## -*- coding: utf-8;mode: org; -*- ## This file is best viewed with GNU Emacs Org-mode: http://orgmode.org/
«A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away.» (Antoine de Saint-Exupéry)
lazyblorg – blogging with Org-mode for very lazy people
This is a web log (blog) environment for GNU Emacs with Org-mode which generates static HTML5 web pages. It is much more superior to any other Org-mode-to-blog-solution I have seen so far!
<(All?) your Org-mode files> --lazyblorg--> static HTML pages | v optional upload (shell) script | v your web space
There is a list of similar/alternative Org-mode blogging projects whose workflows seem really tedious to me.
See my original post to the Org-mode ML for how this idea of lazyblorg started in 2011.
This awesome piece of software is a sheer beauty with regard to:
- simplicity of creating a new blog entry
- I mean it: there is no step which can be skipped!
- add heading+content anywhere, add ID, tag, invoke lazyblorg
- I mean it: there is no step which can be skipped!
- integration into my personal, published workflows
- here, you have to either adapt my totally awesome workflows or you
have to find alternative ways of doing following things:
- linking/including image files or attachments in general (I use this Memacs module)
- advantage of my method: I simply add an image file by typing
tsfile:2014-03-03-this-is-my-file-name.jpg
in double-brackets and I really don’t care in which folder the file is currently located on my system
- advantage of my method: I simply add an image file by typing
- copying resulting HTML files to webspace (I do it using unison/rsync)
- probably more to come
- linking/including image files or attachments in general (I use this Memacs module)
- here, you have to either adapt my totally awesome workflows or you
have to find alternative ways of doing following things:
Target group
Lazy users of Org-mode who want to do blogging very easily and totally cool.
Or simply wannabes. I’m perfectly fine with this as long as they use lazyblorg.
Skills necessary
- modifying configuration settings, e.g., in script files
- optional: creating scheduled tasks (cronjob, …) if you are a really lazy one (and if you trust lazyblorg to do its job in the background)
System requirements
Development platform is Debian GNU/Linux. So with any decent GNU/Linux you should be fine as well.
It might work on OS X but I never tried it so far.
I definitely does not work with Microsoft Windows. Although a
programmer can add a couple of os.path.thisorthat()
here and there
and it should be good to go. Please consider sending a pull-request if
you are fixing this issue. Thanks!
Version
Currently (2017-06-17), I consider lazyblorg in beta-status with version 0.95 or so.
Why lazyblorg?
Minimum effort for blogging.
And: your blog entries can be written anywhere in your Org-mode files. They will be found by lazyblorg. :-)
Further advantages are listed below.
Example workflow for creating a blog entry
- write a blog entry anywhere in your Org-mode files
- With lazyblorg, you can, e.g., write a blog article about an event as a sub-heading of the event itself!
- tag your entry with
:blog:
- add an unique ID in the PROPERTIES drawer
- You might want to use a package that automatically generates unique IDs to your headings (I don’t).
- You might want to take a look at this solution using file or directory variables.
- set the state of your entry to
DONE
- make sure that a
:LOGBOOK:
drawer entry will be created that contains the time-stamp
- make sure that a
An example blog entry looks like this:
** DONE An Example Blog Post :blog:lazyblorg:software: CLOSED: [2017-06-18 Sun 00:16] :PROPERTIES: :ID: 2017-07-17-example-posting :CREATED: [2017-06-17 Sat 23:45] :END: :LOGBOOK: - State "DONE" from "NEXT" [2017-06-18 Sun 00:16] :END: […] Today, I found out that…
That’s it. lazyblorg does the rest. It feels like magic, doesn’t it? :-)
Advantages
These things make a blogger a happy one:
No other Org-mode blogging system I know of is able to process blog entries which are scattered across all your Org-mode documents.
No other Org-mode blogging system I know of is able to generate a blog entry with that minimum effort to the author.
You do not need to maintain a specific Org-mode file that contains you blog posts only. *Create* blog posts anywhere in between your notes, todos, contacts, …
And there are some technological advantages you might consider as well:
- You don’t need to write or correct HTML code by yourself.
- produces static, state-of-the-art HTML5
- it’s super-fast on delivery to browsers
- very low computing requirements on your web server: minimum of server load
- No in-between format or tool.
- Direct conversion from Org-mode to HTML/CSS.
- dependencies have the tendency to cause problems when the dependent tools change over time
- lazyblorg should be running fine for a long time after it is set up properly
- Decide by yourself how and where you are hosting your blog files and log files.
- you will find more advantages when running and using lazyblorg - I am very confident about that ;-)
Disadvantages
Yes, there are some disadvantages. I am totally honest with you since we are becoming close friends right now:
- lazyblorg re-generates the complete set of output pages on every run
- this will probably changed in a future release (to me: no high priority)
- most of the time this is not an issue at all
- if pages are generated on a different system as the web server runs on, performance is a minor issue
- if you don’t have thousands of pages, this will not take long
- lazyblorg is implemented in Python:
- Its Org-mode parser supports only a (large) sub-set of Org-mode syntax
and features.
- Basic rule: use an empty line between two different syntax elements such as paragraphs, lists, tables, and so on.
- Whenever I think that an additional Org-mode syntax element is needed for my blog, I start thinking of implementing it
- I am using Pandoc as a fall-back for all other Org-mode syntax elements which works pretty fine
- For a list of general Org-mode parsers please read this page
- Its Org-mode parser supports only a (large) sub-set of Org-mode syntax
and features.
- lazyblorg is using state-of-the art HTML5 and CSS3
- No old HTML4.01 transitional stuff or similar
- Results might not be compatible with browsers such as Internet
Explorer or mobile devices.
- tell your Internet Explorer friends that they should do themselves a favor and switch to a real browser
- You have to accept the one-time setup effort which requires
knowledge of:
- using command-line tools
- modifying configuration files
- summary: getting this beautiful thing to work in your environment
Features
«Technology develops from the primitive via the complex to the simple.»
(Antoine de Saint-Exupéry; note: lazyblorg is currently “primitive” but with a great outlook up to the status of being simple)
Here is a selection of features of lazyblorg which helps you to blog efficiently:
- Converts Org-mode To HTML5: lazyblorg supports a (large sub-)set of
syntax elements of Org-mode
- also see FAQs for “What Org-mode elements are supported by lazyblorg?”
- Different page types allow you to create:
- articles related to a specific date (temporal pages)
- Those articles are published and hardly updated.
- articles not related to a specific date (persistent pages)
- Frequent updates or the absence of any day-relation makes this page type very sexy to use.
- articles describing a tag you are using (tag pages)
- Yes, with lazyblorg, you are (optionally) able to explain how you are using a certain tag. You can link your most important tag-related articles and so forth. Most systems don’t offer any possibility to communicate the meaning of the tags used.
- the entry page of your blog
- You gotta give them a starting page ;-)
- the templates which are used to generate your blog pages
- Hooray, you are able to define all templates of your blog within Org-mode as well. No need to edit source code here. Isn’t this great?
- articles related to a specific date (temporal pages)
- To efficiently notify users of new articles or changes to existing articles, lazyblorg generates RSS/ATOM feeds.
- Really fast to use linking to other blog articles using their ID property.
- At the bottom of each article, there is a list of related articles that back-link to here.
- You can very easily embed image files with automatically scaling to
their desired width
- This feature is hardened against image file renaming and broken links because of moving images files to different folders
- Users of Memacs do have advanced possibilities here as well
- An optional image cache directory holds previously resized image file and therefore prevents resizing effort for each run.
- For navigating through the blog articles I do recommend using the tags. Articles related to one topic share common tags whereas a date-oriented archive has only very limited use. The tag cloud which is on the tag overview page offers a quick overview of your most used tags.
- There is a search feature which brings you to the content by searching for keywords or phrases.
- Easy embedding of external content such as Tweets or YouTube videos.
- You can exclude content from being published with various features:
- Comment lines
- Mark an article/heading as hidden using the tag NOEXPORT
- The hidden tag does publish an article but hides it from the entry page, navigational pages, and the feeds. This way, you can publish pages who can only be access by people knowing its URL.
FAQs
See https://github.com/novoid/lazyblorg/wiki/FAQs
Installing and Starting with lazyblorg
Currently, lazyblorg is in beta status. It’s not finished yet. However, I am using it for my own blog and therefore it gets more and more ready to use as I add new features.
What’s working so far:
- parsing a very basic sub-set of Org-mode
- parsing the HTML templates
- generating HTML5 pages with a sub-set of the sub-set of the Org-mode syntax elements
External dependencies
The number of external dependencies is kept at a minimum.
This is a list of the most important dependencies:
- Werkzeug
- for sanitizing path components
- I installed it on Debian GNU/Linux with
apt-get install python-werkzeug
- pickle
- object serialization
- most likely: should be part of your Python distribution
- pypandoc
- some Org-mode syntax elements are being converted using Pandoc and its Python binding pypandoc
- you can get it via
sudo apt-get install pandoc
andsudo pip install pypandoc
- Note: Debian GNU/Linux 8 (Jessie) comes with a Pandoc version
which is has bugs. Please install a more recent version. I
upgraded to
pandoc-1.15.1-1-amd64.deb
from: http://pandoc.org/installing.html
- opencv-python
- lazyblorg scales embedded images according to the HTML export attributes
- Install using
sudo apt-get install python-opencv
- Sass (optional) if you want to generate your CSS from the scss-file
All other libraries should be part of a standard Python distribution.
How to Start
- Get the source
git clone https://github.com/novoid/lazyblorg.git
or download current version as ZIP file
- Adapt
config.py
to meet your settings. - Do a technological test-drive
- start:
lazyblorg/example_invocation.sh
- this should work with GNU/Linux (and most probably OS X)
- if not, there is something wrong with the set-up; maybe missing external libraries, wrong paths, …
- start:
- Study, understand, and adopt the content of example_invocation.sh
- with this, you are able to modify command line parameters to meet your requirements
- if unsure, ask for help using
lazyblorg.py --help
- Get yourself an overview on what defines a lazyblorg blog post and
write your own blog posts. A (normal temporal) blog article consists of:
- A (direct) tag has to be
blog
- Sorry, no tag inheritance. Every blog entry has to be explicitely tagged.
- You have to add an unique
:ID:
property - The entry has to be marked with
DONE
- A
:LOGBOOK:
entry has to be found with the time-stamp of setting the entry toDONE
- in my set-up, this is created automatically
- Do not use Org-mode elements that lazyblorg does not understand
- You should not get a disaster if you are using new elements. The result might disappoint you, that’s all.
- However, many Org-mode elements are automatically converted through pandoc.
- A (direct) tag has to be
- OPTIONAL: Write your own CSS file
- you can take a look on mine if you do not care that I am not really into Web design :-)
- please replace hard-coded URL to CSS file in lazyblorg/templates/blog-format.org and link it to your CSS file
- OPTIONAL: Adopt the blog template
- default template is defined in lazyblorg/templates/blog-format.org
- OPTIONAL: Create tag pages for your most important tags where you describe how you are using this tag, what are the most important blog entries related to the tag and so forth.
- Publish your pages on a web space of your choice
- publishing can be done in various ways. This is how I do it using
lazyblorg/make_and_publish_public_voit.sh
which is an adopted version oflazyblorg/example_invocation.sh
:- invoking
testall.sh
- this is for checking whether or not recent code changes did something harmful to my (unfortunately very limited) set of unit tests
- invoking
lazyblorg
with my more or less fixed set of command line parameters - invoking
rsync -av testdata/2del/blog/* $HOME/public_html/
- it synchronizes the newly generated blog data to the local copy of my web space data
- this separation makes sense to me because with this, I am able to do test drives without overwriting my (local copy of my) blog
- invoking unison
- in order to transfer my local copy of my web space data to my public web space
- invoking
- This method has the advantage that generating (invoking
lazyblorg
) and publishing (invokingunison
) are separate steps. This way, I can locally re-generate the blog (for testing purposes) as often I want to. However, as long as I do not sync it to my web space, I keep the meta-data (which is in the local web space copy) of the published version (and not the meta-data of the previous test-run).
- publishing can be done in various ways. This is how I do it using
- Have fun with a pretty neat method to generate your blog pages
Because we are already close friends now, I tell you a hidden feature of lazyblorg nobody knows yet: whenever you see a π-symbol in the upper right corner of a blog entry on my blog: this is a link to the original Org-mode source of that page. This way, you can compare Org-mode-source and HTML-result right away. Isn’t that cool? :-)
Five categories of page type
There are five different types of pages in lazyblorg. Most of the time, you are going to produce temporal pages. However, it is important to understand the other ones as well.
In order to process a blog-heading to its HTML5 representation, its
Org-mode file has to be included in the --orgfiles
command line
argument of lazyblorg.py
. Do not forget to include the archive files
as well.
- temporal
- persistent
- tags
- entry page
- templates
Please do read https://github.com/novoid/lazyblorg/wiki/Page-Types for important details.
BONUS: Preview Blog Article
It is tedious to re-generate the whole blog and even upload it to your web-space just to check the HTML version of the article you are currently writing.
Yeah, this also sucks at my side.
Good news everybody: There is a simple method to preview the article
under the cursor. The script preview_blogentry.sh contains an ELISP
function that extracts the current blog article (all lazyblorg criteria
has to be fulfilled: ID, blog
tag, status DONE
), stores it into a
temporary file, and invokes lazyblorg via preview_blogentry.sh
with
this temporary file and the Org-mode file containing the format
definitions.
If this worked out, your browser shows you all generated blog articles.
Please do adopt the mentioned scripts to you specific requirements - the ones from the repository are for my personal set-up which is unlikely to fit yours (directory paths mostly).
Bang! Another damn cool feature of lazyblorg. This is going better and better. :-)
BONUS: Jump From URL to Blog Article
Imagine, you’re looking at a blog article of your nice lazyblorg-generated blog. Now you want to go to the corresponding Org-mode source to fix a typo.
The issue here is, that you have to either know, where your heading is located or you have to go to the HTML page source, extract the ID, and jump to this ID.
I’ve got a better method: put the URL of your blog article into your
clipboard (via C-l C-c
), press a magic shortcut in Emacs, and BAAAM!
you’re right on spot.
How’s that magic happening?
Just use the following Emacs lisp code snippet, adapt the domain
string, and assign a keyboard shortcut:
(defun my-jump-to-lazyblorg-heading-according-to-URL-in-clipboard ()
"Retrieves an URL from the clipboard, gets its Org-mode source,
extracts the ID of the article and jumps to its Org-mode heading"
(interactive)
(let (
;; Getting URL from the clipboard. Since it may contain
;; some text properties we are using substring-no-properties
;; function
(url (substring-no-properties (current-kill 0)))
;; This is a check string: if the URL in the clipboard
;; doesn't start with this, an error message is shown
(domain "http://karl-voit.at")
)
;; Check if URL string is from my domain (all other strings do
;; not make any sense here)
(if (string-prefix-p (upcase domain) (upcase url))
;; Retrieving content by URL into new buffer asynchronously
(url-retrieve url
;; call this lambda function when URL content is retrieved
(lambda (status)
;; Extrating and preparing the ID
(let* (
;; Limit the ID search to the top 1000 characters of the buffer
(pageheader (buffer-substring 1 1000))
;; Start index of the id
(start (string-match "<meta name=\"orgmode-id\" content=\"" pageheader))
;; End index of the id
(end (string-match "\" />" pageheader start))
;; Amount of characters to skip for the openning tag
(chars-to-skip (length "<meta name=\"orgmode-id\" content=\""))
;; Extract ID
(lazyblorg-id (if (and start end (< start end))
;; ... extract it and return.
(substring pageheader (+ start chars-to-skip) end)
nil))
)
(message (concat "Looking for id:" lazyblorg-id " ..."))
(org-open-link-from-string (concat "id:" lazyblorg-id))
)
)
)
(message (concat "Sorry: the URL \"" (substring url 0 (length domain)) "...\" doesn't start with \"" domain "\". Aborting."))
)
)
)
BONUS: Embedding External Things
- Do read the Wiki for embedding external stuff like Tweets or YouTube videos.
How to Thank Me
I’m glad you like my tools. If you want to support me:
- Send old-fashioned postcard per snailmail - I love personal feedback!
- see my address
- Send feature wishes or improvements as an issue on GitHub
- Create issues on GitHub for bugs
- Contribute merge requests for bug fixes
- Check out my other cool projects on GitHub
If you want to contribute to this cool project, please fork and contribute!
Issues, bugs,… are maintained in the GitHub issue tracker.
I am using Python PEP8 and some ideas from Test Driven Development (TDD).