Merge pull request #127 from anagainaru/cleanup

Remove deprecated code and clean up the configuration options

Author: anagainaru

CLAUDE.md

@@ -29,18 +29,16 @@ poetry run mypy .
 The installable package lives under `src/apeiron/` (imported as `apeiron`; see `pyproject.toml` `packages = [{ include = "apeiron", from = "src" }]`).
 
 ### Core Pipeline
-1. **Config** (`src/apeiron/config/configuration.py`): TOML-based config parsed into frozen dataclasses (`Config`, `ModelCfg`, `DataCfg`, `TrainCfg`, `ContinualLearningCfg`, `DriftDetectionCfg`, `LoggingCfg`, `VisualizationCfg`). Supports `--set key=val` CLI overrides and `APP_` env var overrides.
+1. **Config** (`src/apeiron/config/configuration.py`): TOML-based config parsed into frozen dataclasses (`Config`, `ModelCfg`, `DataCfg`, `TrainCfg`, `ContinualLearningCfg`, `DriftDetectionCfg`, `LoggingCfg`). Supports `--set key=val` CLI overrides and `APP_` env var overrides.
 2. **Model Harness** (`src/apeiron/model/torch_model_harness.py`): Abstract `BaseModelHarness` providing `get_stream_dataloader()`, `get_train_dataloaders()`, `get_hist_dataloaders()`, `update_data_stream()`, `get_criterion()`, `get_optmizer()`, and `eval_metrics` dict. Also keeps a per-task registry for transfer metrics -- `register_task()`, `eval_past_tasks()`, `task_diagonals` -- which subclasses inherit unchanged (see `docs/tracking.md` "Transfer Metrics").
 3. **Driver** (`src/apeiron/driver/continuous_monitor.py`): `ContinuousMonitor` orchestrates the monitoring loop -- evaluates batches, checks drift at intervals, dispatches CL training on drift.
 4. **Drift Detection** (`src/apeiron/drift_detection/`): `BaseDriftDetector` ABC with `update(value) -> DriftSignal`. Implementations: ADWINDetector, KSWINDetector, PageHinkleyDetector, ModelPerformanceDetector, ModelEvalDetector, EnsembleDetector.
 5. **Training** (`src/apeiron/training/continuous_trainer.py`): `ContinuousTrainer` runs outer/inner CL loops with gradient accumulation.
 6. **Updaters** (`src/apeiron/training/updater/`): `BaseUpdater` with hooks `cl_preprocessing()`, `fwd_bwd()`, `update_pre_fwd_bwd()`, `update_post_fwd_bwd()`, `update_post_optimizer_call()`, `cl_postprocessing()`. Implementations: base (vanilla), jvp_reg (JVP updater -- now a first-order SAM robust update), ewc_online (EWC), kfac_online (KFAC), none (no-op).
 7. **Evaluation** (`src/apeiron/evaluation/metrics.py`): `accuracy()` and `accuracy_topk()`.
-8. **Logger** (`src/apeiron/logger/`): `Logger` with pluggable metrics backends -- `WandBLogger` and `MLFlowLogger` (configured via `[logging] backend = "wandb"|"mlflow"|"none"`), plus console output. Stages: eval, drift, cl.
+8. **Logger** (`src/apeiron/logger/`): `Logger` with pluggable metrics backends -- `WandBLogger` and `MLFlowLogger` (configured via `[logging] backend = "wandb"|"mlflow"|"none"`), plus console output. Stages: eval, drift, cl. Metrics are written to a CSV file at `[logging] metrics_output_path` for external analysis.
 9. **Profilers** (`src/apeiron/profilers/`): `FLOPSProfiler` (`count_flops.py`) using PyTorch FlopCounterMode.
 
-Note: `[visualization]` config (`VisualizationCfg`) is parsed but there is no bundled dashboard/renderer in the current package; runs emit a CSV at `visualization.input` for external plotting.
-
 ### Example Harnesses
 - `examples/mnist/model.py`: `MNIST_CNN` -- CNN on MNIST with affine drift simulation.
 - `examples/cifar/model.py`: `CIFAR_VISION` -- ViT/VGG on CIFAR-10 with affine drift.
@@ -49,7 +47,7 @@ Note: `[visualization]` config (`VisualizationCfg`) is parsed but there is no bu
 
 ### Configuration Format (TOML)
 Required sections: `[model]` (name, pretrained_path), `[data]` (name, path), `[train]` (batch_size, num_workers, init_lr), `[drift_detection]` (detector_name, detection_interval, etc).
-Optional sections: `[continual_learning]` (update_mode, lambda params), `[logging]` (backend = "wandb"|"mlflow"|"none", experiment_name, mlflow_tracking_uri), `[visualization]` (baseline, input, output -- parsed but not rendered by the package).
+Optional sections: `[continual_learning]` (update_mode, lambda params), `[logging]` (backend = "wandb"|"mlflow"|"none", experiment_name, mlflow_tracking_uri, metrics_output_path).
 Top-level keys: `seed`, `device` ("auto"|"cpu"|"cuda"|"mps"), `multi_gpu`.
 
 ### Available Drift Detectors

examples/cifar/cifar10_vgg11.toml

@@ -46,5 +46,5 @@ adwin_delta = 0.002
 adwin_minor_threshold = 0.3
 adwin_moderate_threshold = 0.6
 
-[visualization]
-input = "output/cifar.csv"
+[logging]
+metrics_output_path = "output/cifar.csv"

examples/cifar/cifar10_vit.toml

@@ -47,5 +47,5 @@ adwin_delta = 0.002
 adwin_minor_threshold = 0.3
 adwin_moderate_threshold = 0.6
 
-[visualization]
-input = "output/cifar.csv"
+[logging]
+metrics_output_path = "output/cifar.csv"

examples/mnist/mnist.toml

@@ -58,6 +58,4 @@ adwin_moderate_threshold = 0.6
 backend = "wandb"  # "wandb", "mlflow", or "none"
 experiment_name = "mnist-continual-learning"  # Optional: project/experiment name
 # mlflow_tracking_uri = "http://localhost:5000"  # Optional: MLflow tracking server
-
-[visualization]
-input = "output/mnist.csv"
+metrics_output_path = "output/mnist.csv"

src/apeiron/__init__.py

@@ -7,7 +7,6 @@
     TrainCfg,
     ContinualLearningCfg,
     DriftDetectionCfg,
-    VisualizationCfg,
     LoggingCfg,
     build_config,
 )
@@ -35,7 +34,6 @@
     "TrainCfg",
     "ContinualLearningCfg",
     "DriftDetectionCfg",
-    "VisualizationCfg",
     "LoggingCfg",
     "build_config",
     "BaseModelHarness",

src/apeiron/config/configuration.py

@@ -182,18 +182,14 @@ def __post_init__(self) -> None:
         object.__setattr__(self, "ensemble_detectors", tuple(names))
 
 
-@dataclass(frozen=True)
-class VisualizationCfg:
-    input: str = "output/output.csv"  # CSV path where run metrics are written
-
-
 @dataclass(frozen=True)
 class LoggingCfg:
-    backend: MetricsBackend = "wandb"  # "wandb", "mlflow", or "none"
+    backend: MetricsBackend = "none"  # "wandb", "mlflow", or "none"
     experiment_name: str | None = (
         None  # Project name for WandB/Experiment name for MLflow
     )
     mlflow_tracking_uri: str | None = None  # MLflow tracking server URI
+    metrics_output_path: str | None = None  # CSV path where run metrics are written
 
 
 @dataclass(frozen=True)
@@ -208,7 +204,6 @@ class Config:
     device: str
     multi_gpu: bool = False
     verbosity: str = "INFO"
-    visualization: VisualizationCfg | None = None
     logging: LoggingCfg | None = None
 
 
@@ -362,7 +357,6 @@ def build_config(argv=None) -> Config:
     train = TrainCfg(**cfg["train"])
     dd = DriftDetectionCfg(**cfg["drift_detection"])
     cl = ContinualLearningCfg(**cfg.get("continual_learning", {}))
-    viz = VisualizationCfg(**cfg["visualization"]) if "visualization" in cfg else None
     log_cfg = LoggingCfg(**cfg["logging"]) if "logging" in cfg else None
 
     raw_device = str(
@@ -385,7 +379,6 @@ def build_config(argv=None) -> Config:
         "train",
         "continual_learning",
         "drift_detection",
-        "visualization",
         "logging",
         "device",
         "multi_gpu",
@@ -400,7 +393,6 @@ def build_config(argv=None) -> Config:
         train=train,
         continual_learning=cl,
         drift_detection=dd,
-        visualization=viz,
         logging=log_cfg,
         device=resolved_device,
         multi_gpu=multi_gpu_flag,

src/apeiron/logger/logger.py

@@ -24,20 +24,14 @@ def __init__(
         verbosity: str = "INFO",
         backend: MetricsBackend = "wandb",
         csv_path: str | Path | None = None,
-        # Legacy parameter for backwards compatibility
-        wandb_enabled: bool | None = None,
     ):
         """Initialize unified logger.
 
         Args:
             verbosity: Console level (DEBUG, INFO, INFO:n, WARNING, ERROR, CRITICAL)
             backend: Metrics backend to use ("wandb", "mlflow", or "none")
             csv_path: Save metrics to CSV at this path (disabled if None)
-            wandb_enabled: Deprecated. Use backend="wandb" or backend="none" instead.
         """
-        # Handle legacy wandb_enabled parameter
-        if wandb_enabled is not None:
-            backend = "wandb" if wandb_enabled else "none"
 
         # Initialize metrics backend
         if backend == "mlflow":
@@ -163,16 +157,13 @@ def get_logger(
     verbosity: str = "INFO",
     backend: MetricsBackend = "wandb",
     csv_path: str | Path | None = None,
-    # Legacy parameter for backwards compatibility
-    wandb_enabled: bool | None = None,
 ) -> Logger:
     """Get or create the default Logger instance.
 
     Args:
         verbosity: Console level (DEBUG, INFO, INFO:n, WARNING, ERROR, CRITICAL)
         backend: Metrics backend to use ("wandb", "mlflow", or "none")
         csv_path: Save metrics to CSV at this path (disabled if None)
-        wandb_enabled: Deprecated. Use backend="wandb" or backend="none" instead.
 
     Returns:
         Logger instance
@@ -183,7 +174,6 @@ def get_logger(
             verbosity=verbosity,
             backend=backend,
             csv_path=csv_path,
-            wandb_enabled=wandb_enabled,
         )
     return _default_logger
 

src/cl_only.py

@@ -578,7 +578,7 @@ def main(argv: list[str] | None = None) -> int:
     logger = get_logger(
         verbosity=cfg.verbosity,
         backend=backend,
-        csv_path=cfg.visualization.input if cfg.visualization else None,
+        csv_path=cfg.logging.metrics_output_path if cfg.logging else None,
     )
 
     project_name = "basesim-framework"

src/drift_only.py

@@ -355,7 +355,7 @@ def main(argv: list[str] | None = None) -> int:
     logger = get_logger(
         verbosity=cfg.verbosity,
         backend=backend,
-        csv_path=cfg.visualization.input if cfg.visualization else None,
+        csv_path=cfg.logging.metrics_output_path if cfg.logging else None,
     )
 
     project_name = "basesim-framework"

src/main.py

@@ -17,7 +17,7 @@ def main(argv: list[str] | None = None) -> int:
     logger = get_logger(
         verbosity=cfg.verbosity,
         backend=backend,
-        csv_path=cfg.visualization.input if cfg.visualization else None,
+        csv_path=cfg.logging.metrics_output_path if cfg.logging else None,
     )
 
     modelHarness = get_example(cfg=cfg)