blog: (micro)blogging management

Blog commands are high level tools to handle an XMPP blog. They are using the generic pubsub arguments


publish a blog item.

pubsub commands are used to specify the destination item.

stdin is used to get the body of the blog post.


Create a blog post with a body, a subject, 2 tags, and with comments allowed:

$ echo "This is post body" | li blog set -T "This is a test message" -t test -t li -C

Create a blog post with rich content using markdown syntax, and no subject:

$ echo "This is a **rich** body" | li blog set -S markdown


get command retrieves one or more blog post(s) from specified location (by default the personal blog of the profile).

output can be customised to only retrieve some keys, or to use a specific template. For instance, the following command retrieves only the title and publication date of the personal blog of the profile:

$ li blog get -k title -k published

pubsub commands are used to specify the blog location.


Retrieve personal blog of the profile using fancy output with a verbosity of 1 (to show publication date):

$ li blog get -O fancy -v

Retrieve title and publication date of last 3 blog posts from the blog at

$ li blog get -m 3 -u -k title -k published

Retrieve last 2 posts of personal blog, and output them in browser using default template:

$ li blog get -m 2 -O template --oo browser


With edit command you can create a new blog post or modify an existing one using your local editor (the one set in $EDITOR). You’ll edit 2 things: the body of the post, and the metadata which contain things like title, comments infos, or tags.

For some common editors (like vim or Emacs), the editor will be automatially opened using a split screen with body in one side, and metadata on the other. If the editor is not supported or doesn’t support split screen, you’ll edit first the body, then the metadata. You can also specify editor and arguments in libervia.conf, see configuration below

If you don’t change anything or publish an empty blog post, the edition will be cancelled.

In the metadata (see below for details), you can use "publish": false to forbid the publication (or set it with --no-publish argument). In this case, when you’ll save your modification and quit your editor, the blog post will not be published but saved locally in a draft. To continue your work later, just start your edition with the -D, --current option like this:

$ li blog edit -D

Note that item location must be re-specified if it has been used to create the draft, so you’ll have to reproduce the arguments to specify service, node or item (or the URL), other data like tags will be restored from draft file of metadata.

You can specify the syntax by using -S SYNTAX, --syntax SYNTAX. If not specified, the syntax set in your parameters will be used.

When you edit a blog post, it is often useful to activate the -P, --preview option, this will launch a web browser and refresh the page each time you save a modification in your editor. By default, the browser registered as default in your system will be used, and a new tab will be opened on each modification. This is not ideal, and we recommand to set you configuration to activate automatic refreshing of the page instead, see preview configuration below to see how to do.


If –preview doesn’t work, use li blog preview (see below) to get error messages. On GNU/Linux, Be sure that inotify Python module is installed correctly.


Edit a new blog post with comments on your personal blog, using default syntax and preview:

$ li blog edit -P --comments

Modifiy a draft previously saved using the "publish": false metadata:

$ li blog edit -D

Correct a typo in your last published blog post:

$ li blog edit --last-item

Edit the blog item at an HTTPS URL using XHTML syntax:

$ li blog edit -u -S xhtml

Create a new blog post on a XMPP blog node using its HTTPS URL (supposing here that is a XMPP blog node):

$ li blog edit -u


Metadata is set using a JSON object. The key you can use are:


boolean indicating if item can be published. Set to false if you want to work on a draft and to avoid accidental publication.


atom entry identifier. This should not be modified manually.


time of initial publication (unix time). This should not be modified manually.


language of the content


array of URIs to the comments node, if any.


array of tags, if any


human readable name of the entry author


jid of the author. This should notbe modified manually.


true if the pubsub service confirmed that author_jid is the one of the publisher. It is useless to modify this variable.


the title of the message


the rich title of the message, in current text syntax. It will be automatically converted to xhtml.



Local editor used is by default the one set in $EDITOR environment variable, but you can specify one in libervia.conf. To do so, you have to set the name of an editor executable in the editor option in [li] section.

You can specify the args to use by using blog_editor_args option. Use {content_file} to get the path of the main content file (the body of the blog post), and {metadata_file} to get the path of the json metadata.

editor = kate
blog_editor_args = {content_file} {metadata_file}


To set the preview, you can use the options blog_preview_open_cmd and blog_preview_update_cmd in your [li] section. the former is the command to use to open your browser when edition starts, and the later is the command to use when a modification is saved. In both cases you may use {url} to set the location of local HTML file.

This can be used to activate automatic refreshing of the page.

For Konqueror, you can use its D-Bus API to do refreshing. Ensure that qdbus is installed on your system, and enter the following lines in your libervia.conf:

blog_preview_open_cmd = konqueror {url}
blog_preview_update_cmd = /bin/sh -c "qdbus $(qdbus org.kde.konqueror\*) /konqueror/MainWindow_1 reload"

For Firefox, you may use xdotool on X11. Once you have installed this tool, enter the following lines in your libervia.conf:

blog_preview_open_cmd = firefox -new-tab {url}
blog_preview_update_cmd = /bin/sh -c "WID=$(xdotool search --name 'Mozilla Firefox' | head -1); xdotool windowactivate $WID; xdotool key F5"

This xdotool technique can be adapted to other browsers.

syntax extensions

A dictionary with a mapping from syntax name to file extension can be used. This can be useful to activate the right syntax highlighting in your editor. There is a default mapping which can be overriden.

The mapping is set in the syntax_ext_dict option of the [li] section of your libervia.conf file. For instance, if your prefer do your .markdown for temp files instead of the default .md, you can use this:

syntax_ext_dict = {"markdown": "markdown"}

the first markdown is the name of the syntax (could be an other syntax like xhtml), while the second if the file extension.


Move a item to a new ID. This works the same as rename, the only difference is that node defaults to microblog default namespace.


After a typo, a use want to rename the blog item with ID my_great_tittle_123a to my_great_title_123a:

$ li blog rename -i my_great_tittle_123a my_great_title_123a


This command will show the specified file in browser, and refresh it when changes are detected. Configuration is the same as for edit preview. This can be used if you have already started an edition with li blog edit but forgot to use the -P, --preview arguments.


Preview the draft at ~/.local/share/libervia/blog/some_name/

$ li blog preview ~/.local/share/libervia/blog/some_name/


With this command you can import an external blog in a XMPP blog at the specified pubsub location.

The import is done using an importer name and a location which depends of the importer (it can be a path to a file, an URL to a blog, or something else). Let empty to get list of importers, and specify only importer name to get its description.

By default, found images are re-uploaded to XMPP server, if you want to keep original URLs, use the --no-images-upload option.

Alternatively, you can re-upload images except for a specific host with --upload-ignore-host UPLOAD_IGNORE_HOST. The images for the specified host will keep there original URLs while other will be uploaded to XMPP server.

You shoud specify original blog host using --host HOST argument, this is used notably to reconstruct relative URLs of media.

Importers may have specific options, you can set them using the -o NAME VALUE, --option NAME VALUE argument. Check the importer description for details.


List available importers:

$ li blog import

Get description of dotclear importer:

$ li blog import dotclear

Import a Dotclear blog:

$ li blog import dotclear /path/to/dotclear.dump

Import a Dotclear blog without uploading images:

$ li blog import --no-images-upload dotclear /path/to/dotclear.dump