-  #   mimo-v2-flash    - 262K context, 65K output, fast MoE (open-source)
-  #   mimo-v2-omni     - 262K context, 65K output, omni-modal
-  - id: mimo
-    provider: openrouter
-    api_key: "{{ env.OPENROUTER_API_KEY }}"
-    model: xiaomi/mimo-v2.5-pro
-
-  - id: mimo-flash
-    provider: openrouter
-    api_key: "{{ env.OPENROUTER_API_KEY }}"
-    model: xiaomi/mimo-v2-flash
-
-  # -- Direct provider (not through OpenRouter) -----------------------
-  # For providers not in pi-ai's model registry, set max_output_tokens
-  # to match your model's actual output limit. Without this, the default
-  # is 16K which may cap output below the model's capability.
-  # - name: mimo-direct
-  #   provider: openai
-  #   base_url: https://token-plan-sgp.xiaomimimo.com/v1
-  #   api_key: "{{ env.XIAOMI_MIMO_API_KEY }}"
-  #   model: xiaomi/mimo-v2.5-pro
-  #   max_output_tokens: 131072