Generate HTML documentation from SPARQL queries

sparql-doc provides a simple tool for generating browsable HTML documentation from SPARQL queries.

This is particularly useful when preparing training materials or resources to help developers to use a particular dataset or SPARQL endpoint.

sparql-doc supports markup for documenting sparql queries, as well as a package metadata for a collection of queries.

These features are described in the next sections.

sparql-doc processes your SPARQL queries and looks for comments. Like Javadoc, rdoc and similar tools, the content of the comments are used to provide metadata

All simple documentation lines at the start of a query will be treated as its description. E.g:

#This is a description
#of my query. It has multiple
#lines
DESCRIBE ?x

Special tag can be used to specify other metadata, such as the title of a query:

#This is a description
#of my query. It has multiple
#lines
# @title My Query Title
DESCRIBE ?x

The full list of supported tags is:

• @title: should have a single value. Last title tag wins
• @author: author(s) of the query. Can have multiple uses
• @see: add a link from the documentation
• @param: parameter. Not yet used in template
• @endpoint: suggest an endpoint for using query. Automatically adds query link to docs
• @tag: add a tag to a query. Displayed but not yet used, but this will (eventually) be used to organise queries

Here's an example that uses all these:

#This query illustrates how to describe a resource which is identified
#by matching one of its properties. In this case a work is being identified
#using its ISBN-10 value.
# @title Describe via ISSN
# @author Leigh Dodds
# @see http://www.isbn.org/
# @tag book
# @tag isbn
# @endpoint http://bnb.data.bl.uk/sparql
PREFIX bibo: <http://purl.org/ontology/bibo/>
DESCRIBE ?uri WHERE {
  ?uri bibo:isbn10 "0261102214".
}

The query description can be written in Markdown. So you can include embedded markup, e.g. links, that help to further document a query. For example:

#This query illustrates how to describe a resource which is identified
#by matching one on an [ISBN](http://www.isbn.org/)
# @title Describe via ISSN
# @author Leigh Dodds
# @tag book
# @tag isbn
# @endpoint http://bnb.data.bl.uk/sparql
PREFIX bibo: <http://purl.org/ontology/bibo/>
DESCRIBE ?uri WHERE {
  ?uri bibo:isbn10 "0261102214".
}

When processing a directory of queries, sparql-doc will automatically look for a file called overview.md. If found, this file will be automatically parsed as Markdown and its contents included in an Overview section on the homepage of the documentation.

While the description in the package.json file is intended to provide a one line summary of the package, the overview.md file is intended to provide a more detailed introduction. Both are optional, so authors can choose which approach they prefer.

sparql-doc considers a directory of SPARQL queries to be a package. Metadata that describes a package and how its documentation should be generated is provided by a valid JSON file called package.json which is found in the same directory.

The following example of a package.json file shows how to provide a title and a short description of a package of files. The title and description will automatically be injected into the documentation.

{
 "title": "BNB SPARQL Queries",
 "description": "A collection of SPARQL queries for the British National Bibliography"
}

It is common for a collection of queries to be written by the same person, be tagged in the same way, or be useful against the same collection of endpoints. Rather than repeatedly apply the @author, @tag and @endpoint annotations to all queries in a package, default values can be specified in the package.json file.

The following example shows how to do this:

{
 "title": "BNB SPARQL Queries",
 "description": "A collection of SPARQL queries for the British National Bibliography",
 "author": ["Leigh Dodds"],
 "endpoint": ["http://bnb.data.bl.uk/sparql"]
}

Note that because @author, @tag and @endpoint are all multi-valued annotations, their values must be specified as a JSON array.

The package.json file can also be used to indicate that extra files in the query directory should be processed and included in the documentation. E.g.:

{
 "title": "BNB SPARQL Queries",
 "description": "A collection of SPARQL queries for the British National Bibliography"
 "extra-files": ["more-info.md"]
}

This will trigger sparql-doc to process the more-info.md file as Markdown, converting it to more-info.html which is added to the output directory. A link to more-info will be automatically added to the header navigation

Here's the example output using the example queries included in the project.

sparql-doc is available as a gem:

[sudo] gem install sparql-doc

You'll need to make sure you have the following installed:

• Ruby 2.5.0+
• RubyGems
• Bundler
• Rake

Once you have those installed, clone the repository and run the provided rake targets to build and install the gem locally:

git clone https://github.com/ldodds/sparql-doc.git
cd sparql-doc

The code uses two gems which you'll need to have installed: JSON and Redcarpet:

bundle install

Once installed you should be able to run the bin/sparql-doc command-line tool.

Note: the command-line syntax is likely to change in future. E.g. to add more options and/or other commands

This takes two parameters:

• The input directory. The tool will process all .rq files in that directory
• The output directory. All HTML output and required assets will be placed here

E.g. you can run:

sparql-doc examples/bnb /your/output/directory

This will generate documentation from the bundled examples and place it into the specified directory.

Later versions will support additional command-line options

See the issue list. Got a suggestion? File an issue!

This work is hereby released into the Public Domain.

To view a copy of the public domain dedication, visit http://creativecommons.org/licenses/publicdomain or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.