Skip to content

About

Julia solution example for SISAP 2026 Indexing Challenge

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

SISAP 2026 Challenge: Working example on Julia

This repository is a working example for the SISAP 2025 Indexing Challenge https://sisap-challenges.github.io/, working with Julia and GitHub Actions, as specified in Task's descriptions. It is based on the previous year example.

Steps for running

It requires a working installation of Julia (better works with v1.10.10. It also wroks with v1.11 and v1.12 series but you could find performance regressions for our similarity search package. You can download Julia from https://julialang.org/downloads/. You also need git tools and internet access for cloning, install package dependencies and downloading datasets.

The steps are the following:

  1. Clone the example repository
  2. Fetch the datasets
  3. Instantiate the project
  4. Run
  5. Evaluation

the fullset of instructions are listed in the GitHub Actions workflow

https://github.com/sisap-challenges/sisap2026-julia-example/blob/main/.github/workflows/ci.yml

Note that you will need to adjust your scripts to hold the correct hyperparameters for any benchmark you use

Clone this repository

git clone https://github.com/sisap-challenges/sisap2026-julia-example
cd sisap2026-julia-example

Fetch the datasets

You can clone our dataset repository or fetch dataset by dataset; run the prepare-data.sh script to download the necessary testing data. Note that we expect to use small dataset for your GitHub Actions.

https://huggingface.co/datasets/SISAP-Challenges/SISAP2026/tree/main

Instantiate the project

Julia requires to prepare the working directory through an instantiation, a.k.a. installing dependencies, as follows:

JULIA_PROJECT=. JULIA_NUM_THREADS=8 julia -e 'using Pkg; Pkg.instantiate()'

You need internet access for this step.

Run

A similar procedure is needed to run; note that Julia may compile many packages in the first run, so please be patient.

JULIA_PROJECT=. JULIA_NUM_THREADS=8 julia -L sisap2026.jl -e 'main_task1(); main_task2()'

You should modify the number of threads to adapt your hardware; you can also try to add optimization flags to julia, e.g., julia -O3 -Cnative.

Evaluation

JULIA_PROJECT=. julia -L eval.jl -e 'eval_task1()'
JULIA_PROJECT=. julia -L eval.jl -e 'eval_task2()'

Two result files will be created: result-task1.csv and result-task2.csv.

How to take this to create my own system

You can fork this repository and polish it to create your solution. Please also take care of the ci workflow (see below).

GitHub Actions: Continuous integration

You can monitor your runnings in the "Actions" tab of the GitHub panel: for instance, you can see some runs of this repository: https://github.com/sisap-challenges/sisap2026-julia-example/actions

About

Julia solution example for SISAP 2026 Indexing Challenge

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages