cli refactoring and merge into app codebase

.github/workflows/cli-deploy.yml

@@ -43,9 +43,6 @@ jobs:
             zip_name: macos-arm64.zip
             smoke_command: ./out/macos-arm64/anywaydata --help
             bun_build_flags: --compile
-    defaults:
-      run:
-        working-directory: ./cli
     steps:
       - uses: actions/checkout@v4
       - uses: oven-sh/setup-bun@v2
@@ -56,7 +53,7 @@ jobs:
       - name: build binary
         shell: bash
         run: |
-          bun build ./index.ts ${{ matrix.bun_build_flags }} --target=${{ matrix.bun_target }} --outfile "${{ github.workspace }}/${{ matrix.bin_relpath }}"
+          bun build ./apps/cli/src/bun-entry.js ${{ matrix.bun_build_flags }} --target=${{ matrix.bun_target }} --outfile "${{ github.workspace }}/${{ matrix.bin_relpath }}"
       - name: smoke test binary
         run: ${{ matrix.smoke_command }}
       - name: package zip (windows)

.github/workflows/codex-pr-review.yml

@@ -0,0 +1,39 @@
+name: Codex PR Review
+
+on:
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  pull-requests: write
+  issues: write
+
+concurrency:
+  group: codex-pr-review-${{ github.workflow }}-${{ github.ref_name }}
+  cancel-in-progress: true
+
+jobs:
+  codex-review:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Verify OpenAI key is configured
+        shell: bash
+        run: |
+          if [ -z "${{ secrets.OPENAI_API_KEY }}" ]; then
+            echo "OPENAI_API_KEY is not configured for this run."
+            exit 1
+          fi
+
+      - name: Run Codex PR review
+        uses: openai/codex-action@v1
+        with:
+          openai-api-key: ${{ secrets.OPENAI_API_KEY }}
+          prompt: |
+            Review this pull request for:
+            - bugs and behavioral regressions
+            - security or reliability risks
+            - missing or weak tests
+            Keep feedback concise and actionable with file/line references.

.github/workflows/node.js.yml

@@ -74,9 +74,31 @@ jobs:
       - name: Install dependencies
         run: npm ci
 
+      - name: CLI pack smoke test
+        shell: bash
+        run: |
+          set -euo pipefail
+          rm -rf .tmp-pack
+          mkdir -p .tmp-pack
+          CORE_TARBALL="$(npm pack --workspace packages/core --pack-destination ./.tmp-pack | tail -n 1)"
+          CLI_TARBALL="$(npm pack --workspace apps/cli --pack-destination ./.tmp-pack | tail -n 1)"
+          npm exec --yes \
+            --package "./.tmp-pack/${CORE_TARBALL}" \
+            --package "./.tmp-pack/${CLI_TARBALL}" \
+            anywaydata -- --help >/dev/null
+          echo "CLI npm pack smoke test passed with local core+cli tarballs: ${CORE_TARBALL}, ${CLI_TARBALL}"
+
       - name: Build
         run: npm run build --if-present
 
+      - name: Set up Bun
+        uses: oven-sh/setup-bun@v2
+        with:
+          bun-version: 1.2.10
+
+      - name: Bun CLI entry smoke test
+        run: bun run apps/cli/src/bun-entry.js --help
+
       - name: Install Playwright browser
         run: npx playwright install --with-deps chromium
 

.gitignore

@@ -10,3 +10,4 @@ coverage.txt
 test_output.txt
 playwright-report
 test-results
+.tmp

README.md

@@ -176,7 +176,7 @@ This repository now uses npm workspaces:
 
 - `packages/core` -> shared generation engine (`@anywaydata/core`)
 - `packages/core-ui` -> browser UI modules (`@anywaydata/core-ui`)
-- `packages/cli` -> CLI wrapper (`@anywaydata/cli`)
+- `apps/cli` -> npm CLI package (`@anywaydata/cli`)
 - `apps/api` -> REST API (`@anywaydata/api`)
 - `apps/mcp` -> MCP server (`@anywaydata/mcp`)
 
@@ -201,7 +201,7 @@ Run an individual workspace command:
 
 ## npm CLI Usage (`@anywaydata/cli`)
 
-The npm CLI package is the workspace package `@anywaydata/cli` (separate from the Bun CLI in `cli/`).
+The npm CLI package is the workspace app `apps/cli` (`@anywaydata/cli`), and Bun executable builds also use `apps/cli` as the source.
 
 Install globally from npm:
 

apps/cli/README.md

@@ -0,0 +1,162 @@
+# AnyWayData CLI (`apps/cli`)
+
+Unified CLI implementation for Node and Bun, backed by `@anywaydata/core`.
+
+This workspace is the source of truth for:
+
+- npm/npx CLI usage (`anywaydata`)
+- Node runtime entrypoint (`src/node-entry.js`)
+- Bun runtime entrypoint (`src/bun-entry.js`)
+- Bun executable builds
+
+## Quick Start
+
+Install globally:
+
+```bash
+npm install -g @anywaydata/cli
+```
+
+Run with npx:
+
+```bash
+npx @anywaydata/cli --help
+```
+
+Run from repo with Node:
+
+```bash
+node apps/cli/src/node-entry.js --help
+```
+
+Run from repo with Bun:
+
+```bash
+bun run apps/cli/src/bun-entry.js --help
+```
+
+## Example Usage
+
+Generate 3 rows from a spec file:
+
+```bash
+anywaydata generate -i exampleTestDataSpec.txt -n 3
+```
+
+Write to an output file:
+
+```bash
+anywaydata generate -i exampleTestDataSpec.txt -n 3 -o output.csv
+```
+
+Redirect stdout:
+
+```bash
+anywaydata generate -i exampleTestDataSpec.txt -n 3 > output.csv
+```
+
+Use included examples in this workspace:
+
+```bash
+anywaydata generate -i ./apps/cli/examples/company-literal.txt -n 3 -f csv
+anywaydata generate -i ./apps/cli/examples/company.txt -n 3 -f csv
+anywaydata generate -i ./apps/cli/examples/regex-field.txt -n 3 -f csv
+```
+
+## CLI Options
+
+- `-i, --inputfile` input text spec path (required)
+- `-n, --numberOfLines` row count (required, overridden to `1` by `--testMode`)
+- `-f, --format` output format (e.g. `csv`, `json`, `jsonl`, `xml`, `sql`)
+- `-o, --outputfile` optional output file path
+- `-t, --testMode` generate one row and print diagnostics/example output
+- `--show-progress` force progress logs on/off
+- `--stream` enable stream mode when supported
+- `--stream-threshold` auto-enable stream mode when rows >= threshold and output file is set (default `5000`)
+- `--unsafe-faker-expressions` allow expression-style faker args (unsafe for untrusted input)
+
+## Streaming Behavior
+
+Streaming is currently supported for:
+
+- `csv`
+- `jsonl`
+
+Other formats use buffered generation.
+
+Examples:
+
+```bash
+anywaydata generate -i ./apps/cli/examples/company-literal.txt -n 100000 -f jsonl -o output.jsonl --stream
+anywaydata generate -i ./apps/cli/examples/company-literal.txt -n 100000 -f csv -o output.csv --stream-threshold 1000
+```
+
+## Spec File Format
+
+The test data spec format is two lines per field:
+
+1. Column name
+2. Rule/spec for that column
+
+Example:
+
+```text
+Column Name 1
+Data Value Spec for Column Name 1
+Column Name 2
+Data Value Spec for Column Name 2
+```
+
+### Literal Value Example
+
+```text
+Company
+AnyWayData
+```
+
+### Faker Value Example
+
+```text
+Company
+company.name
+```
+
+Faker API docs:
+
+- https://fakerjs.dev/api/
+
+### Regex Value Example
+
+```text
+Regex Generated Field
+[A-Z]{3,6}[0-9]{0,6}
+```
+
+## Safe vs Unsafe Faker Expressions
+
+CLI is safe-by-default. For untrusted specs, keep default behavior.
+
+To allow expression-style faker arguments, opt in explicitly:
+
+```bash
+anywaydata generate -i input.txt -n 10 -f csv --unsafe-faker-expressions
+```
+
+## Build Bun Executables
+
+Build a Bun executable from the shared entrypoint:
+
+```bash
+bun build ./apps/cli/src/bun-entry.js --compile --target=bun-linux-x64 --outfile ./out/linux-x64/anywaydata
+```
+
+Cross-platform targets follow the same pattern:
+
+- `bun-windows-x64`
+- `bun-linux-x64`
+- `bun-linux-arm64`
+- `bun-darwin-arm64`
+
+See Bun executable docs:
+
+- https://bun.sh/docs/bundler/executables

apps/cli/jest.config.cjs

@@ -0,0 +1,8 @@
+module.exports = {
+  verbose: true,
+  testEnvironment: 'node',
+  transform: {
+    '^.+\\.js$': 'babel-jest',
+  },
+  testMatch: ['<rootDir>/src/tests/**/*.test.js'],
+};

packages/cli/package.json → apps/cli/package.json

@@ -1,10 +1,10 @@
 {
   "name": "@anywaydata/cli",
-  "version": "2.1.0",
+  "version": "0.1.0",
   "type": "module",
   "private": false,
   "bin": {
-    "anywaydata": "src/index.js"
+    "anywaydata": "./src/node-entry.js"
   },
   "main": "./src/index.js",
   "exports": {
@@ -17,8 +17,8 @@
     "access": "public"
   },
   "scripts": {
-    "build": "node -e \"console.log('cli: no build step')\"",
-    "test": "node ./src/index.js --help"
+    "build": "node -e \"console.log('cli-next: no build step')\"",
+    "test": "node --experimental-vm-modules ../../node_modules/jest/bin/jest.js --config ./jest.config.cjs"
   },
   "dependencies": {
     "@anywaydata/core": "^1.1.0",

apps/cli/src/bun-entry.js

@@ -0,0 +1,7 @@
+#!/usr/bin/env bun
+import { createBunPlatform, runCli } from './index.js';
+
+const code = await runCli(process.argv, createBunPlatform());
+if (code !== 0) {
+  process.exit(code);
+}

apps/cli/src/cli-options.js

@@ -0,0 +1,70 @@
+import yargs from 'yargs';
+import { hideBin } from 'yargs/helpers';
+import { SUPPORTED_FORMATS } from '@anywaydata/core';
+
+export function parseCliOptions(argvInput = process.argv) {
+  const parsed = yargs(hideBin(argvInput))
+    .usage('Usage: anywaydata generate -i input.txt -n 10 -f csv [-o output.txt]')
+    .command('generate', 'Generate test data from an input spec')
+    .option('i', { alias: 'inputfile', type: 'string', demandOption: true, describe: 'Text file with a data spec' })
+    .option('n', {
+      alias: 'numberOfLines',
+      type: 'number',
+      demandOption: true,
+      default: 1,
+      describe: 'Number of lines to generate',
+    })
+    .option('f', {
+      alias: 'format',
+      type: 'string',
+      default: 'csv',
+      describe: `Output format e.g. ${SUPPORTED_FORMATS.join(',')}`,
+    })
+    .option('o', { alias: 'outputfile', type: 'string', describe: 'Output file' })
+    .option('t', {
+      alias: 'testMode',
+      type: 'boolean',
+      default: false,
+      describe: 'Generate one line, report diagnostics and include example output',
+    })
+    .option('show-progress', {
+      type: 'boolean',
+      default: undefined,
+      describe: 'Show progress logs even when writing to file',
+    })
+    .option('stream', {
+      type: 'boolean',
+      default: false,
+      describe: 'Use streaming generation path when available (exporting CSV or JSONL)',
+    })
+    .option('stream-threshold', {
+      type: 'number',
+      default: 5000,
+      describe: 'Auto-enable stream mode when rows >= threshold and output file is provided',
+    })
+    .option('unsafe-faker-expressions', {
+      type: 'boolean',
+      default: false,
+      describe: 'Allow expression-style faker arguments (unsafe for untrusted input)',
+    })
+    .help('h')
+    .alias('h', 'help')
+    .parseSync();
+
+  const rowCount = parsed.testMode ? 1 : parsed.numberOfLines;
+  const streamThreshold = Number.isFinite(parsed.streamThreshold) ? Math.max(0, parsed.streamThreshold) : 5000;
+  const shouldStream = parsed.stream === true || (!!parsed.outputfile && rowCount >= streamThreshold);
+  const showProgress =
+    parsed.showProgress !== undefined ? parsed.showProgress === true : parsed.testMode || !parsed.outputfile;
+
+  return {
+    inputFile: parsed.inputfile,
+    outputFile: parsed.outputfile,
+    format: String(parsed.format || 'csv').toLowerCase(),
+    rowCount,
+    testMode: parsed.testMode === true,
+    showProgress,
+    shouldStream,
+    unsafeFakerExpressions: parsed['unsafe-faker-expressions'] === true,
+  };
+}