PVNet 3.0.43__tar.gz → 3.0.45__tar.gz

{pvnet-3.0.43 → pvnet-3.0.45}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PVNet
-Version: 3.0.43
+Version: 3.0.45
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Description-Content-Type: text/markdown

{pvnet-3.0.43 → pvnet-3.0.45}/PVNet.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PVNet
-Version: 3.0.43
+Version: 3.0.45
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Description-Content-Type: text/markdown

pvnet-3.0.45/pvnet/__init__.py

@@ -0,0 +1,2 @@
+"""PVNet"""
+__version__ = "3.0.45"

{pvnet-3.0.43 → pvnet-3.0.45}/pvnet/models/base_model.py

@@ -79,6 +79,68 @@ def make_clean_data_config(input_path, output_path, placeholder="PLACEHOLDER"):
         yaml.dump(config, outfile, default_flow_style=False)
 
 
+def minimize_data_config(input_path, output_path, model):
+    """Strip out parts of the data config which aren't used by the model
+
+    Args:
+        input_path: Path to input datapipes configuration file
+        output_path: Location to save the output configuration file
+        model: The PVNet model object
+    """
+    with open(input_path) as cfg:
+        config = yaml.load(cfg, Loader=yaml.FullLoader)
+
+    if "nwp" in config["input_data"]:
+        if not model.include_nwp:
+            del config["input_data"]["nwp"]
+        else:
+            for nwp_source in config["input_data"]["nwp"].keys():
+                nwp_config = config["input_data"]["nwp"][nwp_source]
+
+                if nwp_source not in model.nwp_encoders_dict:
+                    # If not used, delete this source from the config
+                    del config["input_data"]["nwp"][nwp_source]
+                else:
+                    # Replace the image size
+                    nwp_pixel_size = model.nwp_encoders_dict[nwp_source].image_size_pixels
+                    nwp_config["nwp_image_size_pixels_height"] = nwp_pixel_size
+                    nwp_config["nwp_image_size_pixels_width"] = nwp_pixel_size
+
+                    # Replace the forecast minutes
+                    nwp_config["forecast_minutes"] = (
+                        model.nwp_encoders_dict[nwp_source].sequence_length
+                        - nwp_config["history_minutes"] / nwp_config["time_resolution_minutes"]
+                        - 1
+                    ) * nwp_config["time_resolution_minutes"]
+
+    if "satellite" in config["input_data"]:
+        if not model.include_sat:
+            del config["input_data"]["satellite"]
+        else:
+            sat_config = config["input_data"]["satellite"]
+
+            # Replace the image size
+            sat_pixel_size = model.sat_encoder.image_size_pixels
+            sat_config["satellite_image_size_pixels_height"] = sat_pixel_size
+            sat_config["satellite_image_size_pixels_width"] = sat_pixel_size
+
+            # Replace the satellite delay
+            sat_config["live_delay_minutes"] = model.min_sat_delay_minutes
+
+    if "pv" in config["input_data"]:
+        if not model.include_pv:
+            del config["input_data"]["pv"]
+
+    if "gsp" in config["input_data"]:
+        gsp_config = config["input_data"]["gsp"]
+
+        # Replace the forecast minutes
+        gsp_config["forecast_minutes"] = model.forecast_minutes
+
+    with open(output_path, "w") as outfile:
+        yaml.dump(config, outfile, default_flow_style=False)
+
+
 class PVNetModelHubMixin(PyTorchModelHubMixin):
     """
     Implementation of [`PyTorchModelHubMixin`] to provide model Hub upload/download capabilities.
@@ -207,7 +269,13 @@ class PVNetModelHubMixin(PyTorchModelHubMixin):
 
         # Save cleaned datapipes configuration file
         if data_config is not None:
-            make_clean_data_config(data_config, save_directory / DATA_CONFIG_NAME)
+            new_data_config_path = save_directory / DATA_CONFIG_NAME
+
+            # Replace the input filenames with place holders
+            make_clean_data_config(data_config, new_data_config_path)
+
+            # Taylor the data config to the model being saved
+            minimize_data_config(new_data_config_path, new_data_config_path, self)
 
         # Creating and saving model card.
         card_data = ModelCardData(language="en", license="mit", library_name="pytorch")
@@ -245,7 +313,7 @@ class PVNetModelHubMixin(PyTorchModelHubMixin):
 
 
 class BaseModel(pl.LightningModule, PVNetModelHubMixin):
-    """Abtstract base class for PVNet submodels"""
+    """Abstract base class for PVNet submodels"""
 
     def __init__(
         self,
@@ -257,6 +325,7 @@ class BaseModel(pl.LightningModule, PVNetModelHubMixin):
         interval_minutes: int = 30,
         timestep_intervals_to_plot: Optional[list[int]] = None,
         use_weighted_loss: bool = False,
+        forecast_minutes_ignore: Optional[int] = 0,
     ):
         """Abtstract base class for PVNet submodels.
 
@@ -270,6 +339,8 @@ class BaseModel(pl.LightningModule, PVNetModelHubMixin):
             interval_minutes: The interval in minutes between each timestep in the data
             timestep_intervals_to_plot: Intervals, in timesteps, to plot during training
             use_weighted_loss: Whether to use a weighted loss function
+            forecast_minutes_ignore: Number of forecast minutes to ignore when calculating losses.
+                For example if set to 60, the model doesnt predict the first 60 minutes
         """
         super().__init__()
 
@@ -292,10 +363,12 @@ class BaseModel(pl.LightningModule, PVNetModelHubMixin):
         self.forecast_minutes = forecast_minutes
         self.output_quantiles = output_quantiles
         self.interval_minutes = interval_minutes
+        self.forecast_minutes_ignore = forecast_minutes_ignore
 
         # Number of timestemps for 30 minutely data
         self.history_len = history_minutes // interval_minutes
-        self.forecast_len = forecast_minutes // interval_minutes
+        self.forecast_len = (forecast_minutes - forecast_minutes_ignore) // interval_minutes
+        self.forecast_len_ignore = forecast_minutes_ignore // interval_minutes
 
         self.weighted_losses = WeightedLosses(forecast_length=self.forecast_len)
 
@@ -334,7 +407,7 @@ class BaseModel(pl.LightningModule, PVNetModelHubMixin):
         y_median = y_quantiles[..., idx]
         return y_median
 
-    def _calculate_qauntile_loss(self, y_quantiles, y):
+    def _calculate_quantile_loss(self, y_quantiles, y):
         """Calculate quantile loss.
 
         Note:
@@ -366,7 +439,7 @@ class BaseModel(pl.LightningModule, PVNetModelHubMixin):
         losses = {}
 
         if self.use_quantile_regression:
-            losses["quantile_loss"] = self._calculate_qauntile_loss(y_hat, y)
+            losses["quantile_loss"] = self._calculate_quantile_loss(y_hat, y)
             y_hat = self._quantiles_to_prediction(y_hat)
 
         # calculate mse, mae

{pvnet-3.0.43 → pvnet-3.0.45}/pvnet/models/multimodal/multimodal.py

@@ -71,6 +71,7 @@ class Model(MultimodalBaseModel):
         timestep_intervals_to_plot: Optional[list[int]] = None,
         adapt_batches: Optional[bool] = False,
         use_weighted_loss: Optional[bool] = False,
+        forecast_minutes_ignore: Optional[int] = 0,
     ):
         """Neural network which combines information from different sources.
 
@@ -131,6 +132,8 @@ class Model(MultimodalBaseModel):
                 the model to use. This allows us to overprepare batches and slice from them for the
                 data we need for a model run.
             use_weighted_loss: Whether to use a weighted loss function
+            forecast_minutes_ignore: Number of forecast minutes to ignore when calculating losses.
+                For example if set to 60, the model doesnt predict the first 60 minutes
         """
 
         self.include_gsp_yield_history = include_gsp_yield_history
@@ -143,6 +146,7 @@ class Model(MultimodalBaseModel):
         self.embedding_dim = embedding_dim
         self.add_image_embedding_channel = add_image_embedding_channel
         self.interval_minutes = interval_minutes
+        self.min_sat_delay_minutes = min_sat_delay_minutes
         self.adapt_batches = adapt_batches
 
         super().__init__(
@@ -154,6 +158,7 @@ class Model(MultimodalBaseModel):
             interval_minutes=interval_minutes,
             timestep_intervals_to_plot=timestep_intervals_to_plot,
             use_weighted_loss=use_weighted_loss,
+            forecast_minutes_ignore=forecast_minutes_ignore,
         )
 
         # Number of features expected by the output_network
@@ -271,7 +276,8 @@ class Model(MultimodalBaseModel):
 
         if self.include_sun:
             self.sun_fc1 = nn.Linear(
-                in_features=2 * (self.forecast_len + self.history_len + 1),
+                in_features=2
+                * (self.forecast_len + self.forecast_len_ignore + self.history_len + 1),
                 out_features=16,
             )
 

pvnet-3.0.43/pvnet/__init__.py

@@ -1,2 +0,0 @@
-"""PVNet"""
-__version__ = "3.0.43"