Entities (lazy)

A lazy dplyr::tbl() over the entities table (named-entity recognition output). Two filters are applied, both pushed down to SQL:

Usage

entropia_entities(con, include_deleted = FALSE, min_confidence = NULL)

Arguments

con

A connection returned by entropia_connect().

include_deleted

Include soft-deleted entities (source = "manual_deleted"). Default FALSE.

min_confidence

Optional numeric threshold in [0, 1]; rows with confidence < min_confidence are excluded. NULL (default) keeps all confidence levels.

Value

A tbl_sql on entities.

Details

• By default rows marked soft-deleted (source = "manual_deleted", the app's hidden marker) are excluded; pass include_deleted = TRUE to keep them. On schemas that predate the source column (migration 0009) there is no marker to honour and the filter is a no-op.

• Set min_confidence to keep only entities at or above a confidence threshold.

created_at uses the magnitude-guarded datetime_auto contract (the app writes epoch milliseconds, the DDL default is seconds); asset_id is NULL for item-level entities.

Examples

con <- entropia_connect(system.file("extdata", "entropia-example.sqlite",
  package = "entropiaR"
))
entropia_collect(entropia_entities(con)) # soft-deleted rows excluded
#> # A tibble: 3 × 16
#>   id         item_id entity_type value start_offset end_offset confidence source
#>   <chr>      <chr>   <chr>       <chr>        <int>      <int>      <dbl> <chr> 
#> 1 77777777-… 222222… person      Juan…            0         10       0.97 ner   
#> 2 77777777-… 222222… place       Plaz…           12         24       0.93 ner   
#> 3 77777777-… 222222… organizati… Sind…            5         26       0.9  ner   
#> # ℹ 8 more variables: model_name <chr>, created_at <dttm>, latitude <dbl>,
#> #   longitude <dbl>, geo_status <chr>, asset_id <chr>, manual_lat <dbl>,
#> #   manual_lon <dbl>
entropia_collect(entropia_entities(con, include_deleted = TRUE))
#> # A tibble: 4 × 16
#>   id         item_id entity_type value start_offset end_offset confidence source
#>   <chr>      <chr>   <chr>       <chr>        <int>      <int>      <dbl> <chr> 
#> 1 77777777-… 222222… person      Juan…            0         10       0.97 ner   
#> 2 77777777-… 222222… place       Plaz…           12         24       0.93 ner   
#> 3 77777777-… 222222… organizati… Sind…            5         26       0.9  ner   
#> 4 77777777-… 222222… person      Pers…            0         14       0.5  manua…
#> # ℹ 8 more variables: model_name <chr>, created_at <dttm>, latitude <dbl>,
#> #   longitude <dbl>, geo_status <chr>, asset_id <chr>, manual_lat <dbl>,
#> #   manual_lon <dbl>
entropia_collect(entropia_entities(con, min_confidence = 0.9))
#> # A tibble: 3 × 16
#>   id         item_id entity_type value start_offset end_offset confidence source
#>   <chr>      <chr>   <chr>       <chr>        <int>      <int>      <dbl> <chr> 
#> 1 77777777-… 222222… person      Juan…            0         10       0.97 ner   
#> 2 77777777-… 222222… place       Plaz…           12         24       0.93 ner   
#> 3 77777777-… 222222… organizati… Sind…            5         26       0.9  ner   
#> # ℹ 8 more variables: model_name <chr>, created_at <dttm>, latitude <dbl>,
#> #   longitude <dbl>, geo_status <chr>, asset_id <chr>, manual_lat <dbl>,
#> #   manual_lon <dbl>
entropia_disconnect(con)