Edge labeling? Edge predicate knowledge?

General
1

I’ve been researching various graphy brain-like systems for a while now, but I’m not clear whether logseq has any labeling or knowledge storage capacity for an edge, a link between two nodes. With traditional RDF you have a subject-predicate-object graph philosophy where, e.g., Tom is a captain where Tom is one vertex (with an ontology tag/URI associated with it), is a is the predicate, the edge connection (also with an ontology tag/URI) and captain is the object, i.e., the second vertex connected by is a to vertex Tom. Does logseq have any sort of edge capabilities? My cursory investigation hasn’t revealed anything.

Another example where edge knowledge is very crucial would be to imagine a city street grid as a set of vertices of cross streets and the streets themselves as the edges between the cross streets. For example, 42nd/5th would be a vertex, and 42nd and 5th would be edges. Then you’d have to have a meta-graph to express all of 42nd and 5th. So in such a situation the edges are just as important as the vertices. Is there in fact any sort of edge “labelling” or knowledge building?

2

Hi and welcome in the Logseq community!

In theory this information can be stored using page properties, i.e. in the first block of a page:

key:: value

in terms of graph/vertex/edge:

edge-type-1:: [[Vertex A]]

There is no edge labeling in Graph View at the moment.

What this structure can be used for is queries. In Simple Queries you can use a filter like:

{{query (property key value)}}

Advanced Queries can make this structure even more useful but you need to learn the Datalog/Datomic query language.

Also query results have a table view where columns are the property keys of the results.

I hope this helps.

3

P.S. let me know if you are interested in adding metadata to blocks too; they wouldn’t appear as nodes in Graph View but they can still be searched with queries, for example:

- The Picture of Dorian Gray
  author:: [[Oscar Wilde]]