SilverBullet automatically builds and maintains an index of objects extracted from all markdown pages in your space. It subsequently allows you to query this database in (potentially) useful ways.
By design, the truth remains in the markdown: all data indexed as objects will have a representation in markdown text as well. This index can be flushed at any time and be rebuilt from its source markdown files kept in your space (and you can do so on demand if you like using the command).
Object representation
Every object has a set of Attributes, some predefined, but you can add any additional custom attributes that you like.
The following attributes are predefined, and you can expect all objects to have them:
ref: a globally unique identifier, often represented as a pointer to the place (page, position) in your space where the object is defined. For instance, a page object will use the page name as its ref attribute, and a task will use page@pos (where pos is the location the task appears in page).
tag: the main type, or “tag” of the page, usually a built-in type of the object (see below).
In addition, many objects will also contain:
tags: an optional set of additional, explicitly assigned tags.
itags: a set of implicit or inherited tags: including the object’s tag, tags as well as any tags assigned to its containing page. This is useful to answer queries like, “give me all tasks on pages where that page is tagged with person“, which would be expressed as task where itags = "person" (although technically that would also match any tags that have the #person explicitly assigned).
Beside these, any number of additional tag-specific and custom Attributes can be defined (see below).
Tags
Every object has a main tag, which signifies the type of object being described. In addition, any number of additional tags can be assigned as well via the tags attribute. You can use either the main tag or any of the tags as query sources in Live Queries — examples below.
Here are the currently built-in tags:
page
Every page in your space is available via the page tag. You can attach additional tags to a page, by either specifying them in the tags attribute Frontmatter, or by putting additional Tags in a stand alone paragraph with no other (textual) content in them, e.g.:
#example-tag#another-tag
In addition to ref and tags, the page tag defines a bunch of additional attributes as can be seen in this example query:
page where name = @page.name
Note that you can also query this page using the example-tag directly:
example-tag
table
Markdown table rows are indexed using the table tag, any additional tags can be added using Tags in any of its cells.
Title
Description Text
This is some key
The value contains a #table-tag
Some Row
This is an example row in between two others
Another key
This time without a tag
table
Table headers will be normalized by converting them to lowercase and replacing all non alphanumeric characters with _.
task
task Every task in your space is tagged with the task tag by default. You tag it with additional tags by using Tags in the task name, e.g.
My task #upnext
And can then be queried via either task or upnext.
The following query shows all attributes available for tasks:
upnext
Although you may want to render it using a template such as Task instead:
upnext render [[Library/Core/Query/Task]]
taskstate
Tasks support the default x and states (done and not done), but custom states as well. Custom states used across your space are kept in taskstate:
NOT STARTED Task 1
IN PROGRESS Task 2
And can be queried as follows:
taskstate where page = @page.name
template
Indexes all pages tagged with #template. See Templates for more information on templates.
template select name limit 5
item
List items (both bullet point and numbered items) are indexed with the item tag, and additional tags can be added using Tags.
Here is an example of a #quote item using a custom attribute:
“If you don’t know where you’re going you may not get there.” #quote
And then queried via the #quote tag:
quote where page = @page.name and tag = "item" select name, by
paragraph
Top-level paragraphs (that is: paragraphs not embedded in a list) are indexed using the paragraph tag, any additional tags can be added using Tags.
A paragraph with a #paragraph-tag.
paragraph-tag
data
You can also embed arbitrary YAML data blocks in pages via fenced code blocks and use a tag as a coding language, e.g.
name: Pete age: 55
Which then becomes queriable via the person tag:
person
link
All page links are tagged with link. You cannot attach additional tags to links. The main two attributes of a link are:
toPage the page the link is linking to
page the page the link appears on
In addition, the snippet attribute attempts to capture a little bit of context on where the link appears.
Note: this is the data source used for the feature as well page .
Here is a query that shows all links that appear in this particular page:
link where page = @page.name
anchor
Anchors use the notation to allow deeplinking into a page and are also indexed and queryable. It is not possible to attach additional tags to an anchor.
Here is an example query:
anchor where page = @page.name
header
Headers (lines starting with #, ## etc.) are indexed as well and queriable.
header where page = @page.name limit 3
tag
The ultimate meta tag is tag itself, which indexes for all tags used, in which page they appear and what their “parent tag” is (the context of the tag: either page, item or task).
Here are the tags used/defined in this page:
tag where page = @page.name select name, parent
attribute
This is another meta tag, which is used to index all Attributes used in your space. This is used by e.g. attribute completion in various contexts. You likely don’t need to use this tag directly, but it’s there.
attribute where page = @page.name limit 1
System tags
The following tags are technically implemented a bit differently than the rest, but they are still available to be queried.
command
Enables querying of all Commands available in SilverBullet as well as their assigned keyboard shortcuts.
command order by name limit 5
syscall
Enables querying of all PlugOS syscalls enabled in your space. Mostly useful in the context of Plugs and Space Script development.