Datasetpublic

OpenMOSS-Team/SWE-bench-Science

SWE-bench Science SWE-bench Science evaluates coding agents on software-engineering tasks drawn from scientific-computing repositories. The release contains 119 tasks across 20 scientific domains, with isolated environments and separate programmatic verifiers. GitHub release repository: OpenMOSS/SWE-bench-Science Runtime images: Docker Hub, pinned by immutable linux/amd64 digests Evaluation framework: Pier, compatible with Harbor task format Dataset Summary… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/SWE-bench-Science.

sourceHugging Facemitupdated 29d agoView on Hugging Face

7likes3.1kdownloads

rootscriptspier_adapters.py

pier_adapters.py20 linesDownload Raw Back to scripts

1#!/usr/bin/env python32"""Small runtime-only Pier adapters shared by all task images."""3 4from __future__ import annotations5 6from pier.agents.installed.codex import Codex7 8 9class ScienceBenchCodex(Codex):10    """Use npm's optional platform package when Pier installs Codex."""11 12    def install_spec(self):13        spec = super().install_spec()14        for step in spec.steps:15            step.run = step.run.replace(16                "npm install -g @openai/codex",17                "npm install -g --include=optional @openai/codex",18            )19        return spec20