nicetoolbox.detectors.config_handler.Configuration¶

class nicetoolbox.detectors.config_handler.Configuration(project_folder: Path, machine_specifics_file: Path, run_config_file: Path)[source]¶

Bases: ProjectConfigHandler

Handles loading and resolving all configurations required for detectors pipeline. This includes: - machine specifics - project config - run configuration file - detectors configuration - dataset properties

Further provides a config factory that produces frozen and resolved runtime configs per video context

Load all static configuration files.

Parameters:

project_folder (Path) – Path to the project folder containing nice_project.toml.
machine_specifics_file (Path) – Path to machine_specific_paths.toml, may contain placeholders.
run_config_file (Path) – Path to detectors_run_file.toml, may contain placeholders.

Methods

`get_all_detector_names`	Returns all detector names defined in the detectors configuration.
`iter_sequence_contexts`	Iterate over all videos and yield frozen runtime configurations.
`save_experiment_config`
`save_video_config`

Attributes

`check_missing_detectors_dependencies`
`error_level`
`log_file`
`log_level`
`save_csv`
`visualize`
`machine_specific_path`
`run_config_file_path`
`machine_specific_config`
`run_config`
`detectors_config`
`dataset_properties`
`predictions_mapping`
`project_folder`
`project_config_path`
`cfg_loader`
`auto_placeholders`
`runtime_placeholders`
`project_config`

get_all_detector_names() → list[str][source]¶: Returns all detector names defined in the detectors configuration.

iter_sequence_contexts() → Generator[SequenceRuntimeConfig, None, None][source]¶

Iterate over all videos and yield frozen runtime configurations.

Each yielded SequenceRuntimeConfig is fully resolved and immutable. It should be discarded after the video is processed.

Yields:: SequenceRuntimeConfig for each video defined in the run configuration