BruJu

Former TODO that will not be executed because they are not relevant anymore :

• Explore the string problems

  • Are UTF8 and UTF16 that different ?
  • Can we disable text encoder (probably not)
  • Can we build a Term type that relies on JsStrings ? (and are they efficient in the wasm bindgen context ?)

• The macro is horrible, we could build :

  • A base struct WasmWrapper that uses a SophiaDataset and builds the RDF JS functions

• A trait WasmWrapperRewritten that uses a base WasmWrapper and recalls its function

BruJu

// At the moment, this code feels bad to write and seems unnecessary
// but it may not be that bad so I keep it here

// https://github.com/BruJu/Portable-Reasoning-in-Web-Assembly
// https://github.com/BruJu/WasmTreeDataset

impl TreedDataset {
    pub fn new_with_same_indexes(&self) -> Self {
        let base_index = vec!(self.base_tree.0.clone());
        let optional_indexes = {

BruJu

Benchmarking RDF JS Stores with Comunica SPARQL Engine using the BSBM benchmark

In this gist, we try to benchmark my web assembly implementation of RDF.JS Stores on SPARQL requests.

The tools we use are :

• Communica's SPARQL Engine which is used as an adapter of the RDF JS Source interface to a SPARQL interface.
• BSBM Tools to build a dataset and benchmark the performances with their test driver

To benchmark RDF JS stores (N3 and my own implementation), we build a SPARQL end point with a code that is mostly a copy / paste from Comunica (see the sparqlendpointfromstore.js file). We use the BSBM Java test driver to benchmark.

BruJu

// https://www.codingame.com/ide/puzzle/detective-pikaptcha-ep2

#include <iostream>
#include <vector>
#include <optional>

enum class Direction {
    TOP, BOT, LEFT, RIGHT
};

BruJu

In the gist https://gist.github.com/BruJu/59c0a9d51d574e9e5d2dad0a74fd11e9 I benchmarked some rdfjs sources against BSBM

In this gist, we badly benchmark exportation in web assembly ( https://github.com/oxigraph/oxigraph/tree/master/js ) using the sparql interface

Same dataset than the other gist, same command, a new script ( oxipoint.js )

$ npm install oxigraph
$ node oxipoint.js

BruJu

// Rust BTreeSet::Range::size_hint testing

use std::collections::btree_set::*;

fn main() {
    let mut treeset = BTreeSet::<u32>::new();
    
    for i in 0..78 {
        if i % 3 != 1 {
            treeset.insert(i);

BruJu

SPARQL End point benchmark, episode 3

Former experiments

• Sparql End Point.js
• Oxipoint.js

This one

In this gist, we benchmark Wasm Tree Store on SPARQL request using the Comunica engine. We compare it to the default parameters of Comunica, which resorts to a n3 store.

BruJu

///////////////////////////////////////////////////
///////////////////////////////////////////////////
///////////////////////////////////////////////////

// Current status : It works
// SophiaExportQuad : https://github.com/BruJu/Portable-Reasoning-in-Web-Assembly/blob/bf9aefade845c9a9b15e8ba6196d33b42ea5701d/sophia-wasm/src/datamodel_quad.rs#L78

///////////////////////////////////////////////////
///////////////////////////////////////////////////

BruJu

// Not yet tested because WeakRef isn't supported on my browsers

// Actually that's bad because Destructors will exist :
// https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/FinalizationRegistry

class Destructor {
  constructor(object, destructor, time) {
    this.object = new WeakRef(object);
    this.destructor = destructor;
    

BruJu

import json
import numpy

with open('forEachEval.json') as json_file:
    data = json.load(json_file)

def avg(l):
	s = sum(l)
	return s / len(l)