Asciidoc to confluence converter & uploader

Motivation for this project has risen from the need to convert AsciiDoc documents with code blocks to confluence. As standard asciidoctor behavior does not let to insert code blocks in format of confluence code macro, we need to postprocess asciidoctor output.

To publish a document to confluence server its meta-parameters should be known. Parser takes these parameters from comments in input file:

// :DOCUMENT-TITLE: Document title [mandatory header]
// :DOCUMENT-TITLE-OLD: Old document title (in case if document is being renamed) [optional header]
// :DOCUMENT-PARENT: Parent page title [optional header]
// :DOCUMENT-TAGS: tag1, long tag 2, tag3 [optional header]

... and from command-line argument --space=CONFLUENCE_SPACE_KEY.

• If document with DOCUMENT-TITLE title already exists, its content/title/tags will be updated
• If document with DOCUMENT-TITLE-OLD title exists, it will be removed to DOCUMENT-TITLE
• If document with DOCUMENT-PARENT title exists, new document will be added to its children

Also confluence server credentials should be provisioned to program:

  --url=http://localhost:8090 --user=login --pass=pa$$w0rd

If confluence credentials are not passed to program, converted document(s) will be printed to STDOUT.

For debugging purposes you can print converted document onto STDOUT in case of publishing error. Use --debug argument for this.

NOTE: From now on DOCUMENT-SPACE parameter is not used to specify document's space in confluence. This means, that confluence space to work with should be provided with --space command-line argument. Which, in turn, means that all the input files should go to the same confluence space.

You can clean up space in confluence before publishing pages. For this use --clean=SPACE1,SPACE2 command-line argument with comma-separated list of space keys to be cleaned up. If this argument is passed to program all the unprotected pages in mentioned spaces will be removed before publishing stage. Only pages tagged with protected label(s) will be kept. Protected labels are set in configuration file and can be overridden with environment variable PROTECTED_LABELS. Like this:

export PROTECTED_LABELS=protected_label_1,protected_label_2

Also you can forcefully clean up confluence space(s), removing all the pages (even protected). For this use --force command line argument.

The data for program can be provisioned with two ways:

• single file (--input=<fileName>)

  In this case only one file will be processed.

• directory (--dir=<directory>)

  In this case directory will be processed recursively. All the *.adoc and *.asciidoc files will be converted (and, if related arguments are set, published to confluence server). Subdirectories are being processed in such a way that at first all files from a directory are being processed, then all the subdirectories are being processed. This means that all the higher-level pages should be placed in higher-level directories, so that parent pages would be created before related siblings.

If --dir and --input arguments are used together, only --dir key will be considered.

If DOCUMENT-TAGS header is set for document, related page on confluence will be labeled with configured tags. Tags should be separated with commas. Spaces in multi-word tags will be converted to underscores ('_'), as confluence does not support multi-word labels (see here).

After being published to confluence server some documents can be later removed from document repository. Such documents are found during directory processing and removed from confluence server.

(Program gets a list of all the documents in local repository and all the documents from confluence server [for given space]. All the documents found on a server and not found in local repository will be removed from server.)

For testing purposes or during documentation preparation you may need to prevent existing document from being published to confluence server. For this you can use DOCUMENT-HIDDEN header like this:

// prevents document from publication
// :DOCUMENT-HIDDEN:    

// prevents document from publication
// :DOCUMENT-HIDDEN: true    

// allows document publication
// :DOCUMENT-HIDDEN: false

If document is already published and then marked with DOCUMENT-HIDDEN flag, it will be removed from server.

Program tries to convert links to native confluence format. The links in source AsciiDoc can be of the following formats:

• http://link.to.some.internet.site[link display text]

  In this case ordinary link to external resource will be added to confluence page.

• link:CONFLUENCESPACE:Page+title[link display text]

  In this case there will be created confluence link, pointing to page titled Page title in space CONFLUENCESPACE on the same confluence server.

• link:Page+title[link display text]

  In this case there will be created confluence link, pointing to page titled Page title in the same space, which the page being processed belongs to.

NOTE: When creating links in link: format, spaces in page title should be replaces with '+' symbols in source AsciiDoc.

To insert TOC macro in resulting confluence page use following syntax:

// :DOCUMENT-SPACE-KEY: TESTSPACE
// :DOCUMENT-TITLE: Main page

= Title
:hardbreaks:
:toc:

{zwsp}

... the rest of document ...

You can insert confluence page tree macro into source asciidoc files to be rendered to confluence page tree. Use following syntax:

// pagetree::Root+page+name[]

You can insert confluence children display macro into source asciidoc files to be rendered to confluence children display. Use following syntax:

// children::Parent+page+name[]

Formula support is implemented with help of MathJax plugin. Formulas can be inline \(\sum_{i=0}^n i^2 = \frac{(n^2+n)(2n+1)}{6}\) or block \[\sum_{i=0}^n i^2 = \frac{(n^2+n)(2n+1)}{6}\]. By default block-formulas are aligned to center. To align them left or right you need to add following definitions to global Confluence CSS:

.math-left .mjx-chtml.MJXc-display {
    text-align: left !important;
 }
.math-right .mjx-chtml.MJXc-display {
    text-align: right !important;
 }

Then you can use \l or \r in start of formula definition in source asciidoc file: \[\l\sum_{i=0}^n i^2 = \frac{(n^2+n)(2n+1)}{6}\] or \[\r\sum_{i=0}^n i^2 = \frac{(n^2+n)(2n+1)}{6}\]