import
numpy as np
import optuna
+
import yaml
+
from omegaconf import DictConfig, OmegaConf
from sklearn.metrics import confusion_matrix
from tqdm import tqdm
@@ -355,6 +357,8 @@
Source code for lidar_prod.tasks.building_validation_optimization
mts_gt = np.array([c.target for c in clusters])
metrics_dict = self.evaluate_decisions(mts_gt, decisions)
log.info(f"\n Results:\n{self._get_results_logs_str(metrics_dict)}")
+ self._save_results_to_yaml(metrics_dict)
+
return metrics_dict
@@ -697,7 +701,31 @@
Source code for lidar_prod.tasks.building_validation_optimization
+ "\nConfusion Matrix (normalized)\n"
+ str(metrics_dict[self.design.metrics.confusion_matrix_norm].round(3))
)
- return results_logs
+
return results_logs
+
+
def _save_results_to_yaml(self, metrics_dict: dict):
+
out_dict = metrics_dict.copy()
+
for k, v in out_dict.items():
+
if isinstance(v, np.ndarray):
+
out_dict[k] = v.tolist()
+
elif isinstance(v, np.float64):
+
out_dict[k] = float(v)
+
+
if self.paths.evaluation_results_yaml:
+
with open(self.paths.evaluation_results_yaml, "w") as f:
+
yaml.safe_dump(out_dict, f)
+
+
+
[docs]
+
def save_config_with_optimized_thresolds(self, config: DictConfig):
+
"""Save config the thresholds in the building_validation.application
+
part replaced by optimized thresholds"""
+
if "optimize" in self.todo:
+
optimized_cfg = config.copy()
+
optimized_cfg.building_validation.application.thresholds = self.thresholds
+
out_path = config.building_validation.optimization.paths.output_optimized_config
+
OmegaConf.save(config=optimized_cfg, f=out_path, resolve=False)
+