add inspect (#261)

* add inspect first version

* run pre-commit

* update specs for inspect

* add more tests for logging

* add more tests for logging

* update test

openeo_processes_dask/process_implementations/__init__.py

@@ -5,6 +5,7 @@
 from .arrays import *
 from .comparison import *
 from .cubes import *
+from .inspect import *
 from .logic import *
 from .math import *
 

openeo_processes_dask/process_implementations/inspect.py

@@ -0,0 +1,27 @@
+import logging
+from typing import Any, Optional
+
+__all__ = ["inspect"]
+
+logger = logging.getLogger(__name__)
+
+
+def inspect(
+    data: Any,
+    message: Optional[str] = "",
+    code: Optional[str] = "User",
+    level: Optional[str] = "info",
+) -> Any:
+    if level.lower() == "info":
+        logger.info(f"{code}: {message} {data}.")
+
+    elif level.lower() == "warning":
+        logger.warning(f"{code}: {message} {data}.")
+
+    elif level.lower() == "error":
+        logger.error(f"{code}: {message} {data}.")
+
+    elif level.lower() == "debug":
+        logger.debug(f"{code}: {message} {data}.")
+
+    return data

tests/test_inspect.py

@@ -0,0 +1,45 @@
+import logging
+
+import numpy as np
+import pytest
+import xarray as xr
+
+from openeo_processes_dask.process_implementations.inspect import inspect
+from tests.mockdata import create_fake_rastercube
+
+
+@pytest.mark.parametrize("size", [(6, 5, 4, 4)])
+@pytest.mark.parametrize("dtype", [np.float32])
+def test_inspect_results(temporal_interval, bounding_box, random_raster_data):
+    raster_cube = create_fake_rastercube(
+        data=random_raster_data,
+        spatial_extent=bounding_box,
+        temporal_extent=temporal_interval,
+        bands=["B02", "B03", "B04", "B08"],
+    )
+
+    assert inspect(np.zeros(1)) == np.zeros(1)
+    xr.testing.assert_equal(raster_cube, inspect(raster_cube))
+
+
+def test_inspect_logs(caplog):
+    with caplog.at_level(logging.INFO):
+        result = inspect(data=1, message="First log", code="Logging")
+        assert result == 1
+        assert "Logging: First log 1." in caplog.text
+
+    with caplog.at_level(logging.WARNING):
+        result = inspect(data=1, message="First log", level="warning")
+        assert result == 1
+        assert "User: First log 1." in caplog.text
+
+    with caplog.at_level(logging.ERROR):
+        result = inspect(data=1, level="error")
+        assert result == 1
+        assert "User:  1." in caplog.text
+
+    with caplog.at_level(logging.DEBUG):
+        result = inspect(data=1, level="debug")
+        assert result == 1
+        assert "User:  1." in caplog.text
+        assert "User:  2." not in caplog.text

tests/test_ml.py

@@ -49,26 +49,6 @@ def test_fit_regr_random_forest_inline_geojson(
     assert isinstance(model, xgb.core.Booster)
 
 
-def test_fit_regr_random_forest_xvec(vector_data_cube, dask_client):
-    predictors_vars = ["value2"]
-    target_var = "value1"
-
-    predictors = (
-        xr.Dataset.from_dataframe(vector_data_cube.set_index("geometry", append=True))
-        .xvec.set_geom_indexes("geometry", crs=DEFAULT_CRS)
-        .to_array(dim="bands")
-    )
-
-    model = fit_regr_random_forest(
-        predictors=predictors,
-        target=vector_data_cube,
-        target_var=target_var,
-        predictors_vars=predictors_vars,
-    )
-
-    assert isinstance(model, xgb.core.Booster)
-
-
 @pytest.mark.parametrize("size", [(6, 5, 4, 3)])
 @pytest.mark.parametrize("dtype", [np.float64])
 def test_curve_fitting(temporal_interval, bounding_box, random_raster_data):