Light-CMS

What does it do?

Light-CMS is a minimalistic Flat-File-CMS capable of loading scripts in an easy-to-use format in order to create articles or blog-posts without using a database. It renders HTML-templates with Jinja2. The produced HTML code can be used with Flask, Tornado or something similar.

Usage

By default, Light-CMS uses the directory /scripts to load the scripts for inserting the content into a HTML-template.

The functions use the dict-datatype to return the parsed contents. For usual usage you should use the ScriptLoader:

sl = ScriptLoader()
script = sl.get_single_script(<script filename>)
rendered_article = sl.render_article(script, <template filename>)

You then can return the rendered article in your function when using Flask.

The keys to use from the resulting dict are:

Key	Value	HTML-Code
Content	The articles content	Yes
TableOfContents	The table of contents	Yes
Title	The articles title	No

| Date | The articles date | No | | Author | The articles author | No | | Description | The articles description | No | | Keywords | The articles keywords | No | | Filename | The filename behind the article | No | | Date_object | A python date-object (for sorting) | No |

When using the values of the dict to insert into your Jinja2-Template you have to make sure that the values containing pure HTML-code have to be inserted into the template without escaping!

{{ TableOfContents|safe }}		
{{ Content|safe }}

If you want to offer a list of all the articles to your users you can use:

articles = sl.get_article_list()

This time the return value is HTML-code so don't escape it.

Script Format

The script starts with tags and values:

{Title}{Example article}
{Description}{Just an example}
{Keywords}{Example, Flask}
{Date}{13/12/2015}
{Author}{Philipp Hagenlocher}

You can insert as much tags as you want. They will all be parsed and added to the returned dictionary. Title, Date and Author are required!

After the tag section the normal content starts. It works like this:

* Sectiontitle

** Subsectiontitle

The sections and subsections will be numbered automatically. The titles (aswell as the title of the article) will be titlecased. If you don't want them to be titlecase you have to add a '_' infront of the title like this:

{Title}{_Example article}

      ...

*_ Sectiontitle

**_ Subsectiontitle

You can use HTML-code within the scripts.

Using Markdown

It is possible to use Markdown for your content. In order to do so you need to create a 'Filetype' tag and set its value to 'Markdown' (not case sensetive) like this:

{Filetype}{Markdown}

Caching

The provided ScriptLoader has a builtin handler of caching at its disposal. Caching works on 2 levels. The ram- and the disc-cache. The developer can choose how many articles should be kept in ram to directly load them from there. If your webserver doesn't have any RAM to spare, you can only work with the disc-cache which is fine.

The first little performance test showed promising results for the ram cache.

This was the result of test.py on a ThinkPad T410 with an SSD.

No caching (reloading files everytime)

Functioncall	What was loaded	Min (in seconds)	Mean (in seconds)	Max (in seconds)
getarticlelist()	All test articles	1.9138715267181396	1.9623020458221436	2.0416059494018555
getsinglescript()	One test article	0.060880184173583984	0.06287442207336426	0.0752251148223877

Caching (loaded from ram cache)

Functioncall	What was loaded	Min (in seconds)	Mean (in seconds)	Max (in seconds)
getarticlelist()	All test articles	0.024015188217163086	0.02435166358947754	0.03130984306335449
getsinglescript()	One test article	0.0007548332214355469	0.0009222960472106933	0.0020666122436523438

Dynamic Caching

Light-CMS checks if the script files have changed automatically and renews cache entries so that the content in your script file is also the content that is presented to the visitors of your website. This is done by checking fingerprints of the loaded files.