Merge pull request #966 from adilraza99/GH965-g123-cache-key-fix

Normalize site_mask in g123_gwss cache parameters and bump cache version

Author: jonbrenas

malariagen_data/anoph/g123.py

@@ -170,7 +170,7 @@ def g123_gwss(
     ) -> Tuple[np.ndarray, np.ndarray]:
         # Change this name if you ever change the behaviour of this function, to
         # invalidate any previously cached data.
-        name = "g123_gwss_v1"
+        name = "g123_gwss_v2"
 
         valid_sites = self.phasing_analysis_ids + ("all", "segregating")
         if sites not in valid_sites:
@@ -181,7 +181,7 @@ def g123_gwss(
         params = dict(
             contig=contig,
             sites=sites,
-            site_mask=site_mask,
+            site_mask=self._prep_optional_site_mask_param(site_mask=site_mask),
             window_size=window_size,
             sample_sets=self._prep_sample_sets_param(sample_sets=sample_sets),
             # N.B., do not be tempted to convert this sample query into integer

tests/anoph/test_g123.py

@@ -240,8 +240,7 @@ def test_g123_calibration(fixture, api: AnophelesG123Analysis):
 
     # Set up test parameters.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    window_sizes = np.random.randint(100, 500, size=random.randint(2, 5)).tolist()
-    window_sizes = sorted([int(x) for x in window_sizes])
+    window_sizes = sorted(random.sample(range(100, 500), k=random.randint(2, 5)))
     g123_params = dict(
         contig=random.choice(api.contigs),
         sites=random.choice(api.phasing_analysis_ids),