Merge pull request #3427 from snbianco/filter-case

Author: bsipocz

CHANGES.rst

@@ -25,6 +25,8 @@ mast
 
 - Raise an error if non-string values are passed to ``utils.resolve_object``. [#3435]
 
+- Filtering by file extension or by a string column is now case-insensitive in ``MastMissions.filter_products``
+  and ``Observations.filter_products``. [#3427]
 
 
 Infrastructure, Utility and Other Changes and Additions

astroquery/mast/missions.py

@@ -517,12 +517,7 @@ def filter_products(self, products, *, extension=None, **filters):
 
         # Filter by file extension, if provided
         if extension:
-            extensions = [extension] if isinstance(extension, str) else extension
-            ext_mask = np.array(
-                [not isinstance(x, np.ma.core.MaskedConstant) and any(x.endswith(ext) for ext in extensions)
-                 for x in products["filename"]],
-                dtype=bool
-            )
+            ext_mask = utils.apply_extension_filter(products, extension, 'filename')
             filter_mask &= ext_mask
 
         # Apply column-based filters

astroquery/mast/observations.py

@@ -594,12 +594,7 @@ def filter_products(self, products, *, mrp_only=False, extension=None, **filters
 
         # Filter by file extension, if provided
         if extension:
-            extensions = [extension] if isinstance(extension, str) else extension
-            ext_mask = np.array(
-                [not isinstance(x, np.ma.core.MaskedConstant) and any(x.endswith(ext) for ext in extensions)
-                 for x in products["productFilename"]],
-                dtype=bool
-            )
+            ext_mask = utils.apply_extension_filter(products, extension, 'productFilename')
             filter_mask &= ext_mask
 
         # Apply column-based filters

astroquery/mast/tests/test_mast.py

@@ -428,7 +428,7 @@ def test_missions_filter_products(patch_post):
     assert all(~((filtered['size'] >= 14400) & (filtered['size'] <= 17280)))
 
     # Negate a string match
-    filtered = mast.MastMissions.filter_products(products, category='!CALIBRATED')
+    filtered = mast.MastMissions.filter_products(products, category='!calibrated')
     assert all(filtered['category'] != 'CALIBRATED')
 
     # Negate one string in a list
@@ -802,7 +802,7 @@ def test_observations_get_product_list(patch_post):
 def test_observations_filter_products(patch_post):
     products = mast.Observations.get_product_list('2003738726')
     filtered = mast.Observations.filter_products(products,
-                                                 productType=["SCIENCE"],
+                                                 productType=["sCiEnCE"],
                                                  mrp_only=False)
     assert isinstance(filtered, Table)
     assert len(filtered) == 7
@@ -812,8 +812,10 @@ def test_observations_filter_products(patch_post):
     assert all(filtered['productGroupDescription'] == 'Minimum Recommended Products')
 
     # Filter by extension
-    filtered = mast.Observations.filter_products(products, extension='fits')
+    filtered = mast.Observations.filter_products(products, extension='FITS')
     assert len(filtered) > 0
+    filtered = mast.Observations.filter_products(products, extension=['png'])
+    assert len(filtered) == 0
 
     # Numeric filtering
     filtered = mast.Observations.filter_products(products, size='<50000')

astroquery/mast/utils.py

@@ -456,6 +456,38 @@ def remove_duplicate_products(data_products, uri_key):
     return unique_products
 
 
+def apply_extension_filter(products, extension, filename_key):
+    """
+    Applies an extension filter to a product table.
+
+    Parameters
+    ----------
+    products : `~astropy.table.Table`
+        The product table to filter.
+    extension : str
+        The extension to filter by (e.g., 'fits', 'csv').
+    filename_key : str
+        The column name representing the filename of a product.
+
+    Returns
+    -------
+    ext_mask : `numpy.ndarray`
+        A boolean mask indicating which rows of the product table have the specified extension.
+    """
+    # Normalize extensions to lowercase
+    extensions = [extension] if isinstance(extension, str) else extension
+    extensions = tuple(ext.lower() for ext in extensions)
+
+    # Build mask
+    ext_mask = np.array(
+        [not isinstance(x, np.ma.core.MaskedConstant)
+         and str(x).lower().endswith(extensions)
+         for x in products[filename_key]],
+        dtype=bool
+    )
+    return ext_mask
+
+
 def _combine_positive_negative_masks(mask_funcs):
     """
     Combines a list of mask functions into a single mask according to:
@@ -582,7 +614,9 @@ def apply_column_filters(products, filters):
                 v = val[1:] if is_negated else val
 
                 def func(col, v=v):
-                    return np.isin(col, [v])
+                    # Normalize both column values and filter to lowercase strings for case-insensitive comparison
+                    col_lower = np.char.lower(col.astype(str))
+                    return np.isin(col_lower, [v.lower()])
                 mask_funcs.append((func, is_negated))
 
             this_mask = _combine_positive_negative_masks(mask_funcs)(col_data)

docs/mast/mast_cut.rst

@@ -12,9 +12,9 @@ Processing Operation's Center (`SPOC <https://archive.stsci.edu/missions-and-dat
 The cutouts are returned in the form of target pixel files that follow the same format as TESS pipeline target
 pixel files. This tool can be accessed in Astroquery by using the Tesscut class.
 
-As of August 2025, the option to create cutouts using TESS Image CAlibration 
-(`TICA <https://ui.adsabs.harvard.edu/abs/2020RNAAS...4..251F/abstract>`__) full frame images 
-was discontinued. Individual TICA full frame images remain available from the 
+As of August 2025, the option to create cutouts using TESS Image CAlibration
+(`TICA <https://ui.adsabs.harvard.edu/abs/2020RNAAS...4..251F/abstract>`__) full frame images
+was discontinued. Individual TICA full frame images remain available from the
 `MAST TICA homepage <https://archive.stsci.edu/hlsp/tica>`__. Cutouts using SPOC data remain available through TESSCut.
 
 **Note:** TESScut limits each user to no more than 10 simultaneous calls to the service.
@@ -97,7 +97,7 @@ simply with either the objectname or coordinates.
      2  APERTURE      1 ImageHDU        97   (2136, 2078)   int32
 
 
-The `~astroquery.mast.TesscutClass.download_cutouts` function takes a coordinate, cutout size 
+The `~astroquery.mast.TesscutClass.download_cutouts` function takes a coordinate, cutout size
 (in pixels or an angular quantity), or object name (e.g. "M104" or "TIC 32449963") and moving target
 (True or False). It uses these parameters to download the cutout target pixel file(s).
 
@@ -170,6 +170,7 @@ The following example requests SPOC cutouts for a moving target.
    tess-s0043-3-3     43      3   3
    tess-s0044-2-4     44      2   4
    tess-s0092-4-3     92      4   3
+   tess-s0097-1-4     97      1   4
 
 
 Zcut
@@ -237,7 +238,7 @@ If a given coordinate appears in more than one Zcut survey, a cutout will be pro
    Downloading URL https://mast.stsci.edu/zcut/api/v0.1/astrocut?ra=189.49206&dec=62.20615&y=200&x=300&units=px&format=jpg to ./zcut_20201202132453.zip ... [Done]
    ...
    >>> print(manifest)
-                                                  Local Path                                              
+                                                  Local Path
    -------------------------------------------------------------------------------------------------------
       ./hlsp_3dhst_spitzer_irac_goods-n_irac1_v4.0_sc_189.492060_62.206150_10.0pix-x-5.0pix_astrocut_0.jpg
    ./hlsp_3dhst_spitzer_irac_goods-n-s2_irac3_v4.0_sc_189.492060_62.206150_10.0pix-x-5.0pix_astrocut_0.jpg