saveAsJSON() method not found

The README says that you can call .saveAsJSON() on any bloom filter to get a serialization of the bloom filter state that can be persisted or sent over the network. When I try to use it, however, my Typescript code fails to compile with the error: error TS2339: Property 'saveAsJSON' does not exist on type 'BloomFilter'.

My code looks like:

import { BloomFilter } from 'bloom-filters'

// [...]
const bloomFilterEntries = new Set<string>()
// [...]

const bloomFilter = BloomFilter.from(bloomFilterEntries, 0.01)
const bloomJSON = bloomFilter.saveAsJSON()

[folkvir]

@Callidon @stbrody Unfortunately with our templating methods and decorators we cannot make a small change without redefining fromJSON in all Filters of the project. So in the last commit I defined the return type of fromJSON to any this way you can have type definitions only in two ways.

• const b: BloomFilter = BloomFilter.fromJSON(bloomJSON)
• const b = BloomFilter.fromJSON(bloomJSON) as BloomFilter

This is the only solution for us.

Working example:

import { BloomFilter } from '../../bloom-filters/dist/api'

// [...]
const bloomFilterEntries = new Set<string>(["test"])
// [...]

const bloomFilter: BloomFilter = BloomFilter.from(bloomFilterEntries, 0.01)
const bloomJSON: Object = bloomFilter.saveAsJSON()
console.log('JSON: ', bloomJSON)
const b: BloomFilter = BloomFilter.fromJSON(bloomJSON)
console.log('B: ', b)
console.log(b.has('test'))

PS: Just a reminder if you are not aware of it, don’t even try to test or add data on a Filter that are initialized with an empty set or array. It will cause RangeError: Maximum call stack size exceeded. T.from functions are used to initialize filters with predefined values. Otherwise use the classic constructor or the T.create function to initialize a filter of fixed size with desired error rate.

[Callidon]

The method exists for all classes in the package, but I think the latest rework might have removed the method from the definition files, which causes your TypeScript compiler to fail.

I can’t promise that we can provide a fix quickly, because we now have very little time to work on open source projects (sadly). But I will definitively look into it, because that 's a rather nasty bug !