ci: Add caches for pre-commit envionment and test data

.github/workflows/create_cache.yml

@@ -0,0 +1,73 @@
+name: "Create GHA cache"
+
+# GitHub puts the following restrictions on cache sharing. PRs can access
+#
+# - caches that were created by the PR / earlier runs of the PR
+# - caches that were created on the target branch
+#
+# To get effective cache sharing between PRs, we create caches on the `develop`
+# branch (which is where almost all PRs merge into).
+
+on:
+  push:
+    branches: [develop]
+
+# Cancel running jobs if there's a newer push
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  # GitHub Actions cache of the pre-commit environment
+  pre-commit:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v5
+
+      - name: Setup Python 3.11
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
+
+      - uses: actions/cache@v4
+        id: cache
+        with:
+          path: ~/.cache/pre-commit
+          key: pre-commit_${{ env.pythonLocation }}_${{ hashFiles('.pre-commit-config.yaml') }}
+          lookup-only: true  # don't actually download the cache
+
+      - name: Populate pre-commit environment (if not cached)
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          pip install pre-commit
+          pre-commit install --install-hooks
+
+  # GitHub Actions cache of pySHiELD test data
+  test-data:
+    runs-on: ubuntu-latest
+    env:
+      TEST_DATA_PATH: "./test_data/8.1.3/c12_6ranks_baroclinic/physics"
+      TEST_DATA_URL: "https://portal.nccs.nasa.gov/datashare/astg/smt/pace-regression-data/8.1.3_c12_6ranks_baroclinic.physics.tar.gz"
+
+    steps:
+      - uses: actions/checkout@v5
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
+
+      - uses: actions/cache@v4
+        id: cache
+        with:
+          path: pySHiELD/test_data
+          key: ${{ env.TEST_DATA_PATH }}
+          lookup-only: true  # don't actually download the cache
+
+      - name: Download test data (if not already cached)
+        if: steps.cache-restore.outputs.cache-hit != 'true'
+        run: |
+          mkdir -p pySHiELD/test_data && cd pySHiELD/test_data
+          wget ${{ env.TEST_DATA_URL }}
+          tar -xzvf 8.1.3_c12_6ranks_baroclinic.physics.tar.gz
+          rm 8.1.3_c12_6ranks_baroclinic.physics.tar.gz

.github/workflows/lint.yaml

@@ -22,6 +22,13 @@ jobs:
         with:
           python-version: '3.11'
 
+      # Only restore (don't save) caches on PRs. New caches created from PRs won't be
+      # accessible from other PRs, see workflows/create_cache.yaml.
+      - uses: actions/cache/restore@v4
+        with:
+          path: ~/.cache/pre-commit
+          key: pre-commit_${{ env.pythonLocation }}_${{ hashFiles('.pre-commit-config.yaml') }}
+
       - name: Install pre-commit
         run: pip install pre-commit
 

.github/workflows/translate.yaml

@@ -26,6 +26,10 @@ concurrency:
 jobs:
   pyshield_translate_tests:
     runs-on: ubuntu-latest
+    env:
+      TEST_DATA_PATH: "./test_data/8.1.3/c12_6ranks_baroclinic/physics"
+      TEST_DATA_URL: "https://portal.nccs.nasa.gov/datashare/astg/smt/pace-regression-data/8.1.3_c12_6ranks_baroclinic.physics.tar.gz"
+
     steps:
       - name: External trigger Checkout pySHiELD
         if: ${{inputs.component_trigger}}
@@ -91,21 +95,29 @@ jobs:
           pip install numpy==1.26.4
           conda remove sqlite
 
-      - name: Download data
+      # Only restore (don't save) caches on PRs. New caches created from PRs won't be
+      # accessible from other PRs, see workflows/create_cache.yaml.
+      - name: Restore test_data (if cached)
+        id: cache-restore
+        uses: actions/cache/restore@v4
+        with:
+          key: ${{ env.TEST_DATA_PATH }}
+          path: pySHiELD/test_data
+
+      - name: Download data (if not cached)
+        if: steps.cache-restore.outputs.cache-hit != 'true'
         shell: bash -l {0}
         run: |
-          cd ${GITHUB_WORKSPACE}/pySHiELD
-          mkdir -p test_data
-          cd test_data
-          wget https://portal.nccs.nasa.gov/datashare/astg/smt/pace-regression-data/8.1.3_c12_6ranks_baroclinic.physics.tar.gz
+          mkdir -p pySHiELD/test_data && cd pySHiELD/test_data
+          wget ${{ env.TEST_DATA_URL }}
           tar -xzvf 8.1.3_c12_6ranks_baroclinic.physics.tar.gz
 
       - name: Numpy Translate Test
         shell: bash -l {0}
         run: |
           cd ${GITHUB_WORKSPACE}/pySHiELD
           pytest \
-            -v -s --data_path=./test_data/8.1.3/c12_6ranks_baroclinic/physics \
+            -v -s --data_path=${{ env.TEST_DATA_PATH }} \
             --backend=numpy \
             --threshold_overrides_file=./tests/savepoint/translate/overrides/standard.yaml \
             ./tests/savepoint
@@ -119,7 +131,7 @@ jobs:
           export OMP_NUM_THREADS=10
           export PACE_LOGLEVEL=Debug
           pytest \
-            -vvv -x -s --data_path=./test_data/8.1.3/c12_6ranks_baroclinic/physics \
+            -vvv -x -s --data_path=${{ env.TEST_DATA_PATH }} \
             --backend=dace:cpu \
             --threshold_overrides_file=./tests/savepoint/translate/overrides/standard.yaml \
             ./tests/savepoint