Batch Requests

Every network client exposes a generic batch() method, and most mixins layer a purpose-built batched variant on top of it (getProfiles, getClans, getLeaderboards, getProfileStatsBatch).

Basic usage

const profiles = await pika.getProfiles(['divena', 'Notch', 'jeb_']);

for (const result of profiles) {
  if (result.status === 'fulfilled') {
    console.log(result.value);
  } else {
    console.error(result.reason);
  }
}

The contract

• Order is preserved. profiles[i] corresponds to usernames[i], regardless of which operation finishes first.
• Failures are isolated. One failing request doesn't cancel the others. Each settles independently into { status: 'fulfilled', value } or { status: 'rejected', reason }, unless you pass throwOnError: true.
• Concurrency is bounded. Defaults to 10 simultaneous in-flight operations (batchConcurrency on the client, overridable per call via options.concurrency).
• Still rate-limited. Every operation in a batch goes through the exact same rate limiter as a single call. Batching doesn't bypass it; it parallelizes within the limiter's constraints.

Options

await pika.getProfiles(usernames, undefined, {
  concurrency: 5,
  throwOnError: true,
  onProgress: (completed, total) => console.log(`${completed}/${total}`),
});

Calling batch() directly

For anything not covered by a purpose-built batched method:

const results = await pika.batch([
  () => pika.getProfile('divena'),
  () => pika.getClan('SomeClan'),
  () => pika.getLeaderboard({ gamemode: 'bedwars', stat: 'wins' }),
]);

Each operation's return type is inferred individually, so a mixed-type batch like the one above stays fully typed instead of widening to a shared union.

runBatch (src/client/batch.ts) is the shared implementation underneath every batch() call. It isn't part of the public API surface, but it's worth a read if you want to see exactly how concurrency and failure isolation are implemented.