Fixes #169, and adds functions to drop cache entries matching regex

In some cases for 2D DATA CDFs, the 2nd axis has a DEPEND0 while it
also is Non Record Vary. Speasy has to consider that those axes are not
time dependent.

Signed-off-by: Alexis Jeandet <[email protected]>

pyproject.toml

@@ -33,6 +33,7 @@ classifiers = [
 ]
 dependencies = ['requests', 'pandas', 'diskcache', 'appdirs', 'numpy', 'packaging', 'python-dateutil',
     'astropy', 'astroquery', 'pyistp>=0.7.0', 'tqdm', 'matplotlib', 'urllib3>=1.26.0', "PyYAML", "scipy"]
+
 [project.urls]
 homepage = "https://github.com/SciQLop/speasy"
 

speasy/core/cache/__init__.py

@@ -1,14 +1,21 @@
+from typing import Union
+import re
 from .cache import Cache, CacheItem
 from ._function_cache import CacheCall
 from ._providers_caches import CACHE_ALLOWED_KWARGS, Cacheable, UnversionedProviderCache
 from ._instance import _cache
+import logging
+
+log = logging.getLogger(__name__)
 
 
 def cache_len():
+    """Return the number of items in the cache"""
     return len(_cache)
 
 
 def cache_disk_size():
+    """Return the size of the cache on disk"""
     return _cache.disk_size()
 
 
@@ -17,16 +24,70 @@ def stats():
 
 
 def entries():
+    """Return all cache entries as a list of keys
+
+    Returns
+    -------
+    list
+        A list of all cache keys
+    """
     return _cache.keys()
 
 
 def add_item(key, item, expires=None):
+    """Add an item to the cache with an optional expiration time in seconds
+
+    Parameters
+    ----------
+    key : str
+        The key under which the item will be stored
+    item : any
+        The item to be stored
+    expires : int, optional
+        The expiration time in seconds, by default None (no expiration)
+    """
     _cache.set(key, item, expires)
 
 
 def drop_item(key):
+    """Drop an item from the cache by key
+
+    Parameters
+    ----------
+    key : str
+        The key of the item to be dropped
+    """
     _cache.drop(key)
 
 
 def get_item(key, default_value=None):
+    """Get an item from the cache by key
+
+    Parameters
+    ----------
+    key : str
+        The key of the item to be retrieved
+    default_value : any, optional
+        The default value to return if the key does not exist, by default None
+
+    Returns
+    -------
+    any
+        The item stored under the key or the default value if the key does not exist
+    """
     return _cache.get(key, default_value)
+
+
+def drop_matching_entries(pattern: Union[str, re.Pattern]):
+    """Drop all cache entries that match a given pattern
+
+    Parameters
+    ----------
+    pattern : str or re.Pattern
+        The pattern to match cache keys against
+    """
+    if isinstance(pattern, str):
+        pattern = re.compile(pattern)
+    for key in filter(pattern.match, _cache.keys()):
+        log.debug(f"Dropping cache entry {key}")
+        _cache.drop(key)

speasy/core/codecs/bundled_codecs/istp_cdf.py

@@ -38,9 +38,9 @@ def _fix_attributes_types(attributes: dict):
 
 def _is_time_dependent(axis, time_axis_name):
     if axis.attributes.get('DEPEND_TIME', '') == time_axis_name:
-        return True
+        return not axis.is_nrv
     if axis.attributes.get('DEPEND_0', '') == time_axis_name:
-        return True
+        return not axis.is_nrv
     return False
 
 

tests/test_cdaweb.py

@@ -232,6 +232,12 @@ def test_get_virtual_parameter_always_falls_back_to_api(self):
             "2021-11-3", "2021-11-3T01", disable_proxy=True, disable_cache=True, method="FILE")
         self.assertIsNotNone(mms1_fgm_b_bcs_srvy_l2_clean)
 
+    def test_wrong_time_dependency_axis(self):
+        result = spz.get_data(
+            "cda/MMS1_FEEPS_SRVY_L2_ELECTRON/mms1_epd_feeps_srvy_l2_electron_bottom_intensity_sensorid_2",
+            datetime(2018, 5, 26, 1, 0, 0), datetime(2018, 5, 26, 1, 10, 1))
+        self.assertIsNotNone(result)
+
 
 @ddt
 class DirectArchiveConverter(unittest.TestCase):

tests/test_common.py

@@ -8,6 +8,7 @@
 from ddt import ddt, data, unpack
 
 import speasy as spz
+from speasy.core.cache import entries
 
 
 @ddt
@@ -85,5 +86,33 @@ def test_epoch_to_dt64(self, input, expected):
             np.allclose(spz.core.epoch_to_datetime64(input).astype('int64'), expected.astype('int64'), atol=1e-10))
 
 
+class TestDroppingKeysFromCache(unittest.TestCase):
+
+    def test_dropping_specific_key(self):
+        spz.core.cache.add_item("key1", "value1")
+        spz.core.cache.add_item("key2", "value2")
+        spz.core.cache.add_item("key3", "value3")
+
+        spz.core.cache.drop_item("key2")
+
+        self.assertEqual(spz.core.cache.get_item("key1"), "value1")
+        self.assertIsNone(spz.core.cache.get_item("key2"))
+        self.assertEqual(spz.core.cache.get_item("key3"), "value3")
+        spz.core.cache.drop_matching_entries("key\\d")
+
+    def test_dropping_matching_keys(self):
+        spz.core.cache.add_item("key1", "value1")
+        spz.core.cache.add_item("key2", "value2")
+        spz.core.cache.drop_matching_entries("key\\d")
+        self.assertIsNone(spz.core.cache.get_item("key1"))
+        self.assertIsNone(spz.core.cache.get_item("key2"))
+
+    def test_dropping_speasy_variables(self):
+        spz.get_data(spz.inventories.tree.ssc.Trajectories.ace, "2008-01-01", "2008-01-02")
+        self.assertGreater(len(list(filter(lambda e: e.startswith("ssc_orbits/ace/gse"), spz.core.cache.entries()))), 0)
+        spz.core.cache.drop_matching_entries("^ssc_orbits/ace/gse")
+        self.assertEqual(len(list(filter(lambda e: e.startswith("ssc_orbits/ace/gse"), spz.core.cache.entries()))), 0)
+
+
 if __name__ == '__main__':
     unittest.main()