Skip to contents

Backfill missing ROIs as "unclassified" in the database

Source: R/database.R
fill_unclassified_db.Rd

After a partial import (e.g. import_png_folder_to_db with only a few selected taxa per sample), the database holds annotations for just the imported ROIs. This helper reads each sample's complete ROI list from its .adc file and inserts the ROIs that are not yet in the database as "unclassified", so the full sample is represented. Existing annotations are never modified.

Usage

fill_unclassified_db(
  db_path,
  roi_folder,
  samples = NULL,
  class_name = "unclassified",
  annotator = "imported"
)

Arguments

db_path

Path to the SQLite database file

roi_folder

Base ROI folder path, following the standard IFCB folder structure (roi_folder/YYYY/DYYYYMMDD/sample_name.adc). Used to locate each sample's .adc file via get_sample_paths.

samples

Optional character vector of sample names to backfill. When NULL (the default), all annotated samples in the database are used.

class_name

Class name to assign to the missing ROIs. Default "unclassified".

annotator

Annotator name recorded for the inserted rows. Default "imported".

Value

Named list with counts: added (ROIs inserted), samples (samples that received at least one new ROI), skipped (samples with no reachable .adc file).

Details

Only ROIs with a real image (non-zero width and height) are added. Inserted rows are marked is_manual = 0 (not yet reviewed).

Examples

if (FALSE) { # \dontrun{
db_path <- get_db_path("/data/manual")
fill_unclassified_db(db_path, "/data/ifcb/raw")
} # }