This is a simple tool, which transforms your sources into markdown documentations for themselves. It just reads a source file, turns block comments into normal text and surrounds code with markdown backticks syntax. So the aim is just to get a readable document from a code, which is written in more or less literate programming style. The name is like "a thing which makes your sources literate", i.e. helps to use literate programming when it's not really supported by the language.
So you can write your code and use markdown syntax in comments (which keeps your sources readable), and then transform it to a markdown document, from which you can generate a nice html or pdf or whatever else, using your favorite markdown processor.
The tool is written in Scala and first of all for Scala, because it doesn't have normal support for literate programming. But it should work for [some other languages][lib/LanguageMap] โ randomly chosen the most popular, of course. Open an issue, if you want support for something else.
Note: Before
v0.7.0
there was a separate library, command line application and an sbt-plugin.
Some nice features:
- it provides convenient list of links definitions for internal references
- it (optionally) generates index of the files in the given directory and appends it to each produced markdown file
Just add following line to your project/plugins.sbt
:
addSbtPlugin("laughedelic" % "literator" % "<version>")
You can find the latest version in the list of releases.
Key | Type | Default |
---|---|---|
docsMap |
Map[File, File] |
src/ โ docs/src/ |
docsOutputDirs |
Seq[File] |
same as docsMap values |
docsAddIndex |
Boolean |
false |
docsCleanBefore |
Boolean |
true |
You can set docsMap
key in your build.sbt
to map source directories to output documentation directories. Using this map you can easily generate docs for several subprojects or integrate certain parts of the source documentation in the general docs (especially if you're using some service like Read the Docs).
To run Literator from sbt, use generateDocs
task.
Note that output directories are cleaned up before generating docs (you can turn it off with docsCleanBefore := false
), so be careful if you're mixing generated docs with handwritten.
If you use sbt-release plugin, you can add docs generation step. See sbt-release documentation for details.
If you have a piece of code like this (it's from the literator sources):
/* ### Block comments parsing */
/* - When parsing the comment opening brace, we remember the offset for the content
Note that [scaladoc-style comments](http://docs.scala-lang.org/style/scaladoc.html)
are ignored.
*/
def commentStart: PS =
spaces ~ (lang.comment.start ^^ { _.replaceAll(".", " ") }) ~
(((spaces ~ guard(eol)) ^^^ None) | // if the first row is empty, ignore it
(guard(not("*")) ~> " ".?)) ^^ // not scaladoc and an optional space
{ case sps ~ strt ~ sp => sps + strt + sp.getOrElse("") }
/* - Closing comment brace is just ignored */
def commentEnd: PS = spaces ~ lang.comment.end ^^^ ""
Then it will be transformed into the following markdown:
- When parsing the comment opening brace, we remember the offset for the content Note that scaladoc-style comments are ignored.
def commentStart: PS =
spaces ~ (lang.comment.start ^^ { _.replaceAll(".", " ") }) ~
(((spaces ~ guard(eol)) ^^^ None) | // if the first row is empty, ignore it
(guard(not("*")) ~> " ".?)) ^^ // not scaladoc and an optional space
{ case sps ~ strt ~ sp => sps + strt + sp.getOrElse("") }
- Closing comment brace is just ignored
def commentEnd: PS = spaces ~ lang.comment.end ^^^ ""
You can see the result of running Literator on its own sources in the docs/src/ folder.
Yes, don't look that there were no recent commits, it's just quite stable. In @ohnosequences we use this tool in everyday development and as a part of the release process.
Of course, there are plenty of docco-like tools, which generate htmls from your sources (also using markdown), but there are several reasons, why I don't like them:
- there is no normal Scala-clone of such tool that you could integrate in the release process
- most of such tools support only one-line comments and ignore block comments, while the opposite makes more sense: write documentation comments as normal text and keep small technical comments inlined
- those tools claim to be "quick and dirty", IMO better to have something simple, but reliable
- Literator helps to keep things simple using markdown which is good as an intermediate format. For example, it's handy to have just markdown documents on github, as it will render them nicely and then generate from them htmls for a web-site if needed using your favorite tool and style templates
There are a couple of interesting projects which approach literate programming in Scala from another side:
- tut reads markdown files and interprets Scala code in
tut
sheds - sbt-scaliterate generates Scala source code from a programming book written in markdown