Skip to main content

kaprese

Collection of APR tools by SAL (Software Analysis Lab) at Korea University.

Quick Start

You can run kaprese by following the steps below.

Prerequisites

kaprese requires the following software to be installed:

  • Python 3.12 or higher (required)
  • Docker 24 or higher (recommended, we did not test it on other versions)

Installation

kaprese can be installed via pip:

pip install kaprese

Usage

You can run saver engine on flint-1 benchmark as follows:

kaprese benchmark preset c
kaprese engine preset saver
kaprese run -b flint-1 spearmint-1 -e saver
# See the logs. If you want to see more logs set LOG_LEVEL environment variable to INFO, e.g.,
# LOG_LEVEL=info kaprese run -b flint-1 spearmint-1 -e saver

Options

kaprese provides the following options:

Benchmarks Repository
benchmark command to do something about benchmark
engine command to do something about engine
run command to run APR tools on benchmarks
eval command to evaluate the results
-h, --help command to show help

Benchmarks and Engines

kaprese provides a set of benchmarks and engines. You can register them by calling kaprese benchmark preset <language> and kaprese engine preset <engine>. The following table shows the list of supported benchmarks and engines.

Benchmarks

Benchmarks Repository
c C benchmarks in kupl/starlab-benchmarks
ocaml OCaml benchmarks in kupl/starlab-benchmarks

Engines

Engines Repository
saver kupl/SAVER_public
cafe kupl/LearnML

How to Add Your Own

All benchmarks and engines are registered to the global registry (by default, registry is located in ~/.kaprese). You can add your own benchmarks and engines by registering them with a small python code.

Add a Benchmark

A benchmark is an instance of kaprese.core.benchmark.Benchmark. You can initialize a benchmark with the following 4 parameters:

  • name: name of the benchmark (must be unique).
  • image: docker image name, e.g., ghcr.io/kupl/starlab-benchmarks/c:flint-1.
  • language_command or _language: command to find the language of the benchmark, if you set _language, language_command is ignored. the command will be run inside the docker container from image.
  • workdir_command or _workdir: command to find the working directory of the benchmark, if you set _workdir, workdir_command is ignored. the command will be run inside the docker container from image. Then by calling register method of Benchmark, you can register the benchmark to the global registry.

For example, the following code defines a benchmark named flint-1:

from kaprese.core.benchmark import Benchmark

flint_1 = Benchmark(
    name="flint-1",
    image="ghcr.io/kupl/starlab-benchmarks/c:flint-1",
    language_command="cat metadata.json | jq -r .language", # c
    workdir_command="cd $(cat metadata.json | jq -r .buggyPath) && pwd", # /workspace/buggy
)
flint_1.register()

Add an Engine

An engine is an instance of kaprese.core.engine.Engine. You can initialize an engine with the following 6 parameters:

  • name: name of the engine (must be unique).
  • supported_languages: list of supported languages, e.g., ["c", "java"].
  • supported_os: list of supported operating systems, e.g., ["ubuntu:20.04", "debian:12"].
  • location: location to the docker context directory of the engine, e.g., ~/saver.
  • build_args: arguments to pass to docker build command to fill ARG variables in the Dockerfile, e.g., {"SOME_ARG": "some_value"}.
  • exec_commands: list of commands to execute the engine, e.g., ["saver ...", "cp ..."], all commands are joined with ;, in other words, all commands are executed and the return code of the last command is returned. Note that, your dockerfile MUST gets the benchmark image as an argument and starts build from it, e.g.,
ARG BENCHMARK_IMAGE
FROM ${BENCHMARK_IMAGE}
...

In the same way, you MUST specify the benchmark image as an argument when you build the engine image, e.g.,

from kaprese.core.engine import Engine

saver = Engine(
    "saver",
    supported_languages=["c"],
    supported_os=["ubuntu:20.04"],
    location="https://github.com/kupl/kaprese-engines#main:context/saver/starlab-benchmarks", # location of saver context for preset benchmarks
    build_args={"BENCHMARK_IMAGE": "ghcr.io/kupl/starlab-benchmarks/c:flint-1"},
    exec_commands= ["saver ...", "cp ..."], # see kaprese/engines/saver.py
)

Then by calling register method of Engine, you can register the engine to the global registry.

saver.register()

How to Contribute

Any contributions are welcome! You can leave an issue or make a pull request.

Metadata

Release files for kaprese 1.0.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for kaprese 1.0.0
File Size Uploaded Provenance Actions
kaprese-1.0.0.tar.gz 18.4 kB With provenance Details for kaprese-1.0.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for kaprese 1.0.0
File Interpreter ABI Platform Actions
kaprese-1.0.0-py3-none-any.whl Python 3 none any Details for kaprese-1.0.0-py3-none-any.whl

Total release size: 43.7 kB

Release files / kaprese-1.0.0.tar.gz

Download URL kaprese-1.0.0.tar.gz
Size 18.4 kB
Tags Source
SHA-256 checksum
How to use checksums
035b63b7f6194422fa1c044305cb45699d168ce2057727d208e40a65fd3f6372
BLAKE2b-256 checksum
How to use checksums
46e7a684659deea250eba6e3b28a7c34068493a165057beb28e8d2f7ae697c60
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2025.

Transparency log

Release files / kaprese-1.0.0-py3-none-any.whl

Download URL kaprese-1.0.0-py3-none-any.whl
Size 25.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
cf21f7cb6196d48e3b55d079b4f99f3f6c6fbe9d0504802609f399fd75ece60c
BLAKE2b-256 checksum
How to use checksums
b95a86f5e4864fc323f78327a6e89f53f0330cc38af8ba554720e83ce9476184
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2025.

Transparency log

Security for kaprese 1.0.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for kaprese-1.0.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2025.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

1.0.0 This release

2 release files Full provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page