The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

animl 3.3.0

New Functions

• active_times(): Get start and stop dates for each camera folder from a manifest or directory.
• export_yolo(): Export a manifest to YOLO format for model training, including train/val/test splits.
• export_camptrapdp(): Export a manifest to Camera Trap Data Package (camtrapdp) format.
• export_camtrapR(): Export data into sorted folders organized by station in camtrapR format.

API Changes

Detection

• load_detector(): Expanded supported model_type values to include "mdv5", "mdv6", "mdv1000-cedar", "mdv1000-larch", "mdv1000-sorrel", "mdv1000-redwood", "mdv1000-spruce", "yolov5", "yolo", and "onnx".
• Detection category mapping (MD_LABELS): numeric categories in the category column now follow the MegaDetector convention — 0 = empty, 1 = animal, 2 = human, 3 = vehicle. This mapping is used consistently by get_empty(), get_animals(), and the visualization functions.
• get_empty(): Returns all non-animal frames (category != 1), including empty (0), human (2), and vehicle (3) detections. NA categories are treated as empty (0). Adds prediction and confidence columns to the result.
• get_animals(): Returns only frames where category == 1 (animal detections).

Classification

• load_classifier(): Changed default architecture from "CTL" to "efficientnet_v2_m".
• classify(): Now automatically unpacks a model if it is passed as a list.
• single_classification(): Renamed parameter predictions_raw to predictions_output; added file_col and failed_files parameters.

Visualization

• plot_box() and plot_all_bounding_boxes(): Replaced label_col parameter with classifier_label_col and added detector_category_col parameter (default "category") for more granular control over bounding box display. The detector_category_col column is used to color-code and label boxes by detection category (0=empty, 1=animal, 2=human, 3=vehicle).

File Management

• sequence_calculation(): Renamed parameter datetime_col to timestamp_col; added sort_columns and file_col parameters.

Export

• export_folders(): Added timestamp_col and station_col parameters.
• remove_link(): Fixed bug where pbapply was used instead of pbsapply.
• export_megadetector(): Output JSON now includes a detection_categories field with the standard category mapping (0=empty, 1=animal, 2=human, 3=vehicle).

Re-ID

Distance metric functions (remove_diagonal(), euclidean_squared_distance(), cosine_distance(), compute_distance_matrix(), compute_batched_distance_matrix()) are now implemented natively in R, removing the dependency on the Python back-end for these calculations.

Other Changes

• video_processing: Increased default num_workers to 4 for faster video processing.
• install_animl(): Added conda environment handling for more flexible Python environment management.
• Added comprehensive testthat unit tests for all core functions.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.