Conversion from markdown files to database entries to use as the backend of a blog
Simple python script for taking a directory of markdown files and generating/storing the backend of a blog.
The goal is to enable quick editing of simple text-files and posting them to a database via a push to a git branch (default ‘publish’).
Upon running, a python JSON object is created from each file found. There is a markdown header extracted from the file indicating post title, date posted, authors, tags, etc. The content of the post is converted automatically to html and added to the final object in the path.
The resulting objects (currently) are sent to a MongoDB session and saved to the specified collection.
This process is strictly a ‘model’ management system, any view and controller must be built/managed by you.
(The name comes from the “Aberdeen” fish hook 🎣)
Copy the post-update, aberdeen.py and config files to the directory hooks in the git repository on the server. Edit the config file to your specifications. This will be updated when uploaded to pypi.
On your server, create a bare git repository, something like ‘blog_data’. This will simply hold all your markdown (or maybe other type) files. Create a ‘publish’ branch in addition to another ‘working’ one (presumably ‘master’). Add the post-update webhook and configuration as explained in Installation. Clone the repo to your working computer.
This program requires a key-value pair header in each of the markdown files that have typical elements required for blogging
--- title : Post Title date : Mar 15, 2015 tags : Example Feeling Happy XYZ author: Me --- # My New Post This is a great post! *All* my markdown works
The ‘tags’ attribute in this example will generate a list of strings; for more information on how the metadata header works, read this.
Aberdeen creates a python ‘time’ object from the ‘date’ attribute. It will try to be smart about the style of the date, and there are a few ways to interpret the datetime from the string, but it has to be accepted in some form or another by the strptime function of python ‘time’ library. The first way to work will be saved, so it rewards consistency. It is recommended you put in a time field if you care about that, else it will default to midnight of the determined date.
Maybe this can be specified in the config file? (that’s not implemented yet.)
This kind of information is great for storing in NoSQL databases, so MongoDB is the only database currently supported. The content of the markdown is converted to HTML and added to the result as ‘html_content’ field. The objects are sorted in terms of date and written to the database. The previous table or collection will dropped and the new items added. (*NO GUARANTEE* that the items will be in the same order).
Remeber this does not have any HTML structure or view-support for a blog. This strictly converts one form of a model (markdown files) to another (database entries). The view/controllers are totally up to you for retreiving and displaying the posts.
Always assume that the database collection/table will be erased upon every push. The idea is the database reflects the files, so changing a file will replace that entry in the database. It is recommended to NOT use fixed links to posts. It is suggested to used date+title as a unique identifier. Alternatively, you could store a unique post id in the metadata field, if you want some assurance that things will be fixed (but it’s up to you to keep track of the and their uniqueness).