Remove from YAML config. #153

Author: JackKelly

gcp.yaml

@@ -1,5 +1,4 @@
 general:
-  cloud: gcp
   description: example configuration
   name: example
 input_data:

nowcasting_dataset/config/gcp.yaml

@@ -1,7 +1,6 @@
 """ Configuration model for the dataset """
 from datetime import datetime
 from typing import Optional
-import logging
 
 import git
 from pathy import Pathy
@@ -21,25 +20,19 @@ class General(BaseModel):
 
     name: str = Field("example", description="The name of this configuration file.")
     description: str = Field(
-        "example configuration", description="Description of this confgiruation file"
-    )
-    cloud: str = Field(
-        "gcp",
-        description=(
-            "local, gcp, or aws.  Deprecated.  Will be removed when issue"
-            " https://github.com/openclimatefix/nowcasting_dataset/issues/153"
-            " is implemented"
-        ),
+        "example configuration", description="Description of this configuration file"
     )
 
 
 class Git(BaseModel):
     """Git model"""
 
-    hash: str = Field(..., description="The git hash has for when a dataset is created.")
-    message: str = Field(..., description="The git message has for when a dataset is created.")
+    hash: str = Field(
+        ..., description="The git hash of nowcasting_dataset when a dataset is created."
+    )
+    message: str = Field(..., description="The git message for when a dataset is created.")
     committed_date: datetime = Field(
-        ..., description="The git datestamp has for when a dataset is created."
+        ..., description="The git datestamp for when a dataset is created."
     )
 
 
@@ -55,7 +48,7 @@ class DataSourceMixin(BaseModel):
     history_minutes: int = Field(
         None,
         ge=0,
-        description="how many historic minutes are used. "
+        description="how many historic minutes to use. "
         "If set to None, the value is defaulted to InputData.default_history_minutes",
     )
 

nowcasting_dataset/config/model.py

@@ -1,15 +1,16 @@
 #!/usr/bin/env python3
 
-"""Pre-prepares batches of data on Google Cloud Storage.
+"""Pre-prepares batches of data.
 
 Usage:
 
-First, manually create the GCS directories given by the constants
+First, manually create the directories given by the constants
 DST_TRAIN_PATH and DST_VALIDATION_PATH, and create the
 LOCAL_TEMP_PATH.  Note that all files will be deleted from
 LOCAL_TEMP_PATH when this script starts up.
 
-Currently caluclating azimuth and elevation angles, takes about 15 mins for 2548 PV systems, for about 1 year
+Currently caluclating azimuth and elevation angles, takes about 15 mins for 2548 PV systems,
+for about 1 year.
 
 """
 from nowcasting_dataset.filesystem import utils
@@ -75,7 +76,6 @@
 LOCAL_TEMP_PATH = Path(config.process.local_temp_path).expanduser()
 
 UPLOAD_EVERY_N_BATCHES = config.process.upload_every_n_batches
-CLOUD = config.general.cloud  # either gcp or aws
 
 # Necessary to avoid "RuntimeError: receieved 0 items of ancdata".  See:
 # https://discuss.pytorch.org/t/runtimeerror-received-0-items-of-ancdata/4999/2

scripts/prepare_ml_data.py

@@ -48,7 +48,6 @@
         maximum_batch_id_train,
         f"gs://{DST_TRAIN_PATH}",
         LOCAL_TEMP_PATH,
-        cloud="gcp",
         configuration=config,
     ),
     **dataloader_config,
@@ -60,7 +59,6 @@
         maximum_batch_id_train,
         f"gs://{DST_VALIDATION_PATH}",
         LOCAL_TEMP_PATH,
-        cloud="gcp",
         configuration=config,
     ),
     **dataloader_config,
@@ -71,7 +69,6 @@
         maximum_batch_id_train,
         f"gs://{DST_TEST_PATH}",
         LOCAL_TEMP_PATH,
-        cloud="gcp",
         configuration=config,
     ),
     **dataloader_config,

scripts/validate_ml_data.py

@@ -1,5 +1,4 @@
 general:
-  cloud: gcp
   description: example configuration
   name: example
 git: null

tests/config/nwp_size_test.yaml

@@ -1,5 +1,4 @@
 general:
-  cloud: gcp
   description: example configuration
   name: example
 git: null