Spark 3.4 (#611)

* Spark 3.4, multi build CI

* escape var

* update frameless, remove 3.4.0 from CI for now

* change block var

* fix reference

* don't use anchors

* use spark version in Jar name

* comment out tests

* add poetry.lock

* don't continue on error

* update poetry lock

* update frameless

* add tests back in

* change ulimit

* set timeZone

* support pyspark 3.2-3.4

* add env var for tests

---------

Co-authored-by: Thomas Maschler <thomas.maschler@cervest.earth>
Co-authored-by: Grigory <gr.pomadchin@gmail.com>

Author: 3 people

.github/actions/init-python-env/action.yaml

@@ -9,6 +9,9 @@ inputs:
   poetry_version:
     description: 'Version of Poetry to configure'
     default: '1.3.2'
+  spark_version:
+    description: 'Version of Spark to configure'
+    default: '3.4.0'
 
 runs:
   using: "composite"
@@ -36,5 +39,7 @@ runs:
 
     - name: Install Poetry project dependencies
       # if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
+      env:
+        SPARK_VERSION:  ${{ inputs.spark_version }}
       shell: bash
       run: make init-python

.github/workflows/ci.yml

@@ -15,6 +15,13 @@ jobs:
   build-scala:
     runs-on: ubuntu-20.04
 
+    strategy:
+      matrix:
+        spark_version:
+          - "3.2.4"
+          - "3.3.2"
+          - "3.4.0"
+
     steps:
       - name: Checkout Repository
         uses: actions/checkout@v3
@@ -25,22 +32,30 @@ jobs:
         uses: ./.github/actions/init-scala-env
 
       - name: Compile Scala Project
+        env:
+          SPARK_VERSION: ${{ matrix.spark_version }}
         run: make compile-scala
 
       - name: Test Scala Project
         # python/* branches are not supposed to change scala code, trust them
         if: ${{ !startsWith(github.event.inputs.from_branch, 'python/') }}
-        run: make test-scala
+        env:
+          SPARK_VERSION: ${{ matrix.spark_version }}
+        run:
+          ulimit -c unlimited
+          make test-scala
 
       - name: Build Spark Assembly
+        env:
+          SPARK_VERSION: ${{ matrix.spark_version }}
         shell: bash
         run: make build-scala
 
       - name: Cache Spark Assembly
         uses: actions/cache@v3
         with:
           path: ./dist/*
-          key: dist-${{ github.sha }}
+          key: dist-${{ matrix.spark_version }}-${{ github.sha }}
 
   build-python:
     # scala/* branches are not supposed to change python code, trust them
@@ -50,7 +65,13 @@ jobs:
 
     strategy:
       matrix:
-        python: [ "3.8" ]
+        python:
+          - "3.8"
+          - "3.9"
+        spark_version:
+          - "3.2.4"
+          - "3.3.2"
+          - "3.4.0"
 
     steps:
       - name: Checkout Repository
@@ -61,6 +82,7 @@ jobs:
       - uses: ./.github/actions/init-python-env
         with:
           python_version: ${{ matrix.python }}
+          spark_version: ${{ matrix.spark_version }}
 
       - name: Static checks
         shell: bash
@@ -69,18 +91,27 @@ jobs:
       - uses: actions/cache@v3
         with:
           path: ./dist/*
-          key: dist-${{ github.sha }}
+          key: dist-${{ matrix.spark_version }}-${{ github.sha }}
 
       - name: Run tests
+        env:
+          SPARK_VERSION: ${{ matrix.spark_version }}
         shell: bash
         run: make test-python-quick
 
-  publish:
-    name: Publish Artifacts
+  publish-scala:
+    name: Publish Scala Artifacts
     needs: [ build-scala, build-python ]
     runs-on: ubuntu-20.04
     if: (github.event_name != 'pull_request') && startsWith(github.ref, 'refs/tags/v')
 
+    strategy:
+      matrix:
+        spark_version:
+          - "3.2.4"
+          - "3.3.2"
+          - "3.4.0"
+
     steps:
       - name: Checkout Repository
         uses: actions/checkout@v3
@@ -94,17 +125,58 @@ jobs:
         shell: bash
         env:
           GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}
+          SPARK_VERSION: ${{ matrix.spark_version }}
         run: make publish-scala
 
+      - name: Build Spark Assembly
+        env:
+          SPARK_VERSION: ${{ matrix.spark_version }}
+        shell: bash
+        run: make build-scala
+
+      - name: Cache Spark Assembly
+        uses: actions/cache@v3
+        with:
+          path: ./dist/*
+          key: dist-${{ matrix.spark_version }}-${{ github.ref }}
+
+
+  publish-python:
+    name: Publish Scala Artifacts
+    needs: [ publish-scala ]
+    runs-on: ubuntu-20.04
+    if: (github.event_name != 'pull_request') && startsWith(github.ref, 'refs/tags/v')
+
+    strategy:
+      matrix:
+        python:
+          - "3.8"
+          - "3.9"
+        spark_version:
+          - "3.2.4"
+          - "3.3.2"
+          - "3.4.0"
+
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
       - uses: ./.github/actions/init-python-env
         with:
-          python_version: "3.8"
+          python_version: ${{ matrix.python }}
+          spark_version: ${{ matrix.spark_version }}
+
+      - uses: actions/cache@v3
+        with:
+          path: ./dist/*
+          key: dist-${{ matrix.spark_version }}-${{ github.ref }}
 
       - name: Build Python whl
         shell: bash
         run: make build-python
 
-
 # TODO: Where does this go, do we need it?
 #      - name: upload artefacts
 #        uses: ./.github/actions/upload_artefacts

Makefile

@@ -1,7 +1,10 @@
-SHELL := /usr/bin/env bash
+SHELL := env SPARK_VERSION=$(SPARK_VERSION) /usr/bin/env bash
+SPARK_VERSION ?= 3.4.0
 
 .PHONY: init test lint build docs notebooks help
 
+DIST_DIR = ./dist
+
 help:
 	@echo "init - Setup the repository"
 	@echo "clean - clean all compiled python files, build artifacts and virtual envs. Run \`make init\` anew afterwards."
@@ -18,27 +21,32 @@ test: test-scala test-python
 ###############
 
 compile-scala:
-	sbt -v -batch compile test:compile it:compile
+	sbt -v -batch compile test:compile it:compile -DrfSparkVersion=${SPARK_VERSION}
 
 test-scala: test-core-scala test-datasource-scala test-experimental-scala
 
 test-core-scala:
-	sbt -batch core/test
+	sbt -batch core/test -DrfSparkVersion=${SPARK_VERSION}
 
 test-datasource-scala:
-	sbt -batch datasource/test
+	sbt -batch datasource/test -DrfSparkVersion=${SPARK_VERSION}
 
 test-experimental-scala:
-	sbt -batch experimental/test
+	sbt -batch experimental/test -DrfSparkVersion=${SPARK_VERSION}
+
+build-scala: clean-build-scala
+	sbt "pyrasterframes/assembly" -DrfSparkVersion=${SPARK_VERSION}
 
-build-scala:
-	sbt "pyrasterframes/assembly"
+clean-build-scala:
+	if [ -d "$(DIST_DIR)" ]; then \
+		find ./dist -name 'pyrasterframes-assembly-${SPARK_VERSION}*.jar' -exec rm -fr {} +; \
+	fi
 
 clean-scala:
-	sbt clean
+	sbt clean -DrfSparkVersion=${SPARK_VERSION}
 
 publish-scala:
-	sbt publish
+	sbt publish -DrfSparkVersion=${SPARK_VERSION}
 
 ################
 # PYTHON
@@ -49,9 +57,11 @@ init-python:
 	./.venv/bin/python -m pip install --upgrade pip
 	poetry self add "poetry-dynamic-versioning[plugin]"
 	poetry install
+	poetry add pyspark@${SPARK_VERSION}
 	poetry run pre-commit install
 
 test-python: build-scala
+	poetry add pyspark@${SPARK_VERSION}
 	poetry run pytest -vv python/tests --cov=python/pyrasterframes --cov=python/geomesa_pyspark --cov-report=term-missing
 
 test-python-quick:
@@ -72,8 +82,10 @@ notebooks-python: clean-notebooks-python
 clean-python: clean-build-python clean-test-python clean-venv-python clean-docs-python clean-notebooks-python
 
 clean-build-python:
-	find ./dist -name 'pyrasterframes*.whl' -exec rm -fr {} +
-	find ./dist -name 'pyrasterframes*.tar.gz' -exec rm -fr {} +
+	if [ -d "$(DIST_DIR)" ]; then \
+		find ./dist -name 'pyrasterframes*.whl' -exec rm -fr {} +; \
+		find ./dist -name 'pyrasterframes*.tar.gz' -exec rm -fr {} +; \
+	fi
 
 clean-test-python:
 	rm -f .coverage

build.sbt

@@ -27,6 +27,7 @@ ThisBuild / dynverSonatypeSnapshots := true
 ThisBuild / publishMavenStyle := true
 ThisBuild / Test / publishArtifact := false
 
+
 addCommandAlias("makeSite", "docs/makeSite")
 addCommandAlias("previewSite", "docs/previewSite")
 addCommandAlias("ghpagesPushSite", "docs/ghpagesPushSite")
@@ -38,13 +39,15 @@ lazy val IntegrationTest = config("it") extend Test
 lazy val root = project
   .withId("RasterFrames")
   .aggregate(core, datasource)
-  .settings(publish / skip := true)
+  .settings(
+    publish / skip := true)
 
 lazy val `rf-notebook` = project
   .dependsOn(pyrasterframes)
   .disablePlugins(CiReleasePlugin)
   .enablePlugins(RFAssemblyPlugin, DockerPlugin)
-  .settings(publish / skip := true)
+  .settings(
+    publish / skip := true)
 
 lazy val core = project
   .enablePlugins(BuildInfoPlugin)
@@ -79,7 +82,7 @@ lazy val core = project
         ExclusionRule(organization = "com.github.mpilquist")
       ),
       scaffeine,
-      sparktestingbase excludeAll ExclusionRule("org.scala-lang.modules", "scala-xml_2.12"),
+      sparktestingbase().value % Test excludeAll ExclusionRule("org.scala-lang.modules", "scala-xml_2.12"),
       `scala-logging`
     ),
     libraryDependencies ++= {

datasource/src/test/scala/org/locationtech/rasterframes/datasource/slippy/SlippyDataSourceSpec.scala

@@ -34,7 +34,6 @@ class SlippyDataSourceSpec extends TestEnvironment with TestData with BeforeAndA
 
   def tileFilesCount(dir: File): Long = {
     val r = countFiles(dir, ".png")
-    println(dir, r)
     r
   }