-

2025-03-27 22:21:47

adminsforum@hubzilla.org

Thorsten wrote the following post Thu, 27 Mar 2025 22:23:19 +0100

@!Hubzilla Support Forum
Hello,
I would like to start the attempt to move our club homepage to Hubzilla. The homepage was created with Grav and, apart from a few static pages, consists of many blog entries in Markdown format. Is there an easy way to port all these blog entries to Hubzilla, including the images used in the posts and the publication date, so that I can move the history with them?

Each blogpost is located in a separted directory and looks like this:

'---
title: 'Some title'
date: '30-06-2009 00:00'
taxonomy:
category:
- blog
tag:
- aktuelles
author:
name: null
'---

[img=pict2700_640x480.jpg]pict2700_640x480.jpg[/img]

Blogtext

I would like to convert these blog posts into the typical Hubilla posts. Any ideas on the best way to get started?

show all 14 comments

2025-03-27 22:34:31

*_jayrope

jrp@hub.kliklak.net

@Thorsten Very good question, which i've myself. However just as a report from experience i've refrained from importing my own blogs into HZ yet, since you will get a database of your blog entries at least 100x bigger, as it was before. Maybe consider this, before you make that step. Otherwise i know of no help, please consider making a request at the git repository, if serious.

2025-03-27 22:41:23

Thorsten

thorsten@hochminuseins.net

@*_jayrope What do you mean by 100 times larger database in terms of blog entries? Does it make a difference whether I publish posts or articles in Hubzilla or import blog entries? I could also copy each post “manually” and publish it on the homepage channel?!? Am I missing something?

2025-03-27 22:45:38

*_jayrope

jrp@hub.kliklak.net

@Thorsten No, it makes a difference compared to your original blog size. As an example i am running blogs on Textpattern (a CMS, where lterally a thousand articles consume maybe 8 MB in my database (media data doesn't need to be saved). If i did the same in Hubzilla I'd end up with gigabytes of data in my database.

2025-03-27 23:19:00

Thorsten

thorsten@hochminuseins.net

@*_jayrope So either Textpattern is incredibly economical with data or Hubzilla is incredibly wasteful with data, which I somehow can't imagine. Would be interesting to know how big the database would be for something like WordPress.

2025-03-28 00:46:46

*_jayrope

jrp@hub.kliklak.net

@Thorsten Not much bigger than Textpattern. That said, Hubzilla can be a CMS, but it isn't build to be just that.

2025-03-28 02:27:15

Scott M. Stolz

scott@loves.tech

Why would it be bigger? Is it storing images in the database instead of the filesystem? I've looked in the database and it appears to just store the text there.

2025-03-28 02:31:32

Scott M. Stolz

scott@loves.tech

@*_jayrope That is the whole point. You can have your blog on the same domain as your fediverse social media presence.

2025-03-28 08:53:49

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

This sounds like a task for a purpose-made 3rd-party app.

The images would have to be downloaded from your blog server to the file space of your Hubzilla channel. They would also have to be indexed so that the app knows the new file name of each image file, and it can re-associate it with each post.

In fact, if you don't only have resized (640x480) copies of the images on your blog server, but the same images at their original resolution along with them, the app should rather only take the original images with itself.

Of course, the markup would have to be converted from Markdown + HTML to Hubzilla-flavoured BBcode.

In the image-embedding code, the original file name would have to be replaced with the new name and path of the file in your Hubzilla file space. Hence the index above.

[img=pict2700_640x480.jpg]pict2700_640x480.jpg[/img]
would have to be converted to
[zrl=https:///photos//image/][zmg=640x480]https:///photo/.jpg[/zmg][/zrl]

Upon adding each post to the database, the post date would have to be modified in the database so that a blog post from 2009 becomes a Hubzilla post from 2009. This would lead to the weird situation of there being posts on Hubzilla that pre-date not only Hubzilla itself, but the Red Matrix and even Mistpark. If you have posts from 2008, you have posts that would pre-date the entire Fediverse on Hubzilla.

The above is necessary to put the posts into the correct order as well.

The above is also necessary so that the Mastodon Home Owners Association and Mastodon's alt-text police won't take these posts for brand-new and demand you add alt-text to all images. I've yet to see them demand alt-text for image posts that are several years old, also because they'll usually never get to see image posts that are that old.

2025-03-28 09:32:44

Harald Eilertsen

harald@hub.volse.no

@Thorsten Like @Jupiter Rowland says, you would have to write a custom script to perform the import. Exactly how to do that depends on your exact requirements. It's definitely doable, but will require some work.

2025-03-28 10:05:08

Thorsten

thorsten@hochminuseins.net

Many thanks to all of you. There are a few points that I need to think about in more detail. Above all, the question of whether all the effort is really worth it.

On the one hand, I would really like to try it out with Hubzilla, but on the other hand, I wonder to what extent this is really future-oriented for our club if someone other than me who is not familiar with Hubzilla should/must maintain the system. But that's another topic.

2025-03-28 12:16:32

*_jayrope

jrp@hub.kliklak.net

@Scott M. Stolz The reason, why HZ's database is so much bigger is unknown to me. Although i suppose, that it has to do with the sheer number of incoming protocol request every day. However i have no detailed insight there. Fact is, that databases are smaller by between 1 to 2 orders of magnitude - on dedicated blog CMS (like Textpattern and i suppose WP) @Thorsten

2025-03-28 12:22:37

Scott M. Stolz

scott@loves.tech

@*_jayrope If you're comparing it to a blog, yes, of course it'll be much bigger. All of the messages, likes, and boosts you've sent are all in there. Anyone who commented on, liked, or boosted your posts would be in there too. And if you follow people, their posts, comments, likes, and boosts are in there too. And this is before you get to any CMS content.

2025-03-28 12:26:03

*_jayrope

jrp@hub.kliklak.net

@Scott M. Stolz That makes sense. I am not running a WP site atm, but it would be interesting to hear about database size there from someone, who federates with it.

2025-03-28 13:06:13

Scott M. Stolz

scott@loves.tech

@*_jayrope I am not sure how their ActivityPub plugin works, but the size of the database depends on a couple of things.

If they are only concerned with people following the blog and do not follow anyone, then they just need to have a list of followers, plus likes, and possibly boosts. It also has to store any comments received for the post. So similar to what we store.

If they follow people on the fediverse and read their posts within WordPress, then they have to store additional information, which includes all of the posts, comments, likes, and boosts of people they follow. This is usually where your storage balloons but it depends on how many people you follow and how active they are.

For similar functionality, you'd need to store the same information.