Compile Sedona source code¶
Compile Scala / Java source code¶
Sedona Scala/Java code is a project with multiple modules. Each module is a Scala/Java mixed project which is managed by Apache Maven 3.
- Make sure your Linux/Mac machine has Java 11/17, Apache Maven 3.3.1+, and Python3.8+. The compilation of Sedona is not tested on Windows machines.
To compile all modules, please make sure you are in the root folder of all modules. Then enter the following command in the terminal:
mvn clean install -DskipTests
mvn clean install
mvn clean install -DskipTests -Dgeotools
Note
By default, this command will compile Sedona with Spark 3.4 and Scala 2.12
Compile with different targets¶
User can specify -Dspark and -Dscala command line options to compile with different targets. Available targets are:
-Dspark:{major}.{minor}: For example, specify-Dspark=3.4to build for Spark 3.4.-Dscala:2.12or2.13
mvn clean install -DskipTests -Dspark=3.4 -Dscala=2.12
3.4 with Spark major.minor version when building for higher Spark versions.
mvn clean install -DskipTests -Dspark=3.4 -Dscala=2.13
3.4 with Spark major.minor version when building for higher Spark versions.
Tip
To get the Sedona Spark Shaded jar with all GeoTools jars included, simply append -Dgeotools option. The command is like this:mvn clean install -DskipTests -Dscala=2.12 -Dspark=3.4 -Dgeotools
Download staged jars¶
Sedona uses GitHub Actions to automatically generate jars per commit. You can go here and download the jars by clicking the commits Artifacts tag.
Run Python test¶
1) Set up the environment variable SPARK_HOME and PYTHONPATH
For example,
export SPARK_VERSION=3.4.0
export SPARK_HOME=$PWD/spark-${SPARK_VERSION}-bin-hadoop3
export PYTHONPATH=$SPARK_HOME/python
2) Install Spark if you haven't already
wget https://archive.apache.org/dist/spark/spark-${SPARK_VERSION}/spark-${SPARK_VERSION}-bin-hadoop3.tgz
tar -xvzf spark-${SPARK_VERSION}-bin-hadoop3.tgz
rm spark-${SPARK_VERSION}-bin-hadoop3.tgz
3) Put JAI jars to SPARK_HOME/jars/ folder.
export JAI_CORE_VERSION="1.1.3"
export JAI_CODEC_VERSION="1.1.3"
export JAI_IMAGEIO_VERSION="1.1"
wget -P $SPARK_HOME/jars/ https://repo.osgeo.org/repository/release/javax/media/jai_core/${JAI_CORE_VERSION}/jai_core-${JAI_CORE_VERSION}.jar
wget -P $SPARK_HOME/jars/ https://repo.osgeo.org/repository/release/javax/media/jai_codec/${JAI_CODEC_VERSION}/jai_codec-${JAI_CODEC_VERSION}.jar
wget -P $SPARK_HOME/jars/ https://repo.osgeo.org/repository/release/javax/media/jai_imageio/${JAI_IMAGEIO_VERSION}/jai_imageio-${JAI_IMAGEIO_VERSION}.jar
4) Compile the Sedona Scala and Java code with -Dgeotools and then copy the sedona-spark-shaded-1.8.0.jar to SPARK_HOME/jars/ folder.
cp spark-shaded/target/sedona-spark-shaded-*.jar $SPARK_HOME/jars/
5) Install the following libraries
sudo apt-get -y install python3-pip python-dev libgeos-dev
sudo pip3 install -U setuptools
sudo pip3 install -U wheel
sudo pip3 install -U virtualenvwrapper
sudo pip3 install -U pipenv
Homebrew can be used to install libgeos-dev in macOS:
brew install geos
6) Set up pipenv to the desired Python version: 3.8, 3.9, or 3.10
cd python
pipenv --python 3.8
7) Install the PySpark version and the other dependency
cd python
pipenv install pyspark==${SPARK_VERSION}
pipenv install --dev
pipenv install pyspark installs the latest version of pyspark.
In order to remain consistent with the installed spark version, use pipenv install pyspark==<spark_version>
8) Run the Python tests
cd python
pipenv run python setup.py build_ext --inplace
pipenv run pytest tests
Compile the documentation¶
The website is automatically built after each commit. The built website can be downloaded here:
MkDocs website¶
The source code of the documentation website is written in Markdown and then compiled by MkDocs. The website is built upon the Material for MkDocs template.
In the Sedona repository, the MkDocs configuration file mkdocs.yml is in the root folder and all documentation source code is in docs folder.
To compile the source code and test the website on your local machine, please read the MkDocs Tutorial and Materials for MkDocs Tutorial.
In short, you need to run:
pip install -r requirements-docs.txt
After installing MkDocs and MkDocs-Material, run these commands in the Sedona root folder:
mkdocs build
mike deploy --update-aliases latest-snapshot -b website -p
mike serve
pre-commit¶
We run pre-commit with GitHub Actions so installation on your local machine is currently optional.
The pre-commit configuration file is in the repository root. Before you can run the hooks, you need to have pre-commit installed.
The hooks run when running git commit and also from the command line with pre-commit. Some of the hooks will auto
fix the code after the hooks fail whilst most will print error messages from the linters. If a hook fails the overall
commit will fail, and you will need to fix the issues or problems and git add and git commit again. On git commit
the hooks will run mostly only against modified files so if you want to test all hooks against all files and when you
are adding a new hook you should always run:
pre-commit run --all-files
Sometimes you might need to skip a hook to commit because the hook is stopping you from committing or your computer
might not have all the installation requirements for all the hooks. The SKIP variable is comma separated for two or
more hooks:
SKIP=codespell git commit -m "foo"
The same applies when running pre-commit:
SKIP=codespell pre-commit run --all-files
Occasionally you can have more serious problems when using pre-commit with git commit. You can use --no-verify to
commit and stop pre-commit from checking the hooks. For example:
git commit --no-verify -m "foo"
If you just want to run one hook for example just run the markdownlint hook:
pre-commit run markdownlint --all-files
We have a Makefile in the repository root which has three pre-commit convenience commands.
For example, you can run the following to setup pre-commit to run before each commit
make checkinstall