Just raise and error if training data not in altaz

ctapipe/reco/sklearn.py

@@ -9,7 +9,7 @@
 import astropy.units as u
 import joblib
 import numpy as np
-from astropy.coordinates import AltAz, SkyCoord
+from astropy.coordinates import AltAz
 from astropy.table import QTable, Table, vstack
 from astropy.utils.decorators import lazyproperty
 from sklearn.metrics import accuracy_score, r2_score, roc_auc_score
@@ -22,7 +22,6 @@
 
 from ..containers import (
     ArrayEventContainer,
-    CoordinateFrameType,
     DispContainer,
     ParticleClassificationContainer,
     ReconstructedEnergyContainer,
@@ -778,34 +777,13 @@ def predict_table(self, key, table: Table) -> Dict[ReconstructionProperty, Table
         fov_lon = table["hillas_fov_lon"].quantity + disp * np.cos(psi)
         fov_lat = table["hillas_fov_lat"].quantity + disp * np.sin(psi)
 
-        self.log.warning("FIXME: Assuming constant and parallel pointing for each run")
-        if np.all(table["subarray_pointing_frame"] is CoordinateFrameType.ALTAZ):
-            pointing_alt = table["subarray_pointing_lat"]
-            pointing_az = table["subarray_pointing_lon"]
-        elif np.all(table["subarray_pointing_frame"] is CoordinateFrameType.ICRS):
-            pointing_altaz = SkyCoord(
-                ra=table["subarray_pointing_lon"],
-                dec=table["subarray_pointing_lat"],
-                frame="icrs",
-            ).transform_to(AltAz())
-            pointing_alt = pointing_altaz.alt
-            pointing_az = pointing_altaz.az
-        elif np.all(table["subarray_pointing_frame"] is CoordinateFrameType.GALACTIC):
-            pointing_altaz = SkyCoord(
-                l=table["subarray_pointing_lon"],
-                b=table["subarray_pointing_lat"],
-                frame="galactic",
-            ).transform_to(AltAz())
-            pointing_alt = pointing_altaz.alt
-            pointing_az = pointing_altaz.az
-        else:
-            raise KeyError("Unknown observation coordinate frame")
-
+        # FIXME: Assume constant and parallel pointing for each run
+        self.log.warning("Assuming constant and parallel pointing for each run")
         alt, az = telescope_to_horizontal(
             lon=fov_lon,
             lat=fov_lat,
-            pointing_alt=pointing_alt,
-            pointing_az=pointing_az,
+            pointing_alt=table["subarray_pointing_lat"],
+            pointing_az=table["subarray_pointing_lon"],
         )
 
         altaz_result = Table(

ctapipe/tools/train_disp_reconstructor.py

@@ -1,6 +1,7 @@
 import astropy.units as u
 import numpy as np
 
+from ctapipe.containers import CoordinateFrameType
 from ctapipe.core import Tool
 from ctapipe.core.traits import Bool, Int, IntTelescopeParameter, Path
 from ctapipe.exceptions import TooFewEvents
@@ -104,6 +105,13 @@ def start(self):
             self.log.info("Loading events for %s", tel_type)
             table = self._read_table(tel_type)
 
+            if not np.all(
+                table["subarray_pointing_frame"] == CoordinateFrameType.ALTAZ.value
+            ):
+                raise ValueError(
+                    "Pointing information for training data has to be provided in horizontal coordinates"
+                )
+
             self.log.info("Train models on %s events", len(table))
             self.cross_validate(tel_type, table)
 

docs/changes/2431.bugfix.rst

@@ -1,2 +1,2 @@
-Check the coordinate frame in which the array pointing is given
-before using it in ``DispReconstructor``.
+Check that the array pointing is given in horizontal coordinates
+before training a ``DispReconstructor``.