Pelican internals

This section describe how pelican is working internally. As you’ll see, it’s quite simple, but a bit of documentation doesn’t hurt :)

Overall structure

What pelican does, is taking a list of files, and processing them, to some sort of output. Usually, the files are restructured text and markdown files, and the output is a blog, but it can be anything you want.

I’ve separated the logic in different classes and concepts:

  • writers are responsible of all the writing process of the files. It’s writing .html files, RSS feeds and so on. Since those operations are commonly used, the object is created once, and then passed to the generators.
  • readers are used to read from various formats (Markdown, and Restructured Text for now, but the system is extensible). Given a file, they return metadata (author, tags, category etc) and content (HTML formated)
  • generators generate the different outputs. For instance, pelican comes with ArticlesGenerator and PageGenerator, into others. Given a configurations, they can do whatever they want. Most of the time it’s generating files from inputs.
  • pelican also uses templates, so it’s easy to write you own theme. The syntax is jinja2, and, trust me, really easy to learn, so don’t hesitate a second.

How to implement a new reader ?

There is an awesome markup language you want to add to pelican ? Well, the only thing you have to do is to create a class that have a read method, that is returning an HTML content and some metadata.

Take a look to the Markdown reader:

class MarkdownReader(Reader):
    enabled = bool(Markdown)

    def read(self, filename):
        """Parse content and metadata of markdown files"""
        text = open(filename)
        md = Markdown(extensions = ['meta', 'codehilite'])
        content = md.convert(text)

        metadata = {}
        for name, value in md.Meta.items():
            if name in _METADATA_FIELDS:
                meta = _METADATA_FIELDS[name](value[0])
                meta = value[0]
            metadata[name.lower()] = meta
        return content, metadata

Simple isn’t it ?

If your new reader requires additional Python dependencies then you should wrap their import statements in try...except. Then inside the reader’s class set the enabled class attribute to mark import success or failure. This makes it possible for users to continue using their favourite markup method without needing to install modules for all the additional formats they don’t use.

How to implement a new generator ?

Generators have basically two important methods. You’re not forced to create both, only the existing ones will be called.

  • generate_context, that is called in a first place, for all the generators. Do whatever you have to do, and update the global context if needed. This context is shared between all generators, and will be passed to the templates. For instance, the PageGenerator generate_context method find all the pages, transform them into objects, and populate the context with them. Be careful to not output anything using this context at this stage, as it is likely to change by the effect of others generators.
  • generate_output is then called. And guess what is it made for ? Oh, generating the output :) That’s here that you may want to look at the context and call the methods of the writer object, that is passed at the first argument of this function. In the PageGenerator example, this method will look at all the pages recorded in the global context, and output a file on the disk (using the writer method write_file) for each page encountered.
Fork me on GitHub