eegdash 0.0.1__py3-none-any.whl

eegdash/SignalStore/tests/data/valid_data/data_arrays/make_fake_data.py

@@ -0,0 +1,59 @@
+"""This module generates fake data for testing purposes."""
+
+import numpy as np
+import xarray as xr
+import json
+import pathlib
+
+
+def make_fake_data():
+    # get records from tests/data/records.json
+    data_dir = pathlib.Path(__file__).parent.parent
+    records_path = data_dir / "records.json"
+    data_arrays_dir = data_dir / "data_arrays"
+    with open(records_path) as f:
+        records = json.load(f)
+
+    # try making one xarray from scratch
+    da = xr.DataArray(
+        name="test",
+        data=np.random.rand(10, 10),
+        dims=("x", "y"),
+        attrs={"schema_ref": "test", "data_name": "test"},
+    )
+    da_path = data_arrays_dir / "test__test.nc"
+    da.to_netcdf(da_path)
+
+    # generate fake data
+    for record in records:
+        if record.get("has_file") == True:
+            # make xarray from record, generating numpy random floats to go in a numpy arrray of shape record["shape"]
+            array = np.random.rand(*record["shape"])
+            dataarray = xr.DataArray(
+                name=record["schema_ref"] + "__" + record["data_name"],
+                data=array,
+                dims=tuple(record["data_dimensions"]),
+            )
+            # every field other than "shape", "data_dimensions" and "data_name" is an attribute
+            for field, value in record.items():
+                dataarray.attrs[field] = adapt(value)
+
+            # save xarray to netcdf file
+            print(f"Saving {dataarray.name} to {data_arrays_dir}")
+            schema_ref, data_name = record["schema_ref"], record["data_name"]
+            data_path = data_arrays_dir / f"{schema_ref}__{data_name}.nc"
+            dataarray.to_netcdf(data_path)
+
+def adapt(value):
+    """Adapt value to be json serializable."""
+    # boolians are strings
+    if isinstance(value, bool):
+        return str(value)
+    if isinstance(value, dict):
+        return json.dumps(value)
+    else:
+        return value
+
+
+if __name__ == "__main__":
+    make_fake_data()