Update astropy to 4.0.2 to fix fits 1d issue

environment.yml

@@ -5,7 +5,6 @@ channels:
   - default
 dependencies:
   - python=3.7
-  - astropy
   - numpy
   - ipython
   - jupyter
@@ -21,6 +20,7 @@ dependencies:
   - sphinx_rtd_theme
   - pip
   - pip:
-      - nbsphinx
-      - sphinx_automodapi
-      - uproot~=3.0
+    - astropy>=4.0.2,<5
+    - nbsphinx
+    - sphinx_automodapi
+    - uproot~=3.0

pyirf/tests/test_eventdisplay_comparison.py

@@ -36,21 +36,11 @@ def test_eventdisplay_example(caplog):
     caplog.set_level(logging.WARNING, logger="fits_schema")
 
     # check that each HDU respects the OGADF schema
-    AEFF_2D.validate_hdu(output_hdul["EFFECTIVE_AREA"], onerror="log")
-    EDISP_2D.validate_hdu(output_hdul["ENERGY_DISPERSION"], onerror="log")
-    PSF_TABLE.validate_hdu(output_hdul["PSF"], onerror="log")
-    BKG_2D.validate_hdu(output_hdul["BACKGROUND"], onerror="log")
-    RAD_MAX.validate_hdu(output_hdul["RAD_MAX"], onerror="log")
-
-    errors_to_ignore = {
-        # error due to astropy bug, which should be fixed in 4.0.2.
-        # TODO: remove when we require astropy >= 4.0.2
-        "Dimensionality of rows is 0, should be 1",
-    }
-
-    assert all(rec.message in errors_to_ignore for rec in caplog.records)
-
-    # compare to reference
+    AEFF_2D.validate_hdu(output_hdul["EFFECTIVE_AREA"], onerror="raise")
+    EDISP_2D.validate_hdu(output_hdul["ENERGY_DISPERSION"], onerror="raise")
+    PSF_TABLE.validate_hdu(output_hdul["PSF"], onerror="raise")
+    BKG_2D.validate_hdu(output_hdul["BACKGROUND"], onerror="raise")
+    RAD_MAX.validate_hdu(output_hdul["RAD_MAX"], onerror="raise")
 
     f = uproot.open(ROOT_DIR / "data" / IRF_FILE)
     sensitivity_ed = f['DiffSens'] * u.Unit('erg s-1 cm-2')

pyirf/version.py

@@ -1 +1 @@
-__version__ = "0.3.0"
+__version__ = "0.3.0.post1"

setup.py

@@ -29,7 +29,7 @@
     packages=find_packages(),
     include_package_data=True,
     install_requires=[
-        "astropy~=4.0",
+        "astropy~=4.0,>=4.0.2",
         "matplotlib",
         "numpy>=1.18",
         "pandas",