Use hugging face to store test data

.github/workflows/apptainer.yml

@@ -43,12 +43,9 @@ jobs:
         with:
           python-version: "3.12"
 
-      - name: Resolve zenodo checksum for cache key
-        id: zenodo-record
-        run: |
-          DOI=$(grep -oP 'DOI="\K[^"]+' download_data_for_tests.sh)
-          uv run zenodo_get --doi "${DOI}" --md5 --output-dir /tmp/zenodo-md5
-          echo "id=$(md5sum /tmp/zenodo-md5/md5sums.txt | cut -d' ' -f1)" >> "$GITHUB_OUTPUT"
+      - name: Resolve test data revision for cache key
+        id: test-data-revision
+        run: echo "id=$(grep -oP 'REVISION="\K[^"]+' download_data_for_tests.sh)" >> "$GITHUB_OUTPUT"
 
       - name: Cache test data
         id: cache-test-data
@@ -62,7 +59,7 @@ jobs:
             tests/io/omni/data
             tests/io/RBMDataSet/data
             tests/io/solar_wind/data
-          key: zenodo-test-data-${{ steps.zenodo-record.outputs.id }}
+          key: hf-test-data-${{ steps.test-data-revision.outputs.id }}
 
       - name: Download test data
         if: steps.cache-test-data.outputs.cache-hit != 'true'

.github/workflows/tests.yml

@@ -34,7 +34,7 @@ jobs:
           enable-cache: true
 
       - name: Set up Python
-        run: uv python install
+        run: uv python install 3.12
 
       - name: Install project and ty
         run: |
@@ -66,12 +66,9 @@ jobs:
           uv sync --locked
           uv pip install coveralls coverage pytest-cov
 
-      - name: Resolve zenodo checksum for cache key
-        id: zenodo-record
-        run: |
-          DOI=$(grep -oP 'DOI="\K[^"]+' download_data_for_tests.sh)
-          uv run zenodo_get --doi "${DOI}" --md5 --output-dir /tmp/zenodo-md5
-          echo "id=$(md5sum /tmp/zenodo-md5/md5sums.txt | cut -d' ' -f1)" >> "$GITHUB_OUTPUT"
+      - name: Resolve test data revision for cache key
+        id: test-data-revision
+        run: echo "id=$(grep -oP 'REVISION="\K[^"]+' download_data_for_tests.sh)" >> "$GITHUB_OUTPUT"
 
       - name: Cache test data
         id: cache-test-data
@@ -85,7 +82,7 @@ jobs:
             tests/io/omni/data
             tests/io/RBMDataSet/data
             tests/io/solar_wind/data
-          key: zenodo-test-data-${{ steps.zenodo-record.outputs.id }}
+          key: hf-test-data-${{ steps.test-data-revision.outputs.id }}
 
       - name: Download test data
         if: steps.cache-test-data.outputs.cache-hit != 'true'

README.md

@@ -128,4 +128,4 @@ All the above `uv` commands assume you have `uv` installed, if not then remove `
 
 ## Test Data
 
-The data used by the test suite is hosted on Zenodo and is not tracked in this repository. Install the `dev` dependency group (`uv sync --group dev`), which pulls in `zenodo_get`, then run `./download_data_for_tests.sh` to fetch and unpack it.
+The data used by the test suite is hosted on Hugging Face ([spsw/SWVO-Test-Data](https://huggingface.co/datasets/spsw/SWVO-Test-Data)) and is not tracked in this repository. Run `./download_data_for_tests.sh` (requires `uv`) to fetch it into `tests/io`.

download_data_for_tests.sh

@@ -6,9 +6,15 @@
 
 set -euo pipefail
 
-DOI="10.5281/zenodo.22983305"
+REPO="spsw/SWVO-Test-Data"
+REVISION="ee8e705c77d4a80c3895f79466c0c47f3ee6c523"
 OUTPUT_DIR="."
 
-uv run zenodo_get --doi "${DOI}" --output-dir "${OUTPUT_DIR}"
-unzip -o "${OUTPUT_DIR}/swvo-test-data.zip" -d "${OUTPUT_DIR}"
-rm "${OUTPUT_DIR}/swvo-test-data.zip"
+uvx --from huggingface_hub hf download "${REPO}" --repo-type dataset --revision "${REVISION}" --local-dir "${OUTPUT_DIR}"
+
+# hf leaves its own bookkeeping behind in the output dir
+rm -rf "${OUTPUT_DIR}/.cache"
+rm -f "${OUTPUT_DIR}/.gitattributes"
+
+# Hugging Face cannot store empty directories
+mkdir -p "${OUTPUT_DIR}/tests/io/kp/data/tmp"

pyproject.toml

@@ -1,3 +1,3 @@
 # SPDX-FileCopyrightText: 2025 GFZ Helmholtz Centre for Geosciences
 #
 # SPDX-License-Identifier: Apache-2.0
@@ -135,7 +135,6 @@
     "coverage>=7.14.0",
     "pytest>=9.0.2",
     "pytest-mock>=3.15.1",
-    "zenodo-get>=3.1.0",
 ]
 docs = [
     "sphinx-copybutton",