Remove data folder

All the data operations/loading will now be done in nowcasting-dataloader or nowcasting-utils

satflow/core/utils.py

@@ -1,22 +1,5 @@
-import logging
-import typing as Dict
 from nowcasting_dataset.config.load import load_yaml_configuration
-
 import yaml
-
-
-def load_config(file_path: str) -> Dict:
-    with open(file_path, "r") as f:
-        config = yaml.load(f)
-    return config
-
-
-def make_logger(name: str, level=logging.DEBUG) -> logging.Logger:
-    logger = logging.getLogger(name)
-    logger.setLevel(level=level)
-    return logger
-
-
 import logging
 import warnings
 from typing import List, Sequence
@@ -28,6 +11,12 @@ def make_logger(name: str, level=logging.DEBUG) -> logging.Logger:
 from pytorch_lightning.utilities import rank_zero_only
 
 
+def make_logger(name: str, level=logging.DEBUG) -> logging.Logger:
+    logger = logging.getLogger(name)
+    logger.setLevel(level=level)
+    return logger
+
+
 def get_logger(name=__name__, level=logging.INFO) -> logging.Logger:
     """Initializes multi-GPU-friendly python logger."""
 
@@ -211,3 +200,8 @@ def log_hyperparameters(
     # this is just a trick to prevent trainer from logging hparams of model,
     # since we already did that above
     trainer.logger.log_hyperparams = empty
+
+
+def load_config(config_file):
+    with open(config_file, "r") as cfg:
+        return yaml.load(cfg, Loader=yaml.FullLoader)["config"]

satflow/models/joint_model.py

@@ -355,7 +355,13 @@ def configure_optimizers(self):
         return {"optimizer": optimizer, "lr_scheduler": lr_dict}
 
     def construct_query(self, x: dict) -> Tuple[torch.Tensor, torch.Tensor]:
-        return self.query(x), self.pv_query(x)
+        sat_fourier_features = x[SATELLITE_DATA + "_position_encoding"][
+            :, :, : -self.forecast_steps
+        ]  # Only want future steps
+        gsp_fourier_features = x[GSP_YIELD + "_position_encoding"][
+            :, :, : -self.gsp_forecast_steps
+        ]  # Only want future features
+        return self.query(x, sat_fourier_features), self.pv_query(x, gsp_fourier_features)
 
     def forward(self, x, mask=None, query=None):
         return self.model.forward(x, mask=mask, queries=query)