Parsers

The library provides four parsers to choose from based on input type. All parsers are zero-dependency and browser-compatible.

StringParser — Parse from String

Handles non-standard formatting (trailing commas, raw newlines in strings):

import { StringParser } from 'adofai'

const parser = new StringParser()
const data = parser.parse(`{ "angleData": [...], "settings": {...}, "actions": [...] }`)

BufferParser — Parse from Byte Stream

Directly parses Uint8Array binary streams, automatically handles BOM:

import { BufferParser } from 'adofai'

const parser = new BufferParser()
const u8 = new Uint8Array(await file.arrayBuffer())
const data = parser.parse(u8)

ArrayBufferParser

Accepts ArrayBuffer or string, handles BOM stripping, trailing comma normalization, and UTF-8 decoding:

import { ArrayBufferParser } from 'adofai'

const parser = new ArrayBufferParser()
const buffer = await response.arrayBuffer()
const data = parser.parse(buffer)

LargeFileParser — Extra Large Levels

Memory-optimized incremental parser, suitable for levels with massive angleData / actions:

import { LargeFileParser } from 'adofai'

const parser = new LargeFileParser((stage, percent) => {
  console.log(`[${stage}] ${percent}%`)
}, {
  skipLargeActions: false, // Skip actions when exceeding 100MB
  maxActions: 0            // Limit the number of parsed actions
})

const result = parser.parse(arrayBuffer)
// result: { settings?, angleData?, pathData?, actions?, decorations? }

Key behaviors:

• < 50MB — Normal full parse
• > 50MB — actions are parsed incrementally, one by one
• > 100MB — Can completely skip actions via skipLargeActions: true
• Any size — angleData is always parsed incrementally, number by number