diff --git a/README.md b/README.md index 822d4be7f359befcc9ef50014836e404a60de342..ad26a9ea27a841298a66009215a2c1b47e9ced91 100644 --- a/README.md +++ b/README.md @@ -213,9 +213,11 @@ Earlier reported results used a `[-1, 1]` normalization; convert with Install `docker`, then for each task: -1. build the image from `environment/Dockerfile` — it `COPY data/ /app/data/`, +1. build the image from `environment/Dockerfile` — it builds on the public + `quay.io/jupyter/scipy-notebook` image (see below), `COPY data/ /app/data/`, creates the `agent` and `verifier` users declared in `task.toml`, and asserts that no ground-truth material is present in the image; + the container is meant to be started and then driven with `docker exec`; 2. run the agent as the `agent` user; it must leave its report at `/app/answer.txt`; 3. run `tests/test.sh` as the `verifier` user — it sources `tests/judge.env`, @@ -233,8 +235,15 @@ ANTHROPIC_MODEL # judge model id The file ships with placeholders (``, ``, ``) — fill in your own endpoint and token before running. -The same applies to `environment/Dockerfile`, which builds from a -`:latest` placeholder and public package mirrors. + +`environment/Dockerfile` needs no such placeholders. It builds from +`quay.io/jupyter/scipy-notebook`, pinned to a dated tag: a public Python +(conda-forge) environment that already ships the data-science stack these tasks +use — pandas, numpy, scipy, statsmodels, scikit-learn, matplotlib, h5py, +openpyxl, pytables, dask — and pulls the remaining analysis packages (pyarrow, +pypdf, pillow) from public PyPI on top. Any image carrying `python3`/`pip` can be +substituted: the apt and pip steps probe for existence first and skip whatever is +already installed. `judge.py` prefers `JUDGE_*` over `ANTHROPIC_*` when both are set, streams the request (so a long judgment is never killed by a total-time timeout) and falls diff --git a/dataworld-p000014-cause-1/environment/Dockerfile b/dataworld-p000014-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000014-cause-1/environment/Dockerfile +++ b/dataworld-p000014-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000014-consequence-2/environment/Dockerfile b/dataworld-p000014-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000014-consequence-2/environment/Dockerfile +++ b/dataworld-p000014-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000049-cause-2/environment/Dockerfile b/dataworld-p000049-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000049-cause-2/environment/Dockerfile +++ b/dataworld-p000049-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000049-consequence-2/environment/Dockerfile b/dataworld-p000049-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000049-consequence-2/environment/Dockerfile +++ b/dataworld-p000049-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000072-cause-2/environment/Dockerfile b/dataworld-p000072-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000072-cause-2/environment/Dockerfile +++ b/dataworld-p000072-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000072-consequence-1/environment/Dockerfile b/dataworld-p000072-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000072-consequence-1/environment/Dockerfile +++ b/dataworld-p000072-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000096-cause-2/environment/Dockerfile b/dataworld-p000096-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000096-cause-2/environment/Dockerfile +++ b/dataworld-p000096-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000096-consequence-1/environment/Dockerfile b/dataworld-p000096-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000096-consequence-1/environment/Dockerfile +++ b/dataworld-p000096-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000129-cause-2/environment/Dockerfile b/dataworld-p000129-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000129-cause-2/environment/Dockerfile +++ b/dataworld-p000129-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000129-consequence-1/environment/Dockerfile b/dataworld-p000129-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000129-consequence-1/environment/Dockerfile +++ b/dataworld-p000129-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000151-cause-2/environment/Dockerfile b/dataworld-p000151-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000151-cause-2/environment/Dockerfile +++ b/dataworld-p000151-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000151-consequence-1/environment/Dockerfile b/dataworld-p000151-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000151-consequence-1/environment/Dockerfile +++ b/dataworld-p000151-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000183-cause-2/environment/Dockerfile b/dataworld-p000183-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000183-cause-2/environment/Dockerfile +++ b/dataworld-p000183-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000183-consequence-1/environment/Dockerfile b/dataworld-p000183-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000183-consequence-1/environment/Dockerfile +++ b/dataworld-p000183-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000211-cause-2/environment/Dockerfile b/dataworld-p000211-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000211-cause-2/environment/Dockerfile +++ b/dataworld-p000211-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000211-consequence-1/environment/Dockerfile b/dataworld-p000211-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000211-consequence-1/environment/Dockerfile +++ b/dataworld-p000211-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000213-cause-2/environment/Dockerfile b/dataworld-p000213-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000213-cause-2/environment/Dockerfile +++ b/dataworld-p000213-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000213-consequence-2/environment/Dockerfile b/dataworld-p000213-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000213-consequence-2/environment/Dockerfile +++ b/dataworld-p000213-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000221-cause-2/environment/Dockerfile b/dataworld-p000221-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000221-cause-2/environment/Dockerfile +++ b/dataworld-p000221-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000221-consequence-2/environment/Dockerfile b/dataworld-p000221-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000221-consequence-2/environment/Dockerfile +++ b/dataworld-p000221-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000231-cause-1/environment/Dockerfile b/dataworld-p000231-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000231-cause-1/environment/Dockerfile +++ b/dataworld-p000231-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000231-consequence-2/environment/Dockerfile b/dataworld-p000231-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000231-consequence-2/environment/Dockerfile +++ b/dataworld-p000231-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000233-cause-1/environment/Dockerfile b/dataworld-p000233-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000233-cause-1/environment/Dockerfile +++ b/dataworld-p000233-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000233-consequence-1/environment/Dockerfile b/dataworld-p000233-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000233-consequence-1/environment/Dockerfile +++ b/dataworld-p000233-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000236-cause-2/environment/Dockerfile b/dataworld-p000236-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000236-cause-2/environment/Dockerfile +++ b/dataworld-p000236-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000236-consequence-1/environment/Dockerfile b/dataworld-p000236-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000236-consequence-1/environment/Dockerfile +++ b/dataworld-p000236-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000249-cause-1/environment/Dockerfile b/dataworld-p000249-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000249-cause-1/environment/Dockerfile +++ b/dataworld-p000249-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000249-consequence-1/environment/Dockerfile b/dataworld-p000249-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000249-consequence-1/environment/Dockerfile +++ b/dataworld-p000249-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000341-cause-2/environment/Dockerfile b/dataworld-p000341-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000341-cause-2/environment/Dockerfile +++ b/dataworld-p000341-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000341-consequence-2/environment/Dockerfile b/dataworld-p000341-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000341-consequence-2/environment/Dockerfile +++ b/dataworld-p000341-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000348-cause-2/environment/Dockerfile b/dataworld-p000348-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000348-cause-2/environment/Dockerfile +++ b/dataworld-p000348-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000348-consequence-1/environment/Dockerfile b/dataworld-p000348-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000348-consequence-1/environment/Dockerfile +++ b/dataworld-p000348-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000350-cause-2/environment/Dockerfile b/dataworld-p000350-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000350-cause-2/environment/Dockerfile +++ b/dataworld-p000350-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000350-consequence-2/environment/Dockerfile b/dataworld-p000350-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000350-consequence-2/environment/Dockerfile +++ b/dataworld-p000350-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000372-cause-2/environment/Dockerfile b/dataworld-p000372-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000372-cause-2/environment/Dockerfile +++ b/dataworld-p000372-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000372-consequence-1/environment/Dockerfile b/dataworld-p000372-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000372-consequence-1/environment/Dockerfile +++ b/dataworld-p000372-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000384-cause-1/environment/Dockerfile b/dataworld-p000384-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000384-cause-1/environment/Dockerfile +++ b/dataworld-p000384-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000384-consequence-1/environment/Dockerfile b/dataworld-p000384-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000384-consequence-1/environment/Dockerfile +++ b/dataworld-p000384-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000419-cause-2/environment/Dockerfile b/dataworld-p000419-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000419-cause-2/environment/Dockerfile +++ b/dataworld-p000419-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000419-consequence-1/environment/Dockerfile b/dataworld-p000419-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000419-consequence-1/environment/Dockerfile +++ b/dataworld-p000419-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000478-cause-1/environment/Dockerfile b/dataworld-p000478-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000478-cause-1/environment/Dockerfile +++ b/dataworld-p000478-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000478-consequence-1/environment/Dockerfile b/dataworld-p000478-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000478-consequence-1/environment/Dockerfile +++ b/dataworld-p000478-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000565-cause-1/environment/Dockerfile b/dataworld-p000565-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000565-cause-1/environment/Dockerfile +++ b/dataworld-p000565-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000565-consequence-1/environment/Dockerfile b/dataworld-p000565-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000565-consequence-1/environment/Dockerfile +++ b/dataworld-p000565-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000619-cause-2/environment/Dockerfile b/dataworld-p000619-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000619-cause-2/environment/Dockerfile +++ b/dataworld-p000619-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000619-consequence-1/environment/Dockerfile b/dataworld-p000619-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000619-consequence-1/environment/Dockerfile +++ b/dataworld-p000619-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000631-cause-1/environment/Dockerfile b/dataworld-p000631-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000631-cause-1/environment/Dockerfile +++ b/dataworld-p000631-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000631-consequence-2/environment/Dockerfile b/dataworld-p000631-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000631-consequence-2/environment/Dockerfile +++ b/dataworld-p000631-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000917-cause-1/environment/Dockerfile b/dataworld-p000917-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000917-cause-1/environment/Dockerfile +++ b/dataworld-p000917-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000917-consequence-1/environment/Dockerfile b/dataworld-p000917-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000917-consequence-1/environment/Dockerfile +++ b/dataworld-p000917-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000923-cause-2/environment/Dockerfile b/dataworld-p000923-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000923-cause-2/environment/Dockerfile +++ b/dataworld-p000923-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000923-consequence-1/environment/Dockerfile b/dataworld-p000923-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000923-consequence-1/environment/Dockerfile +++ b/dataworld-p000923-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000924-cause-2/environment/Dockerfile b/dataworld-p000924-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000924-cause-2/environment/Dockerfile +++ b/dataworld-p000924-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p000924-consequence-2/environment/Dockerfile b/dataworld-p000924-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p000924-consequence-2/environment/Dockerfile +++ b/dataworld-p000924-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001067-cause-1/environment/Dockerfile b/dataworld-p001067-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001067-cause-1/environment/Dockerfile +++ b/dataworld-p001067-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001067-consequence-1/environment/Dockerfile b/dataworld-p001067-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001067-consequence-1/environment/Dockerfile +++ b/dataworld-p001067-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001071-cause-2/environment/Dockerfile b/dataworld-p001071-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001071-cause-2/environment/Dockerfile +++ b/dataworld-p001071-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001071-consequence-1/environment/Dockerfile b/dataworld-p001071-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001071-consequence-1/environment/Dockerfile +++ b/dataworld-p001071-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001083-cause-1/environment/Dockerfile b/dataworld-p001083-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001083-cause-1/environment/Dockerfile +++ b/dataworld-p001083-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001083-consequence-1/environment/Dockerfile b/dataworld-p001083-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001083-consequence-1/environment/Dockerfile +++ b/dataworld-p001083-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001115-cause-1/environment/Dockerfile b/dataworld-p001115-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001115-cause-1/environment/Dockerfile +++ b/dataworld-p001115-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001115-consequence-2/environment/Dockerfile b/dataworld-p001115-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001115-consequence-2/environment/Dockerfile +++ b/dataworld-p001115-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001119-cause-1/environment/Dockerfile b/dataworld-p001119-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001119-cause-1/environment/Dockerfile +++ b/dataworld-p001119-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001119-consequence-1/environment/Dockerfile b/dataworld-p001119-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001119-consequence-1/environment/Dockerfile +++ b/dataworld-p001119-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001147-cause-2/environment/Dockerfile b/dataworld-p001147-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001147-cause-2/environment/Dockerfile +++ b/dataworld-p001147-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001147-consequence-1/environment/Dockerfile b/dataworld-p001147-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001147-consequence-1/environment/Dockerfile +++ b/dataworld-p001147-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001201-cause-1/environment/Dockerfile b/dataworld-p001201-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001201-cause-1/environment/Dockerfile +++ b/dataworld-p001201-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001201-consequence-1/environment/Dockerfile b/dataworld-p001201-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001201-consequence-1/environment/Dockerfile +++ b/dataworld-p001201-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001203-cause-2/environment/Dockerfile b/dataworld-p001203-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001203-cause-2/environment/Dockerfile +++ b/dataworld-p001203-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001203-consequence-2/environment/Dockerfile b/dataworld-p001203-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001203-consequence-2/environment/Dockerfile +++ b/dataworld-p001203-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001219-cause-2/environment/Dockerfile b/dataworld-p001219-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001219-cause-2/environment/Dockerfile +++ b/dataworld-p001219-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001219-consequence-1/environment/Dockerfile b/dataworld-p001219-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001219-consequence-1/environment/Dockerfile +++ b/dataworld-p001219-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001223-cause-2/environment/Dockerfile b/dataworld-p001223-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001223-cause-2/environment/Dockerfile +++ b/dataworld-p001223-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001223-consequence-2/environment/Dockerfile b/dataworld-p001223-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001223-consequence-2/environment/Dockerfile +++ b/dataworld-p001223-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001237-cause-2/environment/Dockerfile b/dataworld-p001237-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001237-cause-2/environment/Dockerfile +++ b/dataworld-p001237-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001237-consequence-1/environment/Dockerfile b/dataworld-p001237-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001237-consequence-1/environment/Dockerfile +++ b/dataworld-p001237-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001290-cause-2/environment/Dockerfile b/dataworld-p001290-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001290-cause-2/environment/Dockerfile +++ b/dataworld-p001290-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001290-consequence-2/environment/Dockerfile b/dataworld-p001290-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001290-consequence-2/environment/Dockerfile +++ b/dataworld-p001290-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001313-cause-2/environment/Dockerfile b/dataworld-p001313-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001313-cause-2/environment/Dockerfile +++ b/dataworld-p001313-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001313-consequence-1/environment/Dockerfile b/dataworld-p001313-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001313-consequence-1/environment/Dockerfile +++ b/dataworld-p001313-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001656-cause-2/environment/Dockerfile b/dataworld-p001656-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001656-cause-2/environment/Dockerfile +++ b/dataworld-p001656-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p001656-consequence-1/environment/Dockerfile b/dataworld-p001656-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p001656-consequence-1/environment/Dockerfile +++ b/dataworld-p001656-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p002876-cause-2/environment/Dockerfile b/dataworld-p002876-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p002876-cause-2/environment/Dockerfile +++ b/dataworld-p002876-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p002876-consequence-1/environment/Dockerfile b/dataworld-p002876-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p002876-consequence-1/environment/Dockerfile +++ b/dataworld-p002876-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p004607-cause-2/environment/Dockerfile b/dataworld-p004607-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p004607-cause-2/environment/Dockerfile +++ b/dataworld-p004607-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p004607-consequence-2/environment/Dockerfile b/dataworld-p004607-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p004607-consequence-2/environment/Dockerfile +++ b/dataworld-p004607-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006499-cause-2/environment/Dockerfile b/dataworld-p006499-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006499-cause-2/environment/Dockerfile +++ b/dataworld-p006499-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006499-consequence-2/environment/Dockerfile b/dataworld-p006499-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006499-consequence-2/environment/Dockerfile +++ b/dataworld-p006499-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006632-cause-1/environment/Dockerfile b/dataworld-p006632-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006632-cause-1/environment/Dockerfile +++ b/dataworld-p006632-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006632-consequence-2/environment/Dockerfile b/dataworld-p006632-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006632-consequence-2/environment/Dockerfile +++ b/dataworld-p006632-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006802-cause-2/environment/Dockerfile b/dataworld-p006802-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006802-cause-2/environment/Dockerfile +++ b/dataworld-p006802-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p006802-consequence-2/environment/Dockerfile b/dataworld-p006802-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p006802-consequence-2/environment/Dockerfile +++ b/dataworld-p006802-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p007107-cause-1/environment/Dockerfile b/dataworld-p007107-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p007107-cause-1/environment/Dockerfile +++ b/dataworld-p007107-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p007107-consequence-2/environment/Dockerfile b/dataworld-p007107-consequence-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p007107-consequence-2/environment/Dockerfile +++ b/dataworld-p007107-consequence-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009238-cause-1/environment/Dockerfile b/dataworld-p009238-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009238-cause-1/environment/Dockerfile +++ b/dataworld-p009238-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009238-consequence-1/environment/Dockerfile b/dataworld-p009238-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009238-consequence-1/environment/Dockerfile +++ b/dataworld-p009238-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009406-cause-1/environment/Dockerfile b/dataworld-p009406-cause-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009406-cause-1/environment/Dockerfile +++ b/dataworld-p009406-cause-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009406-consequence-1/environment/Dockerfile b/dataworld-p009406-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009406-consequence-1/environment/Dockerfile +++ b/dataworld-p009406-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009851-cause-2/environment/Dockerfile b/dataworld-p009851-cause-2/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009851-cause-2/environment/Dockerfile +++ b/dataworld-p009851-cause-2/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors). diff --git a/dataworld-p009851-consequence-1/environment/Dockerfile b/dataworld-p009851-consequence-1/environment/Dockerfile index 980916ac9506a4dacaf908ecfc0adde7b6a9aaf8..fccc2e05a17a2eaa22cf3afb6a2351118202c7ee 100644 --- a/dataworld-p009851-consequence-1/environment/Dockerfile +++ b/dataworld-p009851-consequence-1/environment/Dockerfile @@ -3,10 +3,33 @@ # required. # Leak prevention: the whitelist COPYs only data/ -- WORLD.md, .build and dag.json must # never make it into the image. -# [RELEASE] Replace the base image below with your own generic image (it needs -# python3/pip, otherwise the apt step below installs them); the apt/pip steps all probe -# for existence first, so anything the base image already provides is not reinstalled. -FROM :latest +# Base image: a public, batteries-included Python environment -- the Jupyter scipy stack +# (conda-forge). It already ships the data-science packages these tasks use (pandas, numpy, +# scipy, statsmodels, scikit-learn, matplotlib, h5py, openpyxl, pytables, dask, numexpr); +# pyarrow/pypdf/pillow are covered by the pip step below. The Docker Hub mirror of this +# image is stale (2023-10), so the maintained quay.io registry is used, and the tag is a +# dated release (index digest +# sha256:8f8a5763d9a01e7f488eff8d9379654788e9cc08286dbcbc6bcda129c1b0edcd) so that builds +# are reproducible -- `latest` tracks the same image. Any image with python3/pip works just +# as well; the apt/pip steps probe for existence first, so nothing the base image already +# provides is reinstalled. Two properties of this base image are reset below: it ends on a +# non-root user (this build needs root to create the agent/verifier users and to open up +# /app), and it carries a Jupyter entrypoint (the container is started by the harness with +# its own command and has to stay up for `docker exec`). +FROM quay.io/jupyter/scipy-notebook:2026-10-12 + +USER root +# Drop the inherited Jupyter entrypoint/CMD: exec whatever command the harness passes, and +# stay alive when it passes none. +ENTRYPOINT ["/bin/sh", "-c", "exec \"$@\"", "-"] +CMD ["sleep", "infinity"] + +# The conda prefix is the only Python in this image; link it into the standard /usr/local/bin +# so that python3/pip3 still resolve when the harness switches to the agent/verifier user +# with a login shell, which would otherwise drop /opt/conda/bin from PATH. +RUN set -eux; \ + ln -sf /opt/conda/bin/python3 /usr/local/bin/python3; \ + ln -sf /opt/conda/bin/pip3 /usr/local/bin/pip3 # Package mirrors -- the release build uses public mirrors (the original build # environment had no internet access and went through internal mirrors).