Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/madhead/doktor

Jenkins plugin for automated documentation uploading to Confluence [Mirror]
https://github.com/madhead/doktor

confluence gradle jenkins jenkins-plugin kotlin

Last synced: 4 days ago
JSON representation

Jenkins plugin for automated documentation uploading to Confluence [Mirror]

Lists

README 

        
image:https://travis-ci.org/jenkinsci/doktor-plugin.svg?branch=master["Build Status", link="https://travis-ci.org/jenkinsci/doktor-plugin"]

image:https://codecov.io/gh/jenkinsci/doktor-plugin/branch/master/graph/badge.svg["Build Status", link="https://codecov.io/gh/jenkinsci/doktor-plugin"]



= Doktor



Doktor is a Jenkins plugin for automated documentation uploading to Confluence.



== Usage



=== Features



==== Supported formats



Doktor supports https://daringfireball.net/projects/markdown/syntax[Markdown] and http://asciidoc.org[AsciiDoc].

Markdown support is provided by awesome https://github.com/vsch/flexmark-java[flexmark-java] library.

AsciiDoc is supported thanks to https://github.com/asciidoctor/asciidoctorj[AsciidoctorJ].

Please note that AsciiDoc support is very experimental.



==== Front matter



_Front matter_ is another word for _metadata_, used by bloggers and hipsters widely.

Doktor uses front matter to configure the way pages appear in Confluence.



===== Markdown



Doktor supports http://www.yaml.org[YAML] front matter in your Markdown files.

Front matter looks like a small YAML fragment at the beginning of the file, separated by a triple minus sign (`---`) in this case:



[source,yml]

----

---

key: value

---

----



===== AsciiDoc



Doktor supports http://www.yaml.org[YAML] front matter in your AsciiDoc files as well.

Front matter looks like a small YAML fragment at the beginning of the file, separated by a triple minus sign (`---`) in this case.

Note, that due to a more strict YAML parser logic for AsciiDoc, strings with special characters need to be quoted:



[source, asciidoc]

----

---

key: value

key_with_specials: 'value: with specials'

---

----



===== Supported front matter attributes:



`title`::

Required.

Page title, unsurprisingly.



`parent`::

Optional.

Parent page title, if any.

If omitted, https://confluence.atlassian.com/doc/orphaned-pages-139542.html["orphaned" page] will be created.

If parent page is not found by title, child page will not be created at all.



`labels`::

Optional.

List of labels to add to a page.



==== Tables



===== Markdown



Though Markdown does not have any support for tables, Doktor supports https://help.github.com/articles/organizing-information-with-tables[GitHub Flavored Markdown tables].

You can also create tables by inlining XHTML markup directly in your docs.



===== AsciiDoc



AsciiDoc (thus Asciidoctor and AsciidoctorJ) http://asciidoctor.org/docs/user-manual/#tables[supports tables natively].



==== Images



Doktor supports images.

When an image is referred by relative URL it will be uploaded to a Confluence server as an attachment of a page, given unique name.

When an image is referred by remote URL (Internet link) it will be referred by this URL from a Confluence server.



===== Markdown



https://daringfireball.net/projects/markdown/syntax#img[Markdown syntax] for images:



[source, markdown]

----

![Millennium Falcon](./millennium_falcon.png "The Millennium Falcon, Han Solo's most prized possession")

----



===== AsciiDoc



Images look like http://asciidoctor.org/docs/asciidoc-writers-guide/#images[this] is AsciiDoc:



[source, asciidoc]

----

.The Millennium Falcon, Han Solo's most prized possession

[link=http://starwars.wikia.com/wiki/Millennium_Falcon]

image::./millennium_falcon.png[Millennium Falcon,400,float="right",align="center"]

----



As you see, AsciiDoc is more feature-rich.



==== Diagrams



Diagrams are only supported in AsciiDoc, the markup looks like this:



[seqdiag]

....

seqdiag {

  // normal edge and doted edge

  A -> B [label = "normal edge"];

  B --> C [label = "dotted edge"];



  B <-- C [label = "return dotted edge"];

  A <- B [label = "return edge"];



  // asynchronus edge

  A ->> B [label = "asynchronus edge"];

  B -->> C [label = "asynchronus dotted edge"];



  B <<-- C [label = "return asynchronus doted edge"];

  A <<- B [label = "return asynchronus edge"];



  // self referenced edge

  A -> A [label = "self reference edge"];

}

....



The snippet above will be rendered in this image:



image::https://github.com/madhead/doktor/blob/master/.github/images/seqdiag.png[]



Read more about diagram syntax in http://asciidoctor.org/docs/asciidoctor-diagram[in the official AsciiDoc guide].

Be warned, that most types of diagrams require external tools (like `seqdiag` or `dot`) to be installed and available on the `PATH`.



Currently, these diagrams are supported:



 - `actdiag` / `blockdiag` / `nwdiag` / `packetdiag` / `rackdiag` / `seqdiag`.

These diagrams require http://blockdiag.com[`blockdiag`] and related Python packages to be available on the `PATH`.

 - http://ditaa.sourceforge.net[`ditaa`].

No additional tools needed.

 - `graphviz`.

Obviously, requires http://www.graphviz.org[Graphviz] tool to be on the `PATH`.

 - `mermaid`

Requires https://mermaidjs.github.io[mermaid] (version prior to `7.x`) and http://phantomjs.org[PhantomJS] to be on the `PATH`.

 - http://plantuml.com[`plantuml`]

No additional tools needed.



==== Custom stylesheets



Confluence allows space admins to provide custom stylesheets that override globals.

Doktor supports styling generated content by wrapping it in a `
`, so you can use `.doctor` prefix in your selector to stylize content.



=== Configure Confluence servers



As you might suspect, Confluence REST API requires authentication.

Doktor supports basic authentication (username and password).

So, first thing to do is to https://github.com/jenkinsci/credentials-plugin/blob/master/docs/user.adoc[configure credentials] in Jenkins.



Create a "Username with password" credentials to be used to authenticate on Confluence server:



image::https://github.com/madhead/doktor/blob/master/.github/images/new_credentials.png[]



You may have as many Confluence servers and credentials for them as you need.



Next thing to do is to configure Confluence servers.

Go to global configuration screen ("Manage Jenkins" -> "Configure System") and find "Confluence Servers" section.

Configure the list of available Confluence servers:



image::https://github.com/madhead/doktor/blob/master/.github/images/confluence_servers.png[]



Now, when you have some Confluence servers to publish documentation to, it's time test this plugin!

Yes, I'm using word "test" https://github.com/madhead/doktor/issues/new[intentionally] here.



=== Pipeline step



Using Doktor with https://jenkins.io/doc/book/pipeline[pipelines] is very easy!

Here is the full syntax of `doktor` step:



[source,groovy]

----

doktor

	server : 'Cantina', // <1>

	markdownIncludePatterns: ['glob:**.md'], // <2>

	markdownExcludePatterns: ['glob:README.md'], // <3>

	asciidocIncludePatterns: ['glob:**.adoc', 'glob:**.asc'], // <4>

	asciidocExcludePatterns: ['glob:LICENSE.adoc', 'glob:CONTRIBUTING.asc'] // <5>

----

<1> One of the available Confluence servers

<2> List of Java 8 https://docs.oracle.com/javase/8/docs/api/java/nio/file/FileSystem.html#getPathMatcher-java.lang.String-[PathMatcher specifications] for https://daringfireball.net/projects/markdown/syntax[Markdown] files to include.

<3> List of Java 8 https://docs.oracle.com/javase/8/docs/api/java/nio/file/FileSystem.html#getPathMatcher-java.lang.String-[PathMatcher specifications] for https://daringfireball.net/projects/markdown/syntax[Markdown] files to exclude.

<4> List of Java 8 https://docs.oracle.com/javase/8/docs/api/java/nio/file/FileSystem.html#getPathMatcher-java.lang.String-[PathMatcher specifications] for http://asciidoc.org[AsciiDoc] files to include.

<5> List of Java 8 https://docs.oracle.com/javase/8/docs/api/java/nio/file/FileSystem.html#getPathMatcher-java.lang.String-[PathMatcher specifications] for http://asciidoc.org[AsciiDoc] files to exclude.



You can also try your luck with "Snippet Generator", available at `/pipeline-syntax` path of your Jenkins installation.



=== Classic builds



Doktor plays nice with "classic" builds too!



image::https://github.com/madhead/doktor/blob/master/.github/images/freestyle_config.png[]



Click those question icons on the right if you need any help.



== Limitations



Doktor recreates pages instead of updating them.

Recreating pages has some counterintuitive effects:



 - Any modification will overridden on each Doktor run, obviously

 - Page likes are not preserved

 - Attachments are not preserved

 - There is no support for extra Confluence markup, macroses and features like comments



This may sound shocking to you, but let me explain.



Doktor's idea is just uploading your documentation somewhere, making it available to _read_ by everybody.

Doktor is not about collaborative editing - use VCS for that.

It's a unidirectional flow - from sources to rendered documents - by design.

I was inspired by GitHub's https://pages.github.com[pages] and https://help.github.com/articles/about-github-wikis[wikis], and I sincerely believe in this approach.



At the moment, Doktor supports only Confluence and may never support any other services (unless my employer switches to another vendor).



== Developing



Doktor is built with https://kotlinlang.org[Kotlin], https://gradle.org[Gradle] and Love.

Well, actually with hate to the workflows on my day-time job.



JPI artifact is produced with https://github.com/jenkinsci/gradle-jpi-plugin[Gradle's JPI plugin].

Read its documentation to know more about supported features and options.



Also, take a look at https://github.com/SimpleFinance/jenkins-firebase-test-plugin[this awesome Jenkins plugin], which is build with Gradle and Kotlin too!



=== Building & running



Basically, `./gradlew --rerun-tasks clean jpi server` will spin up a Jenkins with Doktor installed.

`--rerun-tasks` is used to force clean build every time because Gradle aggressively caches build outputs, especially https://kotlinlang.org/docs/reference/kapt.html[Kotlin annotation processing tool] results.

Feel free to tweak CLI arguments, assuming you know what you do.



Debug is supported as well:



[source, bash]

----

GRADLE_OPTS="-agentlib:jdwp=transport=dt_socket,server=y,suspend=y,address=5005" ./gradlew --rerun-tasks clean jpi server

----



Omit `server` task if you just need a JPI file.



=== Testing on remote agents



Once you may want to test how Doktor behaves on agents.

The simplest way to do that is to run an agent in Docker.

There are two images for agents available.



==== jenkinsci/slave



https://hub.docker.com/r/jenkinsci/slave[jenkinsci/slave] is an image meant to be run by Jenkins to start a new agent.

The configuration is very simple:



image::https://github.com/madhead/doktor/blob/master/.github/images/slave.png[]



When you're running Jenkins via Gradle JPI plugin it will be run under you user account, so either your user needs to be able to execute `sudo docker` without password or you will need to type that password in Gradle's terminal session.



==== jenkinsci/ssh-slave



https://hub.docker.com/r/jenkinsci/ssh-slave[jenkinsci/ssh-slave] is another (better) option.

It allows you manage agent container separately and then attach it to Jenkins, thus eliminating the need to provide any password or execute `sudo docker`.

Container's mounts and FS modifications will be preserved between Jenkins restarts.



First, you need to have an SSH key pair that will be used to connect to the agent.

Looks like only RSA keys are supported (public key must start with `ssh-` prefix).

Either https://help.github.com/articles/generating-a-new-ssh-key-and-adding-it-to-the-ssh-agent[create a new one], or use the existing.



Then, install https://wiki.jenkins.io/display/JENKINS/SSH+Slaves+plugin[SSH Slaves plugin] on the master.



Create new "SSH Username with private key" credentials:



image::https://github.com/madhead/doktor/blob/master/.github/images/ssh_slave_credentials.png[]



You can paste private key directly here or use one of the defaults (`~/.ssh/id_ecdsa`, `~/.ssh/id_rsa`, `~/.ssh/id_dsa`, `~/.ssh/identity`).



Next, start agent container by executing `docker run --detach --name jenkins-slave jenkinsci/ssh-slave "$(cat ~/.ssh/jenkins.pub)"` (assuming that `~/.ssh/jenkins.pub` is a public key corresponding to the private key from previous step).



Finally, create new agent with a configuration like this:



image::https://github.com/madhead/doktor/blob/master/.github/images/ssh_slave.png[]



`172.17.0.2` here is the IP of a Docker contaner from the previous step, https://stackoverflow.com/a/20686101/750510[found in `docker inspect` output].

You could also run the container exposing the ports (e.g. `-p 2222:22`) and then use `localhost` as host and `2222` as port.



=== Testing Confluence integration



You'll need to refer to Confluence REST API.

https://docs.atlassian.com/confluence/REST/latest[Here] is the link.

https://developer.atlassian.com/confdev/confluence-server-rest-api/confluence-rest-api-examples[Samples] are also available.



==== Cloud



Probably, the easiest (and CPU / RAM saving) way to run Confluence is to run it in the cloud (AWS EC2, DigitalOcean, ...).

Though, it will cost you some money.



There is an link:.ansible/confluence.yml[Ansible script] in this repo to automate Confluence installation.

It assumes that you already have a running instance that meets https://confluence.atlassian.com/doc/system-requirements-126517514.html[Confluence's minimal system requirements].

Read your cloud provider's documentation to know how to create and manageVMs.



When you have a VM, just follow these steps to install Confluence Server:



. Create inventory file (`.ansible/inventory`) with a content like this:

+

[source, ini]

----

[confluence]

your.confluence.host

----

+

You might want to add additional parameters.

For example, a set of parameters for Ubuntu 16.04 EC2 instance:

+

[source, ini]

----

[confluence]

your.confluence.host ansible_user=ubuntu ansible_ssh_private_key_file=~/.ssh/confluence.pem ansible_python_interpreter=/usr/bin/python3

----

+

Or you can just use http://docs.ansible.com/ansible/latest/intro_dynamic_inventory.html[dynamic inventories].



. Install required roles from https://galaxy.ansible.com[Ansible Galaxy]: `sudo ansible-galaxy install -r requirements.yml --force`.



. After the inventory is configured, just run `./confluence.yml` from the `.ansible` directory.



. Go to `http://your.confluence.host/` (if the DNS and IPs are set) and configure the instance.

Note, that you will need a license key (trial works for 90 days).



==== Docker



You can run Confluence locally as well.

The easiest way here is https://www.docker.com[Docker] (Windows uses should appreciate the joke).



Running Confluence is as simple as:



[source, bash]

----

docker volume create --name confluence-data

docker run --detach --volume confluence-data:/var/atlassian/application-data/confluence --name confluence --publish-all atlassian/confluence-server:latest

----



You might want to add some https://docs.docker.com/engine/reference/run[additional options] or tweak the existing ones.



Note, that you will need a license key (trial works for 90 days).