stopcock

import { pipe, A } from '@stopcock/fp'

const topActive = pipe(
  players,
  A.filter(p => p.active && p.score > 0),
  A.map(p => ({ name: p.name, score: p.score })),
  A.take(10),
)

pipe spots that filter, map, and take can all run in one loop, so it fuses them. take(10) bails out after 10 hits. On a million-element array, that means touching a few dozen items instead of three full passes.

9xfilter→map→take(10) on 100K elements

---

import { pipe, O, R } from '@stopcock/fp'

const port = pipe(
  O.fromNullable(process.env.PORT),
  O.map(s => parseInt(s, 10)),
  O.filter(n => n > 0 && n < 65536),
  O.getOrElse(() => 3000),
)

const parsed = pipe(
  R.tryCatch(() => JSON.parse(raw)),
  R.map(obj => obj.data),
  R.getOrElse(() => []),
)

Option for values that might not exist, Result for things that can fail. Both use numeric _tag instead of strings, so V8 can branch on them fast.

---

A Array S String D Dict N Number G Guards Obj Object M Math B Boolean Logic Logic Lens Lenses O Option R Result

---

Libraries

Standalone packages. They all work with pipe and with each other.

async Tasks & concurrency stream Lazy sequences validate Schema validation date Date utilities compress Compression algorithms encoding Base64, Hex, MsgPack hash SHA-256, MD5, CRC32 struct Data structures search Fuzzy, Trie, Bloom parse Parser combinators zip ZIP archives serve HTTP framework optics Lens, Prism, Iso geo Geospatial math rand RNG & distributions img Image processing

bun add @stopcock/fp