Skip to contents

A lazy dplyr::tbl() over triples joined to the item and collection it came from, so every subject/predicate/object extraction carries its provenance:

Usage

entropia_entity_relations(con, min_confidence = NULL)

Arguments

con

A connection returned by entropia_connect().

min_confidence

Unused in the current schema (triples carry no confidence column); must be NULL (the default).

Value

A tbl_sql.

Details

  • id, subject, predicate, object: the triple itself.

  • item_id, item_title: the owning item and its title.

  • collection_id, collection_name: the item's collection.

  • asset_id: the specific asset the triple is scoped to, NA for item-level triples.

Everything stays lazy: the joins run in SQLite when the result is collected.

min_confidence is accepted for API stability with the plan's research layer, but the current schema's triples table has no confidence column, so passing a non-NULL value errors with guidance rather than silently doing nothing.

Examples

con <- entropia_connect(system.file("extdata", "entropia-example.sqlite",
  package = "entropiaR"
))
entropia_collect(entropia_entity_relations(con))
#> # A tibble: 1 × 9
#>   id          subject predicate object item_id asset_id item_title collection_id
#>   <chr>       <chr>   <chr>     <chr>  <chr>   <chr>    <chr>      <chr>        
#> 1 88888888-8… Juan P… particip… la hu… 222222… NA       Manifiest… 11111111-111…
#> # ℹ 1 more variable: collection_name <chr>
entropia_disconnect(con)