publish SWE-bench Science release (part 2)

tasks/task_041/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-041"
+description = "Repair an inconsistent near-Earth vector conversion"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "041"
+language = "python"
+base_commit_hash = "470d13491dd2c2a9440b3c657af33373ca8d47c1"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-041:v0.1.0@sha256:2d7000ea85b0ce1a1df141d46ccce61e3edcd53e86e9d5136a1b26096799aa77"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_041"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-041:v0.1.0@sha256:7e3b7b1a0ab181ed010a5bdb2bde149d115188f96418edea77331caf6e20b00a"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_041"
+
+[[verifier.collect]]
+command = "cd /app/task_041 && git config --global --add safe.directory /app/task_041 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_041/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-041:v0.1.0@sha256:7e3b7b1a0ab181ed010a5bdb2bde149d115188f96418edea77331caf6e20b00a
+FROM ${VERIFIER_IMAGE}

tasks/task_041/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_042/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.12-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_042
+COPY repo/ /app/task_042/source/
+COPY public/ /app/task_042/
+COPY requirements.lock /app/task_042/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_042:/app/task_042/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_042/instruction.md

@@ -0,0 +1,29 @@
+# Repair an inconsistent solar-coordinate archive workflow
+
+A solar-observation catalogue uses the supplied scientific package to attach
+coordinate objects to a table before archiving the catalogue for later
+analysis. The coordinates describe a short sequence of features observed from
+changing locations. Object construction and the in-memory analysis are valid,
+but the archive/reload workflow does not yield a catalogue that can be reused
+consistently in the same downstream analysis.
+
+Reproduce the problem, inspect the source and public workflow, and repair the
+implementation so that valid solar-observation
+catalogues retain their scientific meaning through supported archive
+workflows. The repair must generalize beyond the supplied catalogue. Do not
+replace scientific objects with fixed outputs, special-case the supplied data,
+or change the public reproduction to hide a failure.
+
+Work only under this task directory. The evaluation environment is Linux and
+has no network access. You may modify files under `source/`; do not modify the
+task statement, reproduction, fixtures, workflow, or evaluation infrastructure.
+
+Run the public smoke test with:
+
+```bash
+python reproduce.py
+```
+
+The command returns `1` when the untouched source reproduces the scientific
+archive failure, `0` after a successful repair, and `2` for a runner or
+environment failure.

tasks/task_042/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "042",
+  "title": "Repair an inconsistent solar-coordinate archive workflow",
+  "domain": "solar-physics-scientific-data-semantics",
+  "language": "python",
+  "source_repository": "https://github.com/sunpy/sunpy",
+  "source_commit": "c961a1507a83b9dab24ccce12ab92a53b30359d7",
+  "source_license": "UNKNOWN",
+  "license_source": "LICENSE.rst",
+  "gpl_family": false,
+  "public_payload_sha256": "2b237d17e4f9409c1a9f77b47e60d782df91ca4aa81205bb312c597781819fe9"
+}

tasks/task_042/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-042"
+description = "Repair an inconsistent solar-coordinate archive workflow"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "042"
+language = "python"
+base_commit_hash = "c961a1507a83b9dab24ccce12ab92a53b30359d7"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-042:v0.1.0@sha256:38b71d095fa30606c636c19aa2145fde1f74993809cff837b6da5e5597895181"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_042"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-042:v0.1.0@sha256:133f5154d63c2614e6e754d68ca91bef2a8ea4af1210de147a794c570c06d8f9"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_042"
+
+[[verifier.collect]]
+command = "cd /app/task_042 && git config --global --add safe.directory /app/task_042 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_042/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-042:v0.1.0@sha256:133f5154d63c2614e6e754d68ca91bef2a8ea4af1210de147a794c570c06d8f9
+FROM ${VERIFIER_IMAGE}

tasks/task_042/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_043/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_043
+COPY repo/ /app/task_043/source/
+COPY public/ /app/task_043/
+COPY requirements.lock /app/task_043/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_043:/app/task_043/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_043/instruction.md

@@ -0,0 +1,19 @@
+# Repair an inconsistent porous-volume characterization
+
+A porous-media research group is reprocessing segmented X-ray tomography
+data from sandstone with this historical PoreSpy source snapshot.  The
+supplied offline workflow completes normally, but the summaries produced from
+the supplied segmented volume and an explicit axis-permuted representation of
+that same volume are not mutually consistent. A radial two-point
+characterization should not depend on which array axis stores a physical
+direction.
+
+Run the reproduction, inspect the supplied sample data and project source, and
+repair the implementation so the
+scientific characterization is consistent with the definitions supported by
+the project.  The repair must generalize to other valid segmented images and
+documented analysis options.
+
+Do not hard-code the supplied volume, permutation, any public observation,
+fixed dimensions, or filename-specific behavior. Do not modify the scientific
+input, workflow, public observations, or generated reports.

tasks/task_043/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "043",
+  "title": "Repair an inconsistent porous-volume characterization",
+  "domain": "porous-media-image-analysis",
+  "language": "python",
+  "source_repository": "https://github.com/PMEAL/porespy",
+  "source_commit": "f2360a52bc77ec917aa07ce0cf6128dc7687ea64",
+  "source_license": "MIT",
+  "license_source": "LICENSE",
+  "gpl_family": false,
+  "public_payload_sha256": "dc0ef803237a7f1beab83b5eeaa8fe68fb3c39a553bb28ce9acde348e912b0be"
+}

tasks/task_043/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-043"
+description = "Repair an inconsistent porous-volume characterization"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "043"
+language = "python"
+base_commit_hash = "f2360a52bc77ec917aa07ce0cf6128dc7687ea64"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-043:v0.1.0@sha256:3e82e859587bdd681272bac9ecb4a415e77fe9f2160afe76a9f0bc641cf7096f"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_043"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-043:v0.1.0@sha256:c9a3a93ea8eca723583d471b478202553f011baaaee8411063a1aabec7c1651e"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_043"
+
+[[verifier.collect]]
+command = "cd /app/task_043 && git config --global --add safe.directory /app/task_043 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_043/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-043:v0.1.0@sha256:c9a3a93ea8eca723583d471b478202553f011baaaee8411063a1aabec7c1651e
+FROM ${VERIFIER_IMAGE}

tasks/task_043/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_044/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_044
+COPY repo/ /app/task_044/source/
+COPY public/ /app/task_044/
+COPY requirements.lock /app/task_044/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_044:/app/task_044/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_044/instruction.md

@@ -0,0 +1,17 @@
+# Repair inconsistent battery-model predictions
+
+A battery-modeling group is using this historical PyBaMM snapshot to study a
+realistic lithium-ion cell parameterised for the supplied study. A
+valid offline discharge workflow completes, but its scientific output does not
+satisfy the discharge behavior required by the model and protocol.
+
+Run the supplied reproduction, inspect the complete project source and task
+workflow. Diagnose the cause of the inconsistent battery prediction and repair
+the implementation so that the documented scientific workflow is correct for
+supported inputs.
+
+The repair must be general. Do not hard-code the supplied result, parameter
+names, model class, protocol, or file paths. Do not modify the scientific
+inputs, workflow, reproduction, or generated reports.
+Validate your change with additional legal parameter values and model paths,
+not only the public run.

tasks/task_044/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "044",
+  "title": "Repair inconsistent lithium-ion discharge predictions",
+  "domain": "electrochemical-battery-modeling",
+  "language": "python",
+  "source_repository": "https://github.com/pybamm-team/PyBaMM",
+  "source_commit": "a42d7a92dd9a2c63240c8ccb6f240e5c3c42de46",
+  "source_license": "BSD-3-Clause",
+  "license_source": "LICENSE.txt",
+  "gpl_family": false,
+  "public_payload_sha256": "735b90c38de21e5576b5f999f18e125f63adcef3b45423c2115fea41665d1bff"
+}

tasks/task_044/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-044"
+description = "Repair inconsistent lithium-ion discharge predictions"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "044"
+language = "python"
+base_commit_hash = "a42d7a92dd9a2c63240c8ccb6f240e5c3c42de46"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-044:v0.1.0@sha256:513707a2950f87958370d940ca542603b5a4b83443e5c712c916abd1cc37eeb0"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_044"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-044:v0.1.0@sha256:61f22d5c7943225471bcdeed86f32470249eaefe57453989c3af1314811f3671"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_044"
+
+[[verifier.collect]]
+command = "cd /app/task_044 && git config --global --add safe.directory /app/task_044 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_044/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-044:v0.1.0@sha256:61f22d5c7943225471bcdeed86f32470249eaefe57453989c3af1314811f3671
+FROM ${VERIFIER_IMAGE}

tasks/task_044/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_045/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_045
+COPY repo/ /app/task_045/source/
+COPY public/ /app/task_045/
+COPY requirements.lock /app/task_045/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_045:/app/task_045/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_045/instruction.md

@@ -0,0 +1,35 @@
+# Repair inconsistent finite-volume operations on a connected multi-face grid
+
+An ocean-modeling group is post-processing a tracer field from a global
+finite-volume calculation. The supplied grid is represented as several
+logically rectangular faces that meet at rotated edges. Interpolation and
+differencing across those edges are part of the same diagnostic workflow used
+to form transport and mixing budgets.
+
+The workflow completes and returns finite arrays, but a connected padding call
+can reorder the local dimensions. Values occupying the expected connected-halo
+positions then no longer come from the neighboring face declared by the
+cubed-sphere topology. The result can also depend on process-level ordering
+even though the physical grid and field are unchanged. A connected edge is a
+geometric interface: padding must preserve the input dimension order, and for a
+scalar face-label field every non-corner halo position in that layout must have
+the label of the declared neighboring face, independent of process hash seed.
+
+Inspect the complete source snapshot, the finite-volume references, the grid
+description, and the public reproduction. Repair the implementation so that
+connected-face operations are geometrically consistent for valid inputs
+supported by the project. The repair must remain correct for other connected
+multi-face grids, fields, widths, orientations, and valid scientific workflows,
+not only for the supplied probes.
+
+Do not hard-code the supplied field, face count, process seeds, edge probes,
+array values, or fixture path. Do not change the scientific input or public
+report to hide a failure. Preserve existing behavior for ordinary single-face
+grids and valid disconnected boundaries.
+
+Run `python reproduce.py` before and after your repair. The command checks a
+small, transparent subset of connected edges by deriving each required
+neighbor directly from the public `face_connections` table. It also reports a
+dimension-order comparison as a non-authoritative observation. Hidden tests
+cover the complete topology and broader field and operation families, so the
+public observations are not the complete correctness oracle.

tasks/task_045/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "045",
+  "title": "Repair inconsistent finite-volume operations on a connected multi-face grid",
+  "domain": "computational-oceanography-finite-volume-grids",
+  "language": "python",
+  "source_repository": "https://github.com/xgcm/xgcm",
+  "source_commit": "3831b46accbff35f06e4ed3831600d71d5d8da18",
+  "source_license": "MIT",
+  "license_source": "LICENSE",
+  "gpl_family": false,
+  "public_payload_sha256": "f7e001f18305b0978520b26cb621a76014f6e5d030992ae0bc4f03f76f4ad82b"
+}

tasks/task_045/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-045"
+description = "Repair inconsistent finite-volume operations on a connected multi-face grid"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "045"
+language = "python"
+base_commit_hash = "3831b46accbff35f06e4ed3831600d71d5d8da18"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-045:v0.1.0@sha256:bdf7ecf2141952410df47d69d36d45b662047a720f6ff26ceeef1c33486b3fa0"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_045"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-045:v0.1.0@sha256:a3cd9863966518102f06bbf956ddfa7b6578debf5bcf92ccfbe0d1252c3a4b20"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_045"
+
+[[verifier.collect]]
+command = "cd /app/task_045 && git config --global --add safe.directory /app/task_045 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_045/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-045:v0.1.0@sha256:a3cd9863966518102f06bbf956ddfa7b6578debf5bcf92ccfbe0d1252c3a4b20
+FROM ${VERIFIER_IMAGE}

tasks/task_045/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_046/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_046
+COPY repo/ /app/task_046/source/
+COPY public/ /app/task_046/
+COPY requirements.lock /app/task_046/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_046:/app/task_046/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_046/instruction.md

@@ -0,0 +1,16 @@
+# Repair an inconsistent X-ray spectral-timing analysis
+
+An X-ray timing group is replaying a two-band quality-control observation with
+the historical scientific-software source supplied in this task. The
+observation loads successfully and the Fourier calculation completes, but the
+returned frequency-dependent products do not satisfy the statistical behavior
+described by the supplied method material.
+
+Run the offline reproduction, inspect the complete observation and historical
+source, and repair the implementation so the reconstructed analysis agrees
+with the required observation semantics. The repair must generalize to
+other valid simultaneous light-curve inputs supported by the project.
+
+Do not hard-code this observation, a filename-specific result, or report
+fields. Do not modify the supplied observation, public reproduction, or
+generated reports.

tasks/task_046/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "046",
+  "title": "Repair an inconsistent X-ray spectral-timing analysis",
+  "domain": "astrophysical-xray-spectral-timing",
+  "language": "python",
+  "source_repository": "https://github.com/StingraySoftware/stingray",
+  "source_commit": "79b8f9bb9d122bd0176bb37f48f0fd4926eb9afb",
+  "source_license": "MIT",
+  "license_source": "LICENSE.rst",
+  "gpl_family": false,
+  "public_payload_sha256": "f358c234dfb9d428b07da73956df43b4e327aa28922de1f7380efd56563d9435"
+}

tasks/task_046/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-046"
+description = "Repair an inconsistent X-ray spectral-timing analysis"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "046"
+language = "python"
+base_commit_hash = "79b8f9bb9d122bd0176bb37f48f0fd4926eb9afb"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-046:v0.1.0@sha256:5113cc358a6f5df25d7b7c09e02cb5b9bd9964248f2d51ead28308c90febd15d"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_046"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-046:v0.1.0@sha256:73e5bc8338655f2d15e7dd871c42544ec9e7d16af3f24e6ecb69030f71edefbf"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_046"
+
+[[verifier.collect]]
+command = "cd /app/task_046 && git config --global --add safe.directory /app/task_046 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_046/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-046:v0.1.0@sha256:73e5bc8338655f2d15e7dd871c42544ec9e7d16af3f24e6ecb69030f71edefbf
+FROM ${VERIFIER_IMAGE}

tasks/task_046/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_047/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_047
+COPY repo/ /app/task_047/source/
+COPY public/ /app/task_047/
+COPY requirements.lock /app/task_047/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_047:/app/task_047/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_047/instruction.md

@@ -0,0 +1,20 @@
+# Repair an inconsistent broadband microwave composite
+
+I am replaying a lossy two-port measurement workflow with the supplied
+historical scientific-library source, two measured microstrip records, and a
+declared replay configuration. The calculation completes, the frequency sweep
+is nonempty, and the reported response is finite. However, the end-to-end
+network behavior is not scientifically consistent with the supplied material
+for the same records and configuration.
+
+Run `python reproduce.py` to reproduce the observation. Inspect the complete
+source tree and public measurement records, then repair the
+library implementation so this workflow and related supported network
+workflows are scientifically consistent. Preserve the established public APIs.
+The repair must generalize beyond the supplied records, frequencies, and replay
+values; do not hard-code the public fixture or replace the scientific
+calculation with stored output.
+
+The task is fully offline. The Touchstone files are authentic upstream
+measurement examples. The replay configuration is a fixed input record, not
+code under repair.

tasks/task_047/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "047",
+  "title": "Repair an inconsistent broadband microwave composite",
+  "domain": "microwave-network-analysis",
+  "language": "python",
+  "source_repository": "https://github.com/scikit-rf/scikit-rf",
+  "source_commit": "a6dae181dbcf76fc738fc7fbc03d7de0a6c22634",
+  "source_license": "BSD-3-Clause",
+  "license_source": "LICENSE.txt",
+  "gpl_family": false,
+  "public_payload_sha256": "fa85f0105f3ab7a955fd95d203fbe54b220508b0f1d04065e60bc9bb1aac505b"
+}

tasks/task_047/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-047"
+description = "Repair an inconsistent broadband microwave composite"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "047"
+language = "python"
+base_commit_hash = "a6dae181dbcf76fc738fc7fbc03d7de0a6c22634"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-047:v0.1.0@sha256:a94de21354f66e978943e80946546812dbf404c4468baf12a84d23c34720e3d6"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_047"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-047:v0.1.0@sha256:8529979be314893a77f9c51cf5d2c0d6266d6f9df10fb33d332eb242f38953db"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_047"
+
+[[verifier.collect]]
+command = "cd /app/task_047 && git config --global --add safe.directory /app/task_047 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_047/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-047:v0.1.0@sha256:8529979be314893a77f9c51cf5d2c0d6266d6f9df10fb33d332eb242f38953db
+FROM ${VERIFIER_IMAGE}

tasks/task_047/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_048/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_048
+COPY repo/ /app/task_048/source/
+COPY public/ /app/task_048/
+COPY requirements.lock /app/task_048/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_048:/app/task_048/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_048/instruction.md

@@ -0,0 +1,27 @@
+# Repair an inconsistent alloy equilibrium assessment
+
+A materials-thermodynamics group is replaying literature-linked alloy studies
+with the supplied historical pycalphad source and databases. The calculations
+complete and return finite datasets, but several observations no longer agree
+with the group's archived results:
+
+- a composition-temperature phase boundary moves when an equivalent database
+  representation is admitted to the same assessment;
+- the associated phase fractions and internal-state evolution cease to form a
+  physically consistent transition;
+- chemical-potential and heat-response curves disagree with the corresponding
+  baseline calculation; and
+- the equilibrium topology depends on a representation that should not create
+  a new thermodynamic state for the requested components.
+
+Run the offline workflow, inspect the databases and source snapshot, and repair
+the scientific
+implementation. Determine the common cause rather than treating the symptoms
+independently. The repair must generalize to other valid CALPHAD database
+representations and supported phase states while preserving public APIs and
+unrelated model behavior.
+
+Do not hard-code the supplied grids or generated artifacts, and do not modify
+the scientific input files. The public workflow only establishes that the real
+project calculation runs and saves results; correctness is evaluated on
+independent scientific cases.

tasks/task_048/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "048",
+  "title": "Repair an inconsistent alloy equilibrium assessment",
+  "domain": "computational-thermodynamics-calphad",
+  "language": "python",
+  "source_repository": "",
+  "source_commit": "",
+  "source_license": "MIT",
+  "license_source": "LICENSE.txt",
+  "gpl_family": false,
+  "public_payload_sha256": "5a22842c3c77c8730dca22e3e922c35f6b0018b981a092d0ef6752cd92855917"
+}

tasks/task_048/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-048"
+description = "Repair an inconsistent alloy equilibrium assessment"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "048"
+language = "python"
+base_commit_hash = ""
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-048:v0.1.0@sha256:38893d5510a1f2c722c39dc851c1f4c0698c52602753e0e9a0531ea466ee9a4b"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_048"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-048:v0.1.0@sha256:66acaa226780910a92acf5076beb045a8f521ec7f222277fb4fa1e37296d46da"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_048"
+
+[[verifier.collect]]
+command = "cd /app/task_048 && git config --global --add safe.directory /app/task_048 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_048/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-048:v0.1.0@sha256:66acaa226780910a92acf5076beb045a8f521ec7f222277fb4fa1e37296d46da
+FROM ${VERIFIER_IMAGE}

tasks/task_048/tests/test.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -eu
+
+# Pier requires a public tests entrypoint when loading a local task. In
+# separate-verifier mode the executable grader remains inside the verifier image.
+exec python /tests/grader.py

tasks/task_049/environment/Dockerfile

@@ -0,0 +1,60 @@
+ARG BASE_IMAGE=python:3.11-slim
+FROM ${BASE_IMAGE}
+
+ARG DEBIAN_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian
+ARG DEBIAN_SECURITY_MIRROR=https://mirrors.tuna.tsinghua.edu.cn/debian-security
+ARG PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
+ARG INSTALL_FORTRAN=0
+ARG INSTALL_OCTAVE=0
+ENV DEBIAN_FRONTEND=noninteractive \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PIP_INDEX_URL=${PIP_INDEX_URL} \
+    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+RUN sed -i \
+        -e "s|http://deb.debian.org/debian-security|${DEBIAN_SECURITY_MIRROR}|g" \
+        -e "s|http://deb.debian.org/debian|${DEBIAN_MIRROR}|g" \
+        /etc/apt/sources.list.d/debian.sources
+
+RUN apt-get -o Acquire::Retries=10 update \
+    && apt-get -o Acquire::Retries=10 install -y --no-install-recommends \
+        build-essential \
+        ca-certificates \
+        curl \
+        git \
+        ripgrep \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN if [ "${INSTALL_FORTRAN}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends gfortran \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi \
+    && if [ "${INSTALL_OCTAVE}" = "1" ]; then \
+        apt-get -o Acquire::Retries=10 update \
+        && apt-get -o Acquire::Retries=10 install -y --no-install-recommends octave \
+        && rm -rf /var/lib/apt/lists/*; \
+    fi
+
+WORKDIR /app/task_049
+COPY repo/ /app/task_049/source/
+COPY public/ /app/task_049/
+COPY requirements.lock /app/task_049/requirements.lock
+
+RUN python -m pip install --no-cache-dir --upgrade pip \
+    && python -m pip install --no-cache-dir -r requirements.lock
+ENV PYTHONPATH=/app/task_049:/app/task_049/source
+
+RUN git init \
+    && git checkout -b main \
+    && git config user.email science-bench@example.invalid \
+    && git config user.name "SWE-bench Science baseline" \
+    && git add -A \
+    && GIT_AUTHOR_DATE=2025-01-01T00:00:00Z \
+       GIT_COMMITTER_DATE=2025-01-01T00:00:00Z \
+       git commit -m baseline \
+    && git reflog expire --expire=now --all \
+    && git gc --prune=now
+
+CMD ["bash"]

tasks/task_049/instruction.md

@@ -0,0 +1,31 @@
+# Restore the stochastic sampling and discretization contract
+
+A computational physics group is maintaining the supplied historical py-pde
+source. Complete stochastic `PDE.solve()` calculations remain numerically
+stable because the current implementation compensates one scaling choice with
+another. The defect is instead at the reusable interface between Gaussian
+sampling and spatial discretization.
+
+The backend Gaussian primitive is a standard-normal sampling service. For the
+same backend, seed, dtype, and array shape, its draw must not change when only
+the physical grid measure changes. Cell volumes belong to the discretization
+of a stochastic field: a variance-based solver converts continuum noise
+variance to a cell increment using the inverse cell measure. Keeping these
+responsibilities separate lets the primitive remain reusable across fields and
+backends while solvers consistently interpret the public noise variance.
+
+The package also documents a lower-level direct-realization interface. A user
+who chooses that route supplies the complete spatially discretized realization,
+including any required measure factor. That does not change the contract of the
+backend standard-normal primitive used by variance-based solvers.
+
+Run the offline reproduction, inspect the task context and complete source
+snapshot, and repair the implementation so that both sides of this contract
+hold for all supported valid inputs. Preserve deterministic PDE
+behavior, public APIs, stochastic interpretations, supported backends, device
+and dtype behavior, and unrelated functionality.
+
+Do not hard-code the public grids, seeds, controlled draw, lifecycle status, or
+generated report. The public experiment states the architecture being studied
+but does not name a target source file or prescribe a patch shape. Hidden tests
+vary geometry, backend, solver, noise interpretation, and realization route.

tasks/task_049/metadata.json

@@ -0,0 +1,12 @@
+{
+  "task_id": "049",
+  "title": "Restore consistent stochastic-field behavior on nonuniform grids",
+  "domain": "finite-difference-stochastic-partial-differential-equations",
+  "language": "python",
+  "source_repository": "https://github.com/zwicker-group/py-pde",
+  "source_commit": "a9f8256c1b4b48ac73f3df35b3e570993501b333",
+  "source_license": "MIT",
+  "license_source": "LICENSE",
+  "gpl_family": false,
+  "public_payload_sha256": "2102d49be26585ee8217bb375347adb25a008f36d62c783554d2a4e930489433"
+}

tasks/task_049/task.toml

@@ -0,0 +1,44 @@
+schema_version = "1.3"
+artifacts = ["/logs/artifacts/model.patch"]
+
+[task]
+name = "openmoss/swe-bench-science-049"
+description = "Restore consistent stochastic-field behavior on nonuniform grids"
+authors = []
+keywords = ["scientific-computing", "software-engineering"]
+
+[metadata]
+task_id = "049"
+language = "python"
+base_commit_hash = "a9f8256c1b4b48ac73f3df35b3e570993501b333"
+
+[agent]
+network_mode = "no-network"
+timeout_sec = 5400.0
+
+[environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-environment-python-task-049:v0.1.0@sha256:5d2c93c90c190f69ed1f8bb05014f34d900f4c68ca483cf428399caa5c80f339"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_049"
+
+[verifier]
+network_mode = "no-network"
+environment_mode = "separate"
+timeout_sec = 1800.0
+
+[verifier.environment]
+docker_image = "docker.io/kevinxulearning/swe-bench-science-verifier-python-task-049:v0.1.0@sha256:7a8babe7c3e8d6121a1fff866e0e6ae5871de381c8e759897a4b40244b6b76b6"
+os = "linux"
+allow_internet = false
+cpus = 2
+memory_mb = 8192
+storage_mb = 20480
+workdir = "/app/task_049"
+
+[[verifier.collect]]
+command = "cd /app/task_049 && git config --global --add safe.directory /app/task_049 && baseline=$(git rev-list --max-parents=0 --reverse HEAD | head -n 1) && git add -A && git diff --cached --binary \"$baseline\" > /logs/artifacts/model.patch"
+timeout_sec = 300.0

tasks/task_049/tests/Dockerfile

@@ -0,0 +1,2 @@
+ARG VERIFIER_IMAGE=docker.io/kevinxulearning/swe-bench-science-verifier-python-task-049:v0.1.0@sha256:7a8babe7c3e8d6121a1fff866e0e6ae5871de381c8e759897a4b40244b6b76b6
+FROM ${VERIFIER_IMAGE}