Save sample annotations

Saves the current annotations for a sample. By default annotations are stored in a local SQLite database (annotations.sqlite in the database folder). Optionally, a MATLAB-compatible .mat file can also be written.

Usage

save_sample_annotations(
  sample_name,
  classifications,
  original_classifications,
  changes_log,
  temp_png_folder,
  output_folder,
  png_output_folder,
  roi_folder,
  class2use_path,
  class2use = NULL,
  annotator = "Unknown",
  adc_folder = NULL,
  save_format = "sqlite",
  db_folder = get_default_db_dir(),
  export_statistics = TRUE
)

Arguments

sample_name

Sample name (e.g., "D20230101T120000_IFCB134")

classifications

Current classifications data frame

original_classifications

Original classifications data frame (for comparison)

changes_log

Changes log data frame from create_empty_changes_log

temp_png_folder

Path to temporary folder with extracted PNG images

output_folder

Output folder path for MAT files and statistics

png_output_folder

PNG output folder path (organized by class)

roi_folder

ROI folder path (for ADC file location, used as fallback)

class2use_path

Path to class2use file

class2use

Character vector of class names. When NULL (default), loaded from class2use_path.

annotator

Annotator name for statistics

adc_folder

Direct path to the ADC folder. When provided, this is used instead of constructing the path via get_sample_paths. This supports non-standard folder structures.

save_format

One of "sqlite" (default), "mat", or "both". Controls which backend(s) are written.

db_folder

Path to the database folder for SQLite storage. Defaults to get_default_db_dir(). Should be a local filesystem path, not a network drive.

export_statistics

Logical. When TRUE (default), validation statistics CSV files are written to a validation_statistics/ subfolder inside output_folder. Set to FALSE to skip this export, e.g. when annotating from scratch.

Value

TRUE on success, FALSE when there is nothing to save (empty changes log) or required inputs are missing. Errors raised while writing any backend propagate to the caller, so callers can distinguish a failed save from an empty one.

Examples

if (FALSE) { # \dontrun{
# Save annotations for a sample (default: SQLite)
success <- save_sample_annotations(
  sample_name = "D20230101T120000_IFCB134",
  classifications = current_classifications,
  original_classifications = original_classifications,
  changes_log = changes_log,
  temp_png_folder = "/tmp/png",
  output_folder = "/data/manual",
  png_output_folder = "/data/png_output",
  roi_folder = "/data/raw",
  class2use_path = "/data/class2use.mat",
  annotator = "John Doe"
)
} # }