<p>Welcome to the Spark documentation!</p>

<p>This readme will walk you through navigating and building the Spark documentation, which is included
here with the Spark source code. You can also find documentation specific to release versions of
Spark at https://spark.apache.org/documentation.html.</p>

<p>Read on to learn more about viewing documentation in plain text (i.e., markdown) or building the
documentation yourself. Why build it yourself? So that you have the docs that correspond to
whichever version of Spark you currently have checked out of revision control.</p>

<h2 id="prerequisites">Prerequisites</h2>

<p>The Spark documentation build uses a number of tools to build HTML docs and API docs in Scala, Java,
Python, R and SQL.</p>

<p>You need to have <a href="https://www.ruby-lang.org/en/documentation/installation/">Ruby</a> and
<a href="https://docs.python.org/2/using/unix.html#getting-and-installing-the-latest-version-of-python">Python</a>
installed. Also install the following libraries:</p>

<div class="language-sh highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="nv">$ </span><span class="nb">sudo </span>gem <span class="nb">install </span>jekyll jekyll-redirect-from rouge
<span class="c"># Following is needed only for generating API docs</span>
<span class="nv">$ </span><span class="nb">sudo </span>pip <span class="nb">install </span>sphinx pypandoc mkdocs
<span class="nv">$ </span><span class="nb">sudo </span>Rscript <span class="nt">-e</span> <span class="s1">'install.packages(c("knitr", "devtools", "rmarkdown"), repos="https://cloud.r-project.org/")'</span>
<span class="nv">$ </span><span class="nb">sudo </span>Rscript <span class="nt">-e</span> <span class="s1">'devtools::install_version("roxygen2", version = "5.0.1", repos="https://cloud.r-project.org/")'</span>
<span class="nv">$ </span><span class="nb">sudo </span>Rscript <span class="nt">-e</span> <span class="s1">'devtools::install_version("testthat", version = "1.0.2", repos="https://cloud.r-project.org/")'</span>
</code></pre></div></div>

<p>Note: If you are on a system with both Ruby 1.9 and Ruby 2.0 you may need to replace gem with gem2.0.</p>

<p>Note: Other versions of roxygen2 might work in SparkR documentation generation but <code class="highlighter-rouge">RoxygenNote</code> field in <code class="highlighter-rouge">$SPARK_HOME/R/pkg/DESCRIPTION</code> is 5.0.1, which is updated if the version is mismatched.</p>

<h2 id="generating-the-documentation-html">Generating the Documentation HTML</h2>

<p>We include the Spark documentation as part of the source (as opposed to using a hosted wiki, such as
the github wiki, as the definitive documentation) to enable the documentation to evolve along with
the source code and be captured by revision control (currently git). This way the code automatically
includes the version of the documentation that is relevant regardless of which version or release
you have checked out or downloaded.</p>

<p>In this directory you will find text files formatted using Markdown, with an &#8220;.md&#8221; suffix. You can
read those text files directly if you want. Start with <code class="highlighter-rouge">index.md</code>.</p>

<p>Execute <code class="highlighter-rouge">jekyll build</code> from the <code class="highlighter-rouge">docs/</code> directory to compile the site. Compiling the site with
Jekyll will create a directory called <code class="highlighter-rouge">_site</code> containing <code class="highlighter-rouge">index.html</code> as well as the rest of the
compiled files.</p>

<div class="language-sh highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="nv">$ </span><span class="nb">cd </span>docs
<span class="nv">$ </span>jekyll build
</code></pre></div></div>

<p>You can modify the default Jekyll build as follows:</p>

<div class="language-sh highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="c"># Skip generating API docs (which takes a while)</span>
<span class="nv">$ SKIP_API</span><span class="o">=</span>1 jekyll build

<span class="c"># Serve content locally on port 4000</span>
<span class="nv">$ </span>jekyll serve <span class="nt">--watch</span>

<span class="c"># Build the site with extra features used on the live page</span>
<span class="nv">$ PRODUCTION</span><span class="o">=</span>1 jekyll build
</code></pre></div></div>

<h2 id="api-docs-scaladoc-javadoc-sphinx-roxygen2-mkdocs">API Docs (Scaladoc, Javadoc, Sphinx, roxygen2, MkDocs)</h2>

<p>You can build just the Spark scaladoc and javadoc by running <code class="highlighter-rouge">./build/sbt unidoc</code> from the <code class="highlighter-rouge">$SPARK_HOME</code> directory.</p>

<p>Similarly, you can build just the PySpark docs by running <code class="highlighter-rouge">make html</code> from the
<code class="highlighter-rouge">$SPARK_HOME/python/docs</code> directory. Documentation is only generated for classes that are listed as
public in <code class="highlighter-rouge">__init__.py</code>. The SparkR docs can be built by running <code class="highlighter-rouge">$SPARK_HOME/R/create-docs.sh</code>, and
the SQL docs can be built by running <code class="highlighter-rouge">$SPARK_HOME/sql/create-docs.sh</code>
after <a href="https://github.com/apache/spark#building-spark">building Spark</a> first.</p>

<p>When you run <code class="highlighter-rouge">jekyll build</code> in the <code class="highlighter-rouge">docs</code> directory, it will also copy over the scaladoc and javadoc for the various
Spark subprojects into the <code class="highlighter-rouge">docs</code> directory (and then also into the <code class="highlighter-rouge">_site</code> directory). We use a
jekyll plugin to run <code class="highlighter-rouge">./build/sbt unidoc</code> before building the site so if you haven&#8217;t run it (recently) it
may take some time as it generates all of the scaladoc and javadoc using <a href="https://github.com/sbt/sbt-unidoc">Unidoc</a>.
The jekyll plugin also generates the PySpark docs using <a href="http://sphinx-doc.org/">Sphinx</a>, SparkR docs
using <a href="https://cran.r-project.org/web/packages/roxygen2/index.html">roxygen2</a> and SQL docs
using <a href="https://www.mkdocs.org/">MkDocs</a>.</p>

<p>NOTE: To skip the step of building and copying over the Scala, Java, Python, R and SQL API docs, run <code class="highlighter-rouge">SKIP_API=1
jekyll build</code>. In addition, <code class="highlighter-rouge">SKIP_SCALADOC=1</code>, <code class="highlighter-rouge">SKIP_PYTHONDOC=1</code>, <code class="highlighter-rouge">SKIP_RDOC=1</code> and <code class="highlighter-rouge">SKIP_SQLDOC=1</code> can be used
to skip a single step of the corresponding language. <code class="highlighter-rouge">SKIP_SCALADOC</code> indicates skipping both the Scala and Java docs.</p>

<h3 id="automatically-rebuilding-api-docs">Automatically Rebuilding API Docs</h3>

<p><code class="highlighter-rouge">jekyll serve --watch</code> will only watch what&#8217;s in <code class="highlighter-rouge">docs/</code>, and it won&#8217;t follow symlinks. That means it won&#8217;t monitor your API docs under <code class="highlighter-rouge">python/docs</code> or elsewhere.</p>

<p>To work around this limitation for Python, install <a href="http://eradman.com/entrproject/"><code class="highlighter-rouge">entr</code></a> and run the following in a separate shell:</p>

<div class="language-sh highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="nb">cd</span> <span class="s2">"</span><span class="nv">$SPARK_HOME</span><span class="s2">/python/docs"</span>
find .. <span class="nt">-type</span> f <span class="nt">-name</span> <span class="s1">'*.py'</span> <span class="se">\</span>
| entr <span class="nt">-s</span> <span class="s1">'make html &amp;&amp; cp -r _build/html/. ../../docs/api/python'</span>
</code></pre></div></div>

<p>Whenever there is a change to your Python code, <code class="highlighter-rouge">entr</code> will automatically rebuild the Python API docs and copy them to <code class="highlighter-rouge">docs/</code>, thus triggering a Jekyll update.</p>