feat: programmatic API with getSuites(), executeSuite(), executeTest()

[DavertMik]

Summary

• Add getSuites(pattern?) to Codecept class — returns parsed suites with tests as plain objects, using a temporary Mocha instance to avoid polluting container state
• Add executeSuite(suite) and executeTest(test) — clean execution methods that accept objects returned by getSuites()
• Refactor workers.js — createGroupsOfTests() and createGroupsOfSuites() now use getSuites() instead of duplicated Mocha boilerplate (~30 lines removed)
• Refactor dryRun.js — printTests() now uses getSuites(), removing direct Container.mocha() dependency
• Export Result class from lib/index.js
• Rewrite "Custom Runner" section in docs/internal-api.md as full "Programmatic API" reference

Usage

import { Codecept, container } from 'codeceptjs'

const codecept = new Codecept(config, { steps: true })
await codecept.init(__dirname)

// List tests
const suites = codecept.getSuites()
for (const suite of suites) {
  console.log(suite.title)
  for (const test of suite.tests) {
    console.log(' -', test.title, test.tags)
  }
}

// Execute
await codecept.bootstrap()
for (const suite of suites) {
  await codecept.executeSuite(suite)
}
const result = container.result()
await codecept.teardown()

Test plan

• dry-run command works with refactored getSuites() (tested with grep filtering)
• Unit tests pass (401 passing, 7 pre-existing AI test failures unrelated)
• Shard tests pass
• Workers should be tested with run-workers 2

🤖 Generated with Claude Code