Merge branch 'master' into GH1237-seed-random

Author: jonbrenas

malariagen_data/adar1.py

@@ -181,7 +181,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Data releases available
                         </th>
-                        <td>{', '.join(self._available_releases)}</td>
+                        <td>{", ".join(self._available_releases)}</td>
                     </tr>
                     <tr>
                         <th style="text-align: left">
@@ -229,7 +229,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Relevant data releases
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                 </tbody>
             </table>

malariagen_data/adir1.py

@@ -181,7 +181,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Data releases available
                         </th>
-                        <td>{', '.join(self._available_releases)}</td>
+                        <td>{", ".join(self._available_releases)}</td>
                     </tr>
                     <tr>
                         <th style="text-align: left">
@@ -229,7 +229,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Relevant data releases
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                 </tbody>
             </table>

malariagen_data/af1.py

@@ -183,7 +183,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Data releases available
                         </th>
-                        <td>{', '.join(self._available_releases)}</td>
+                        <td>{", ".join(self._available_releases)}</td>
                     </tr>
                     <tr>
                         <th style="text-align: left">
@@ -231,7 +231,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Relevant data releases
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                 </tbody>
             </table>

malariagen_data/ag3.py

@@ -278,7 +278,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Data releases available
                         </th>
-                        <td>{', '.join(self._available_releases)}</td>
+                        <td>{", ".join(self._available_releases)}</td>
                     </tr>
                     <tr>
                         <th style="text-align: left">
@@ -332,7 +332,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Relevant data releases
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                 </tbody>
             </table>

malariagen_data/amin1.py

@@ -181,7 +181,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Data releases available
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                     <tr>
                         <th style="text-align: left">
@@ -229,7 +229,7 @@ def _repr_html_(self):
                         <th style="text-align: left">
                             Relevant data releases
                         </th>
-                        <td>{', '.join(self.releases)}</td>
+                        <td>{", ".join(self.releases)}</td>
                     </tr>
                 </tbody>
             </table>

malariagen_data/anoph/base.py

@@ -135,7 +135,7 @@ def __init__(
             storage_options = dict()
         try:
             self._fs, self._base_path = _init_filesystem(self._url, **storage_options)
-        except Exception as exc:  # pragma: no cover
+        except (OSError, ImportError) as exc:  # pragma: no cover
             raise IOError(
                 "An error occurred establishing a connection to the storage system. Please see the nested exception for more details."
             ) from exc
@@ -144,7 +144,7 @@ def __init__(
         try:
             with self.open_file(self._config_path) as f:
                 self._config = json.load(f)
-        except Exception as exc:  # pragma: no cover
+        except (OSError, json.JSONDecodeError) as exc:  # pragma: no cover
             if (isinstance(exc, OSError) and "forbidden" in str(exc).lower()) or (
                 getattr(exc, "status", None) == 403
             ):

malariagen_data/anoph/base_params.py

@@ -10,6 +10,7 @@
     single_contig_param_type,
     single_region_param_type,
     chunks_param_type,
+    Region,
 )
 
 contig: TypeAlias = Annotated[
@@ -47,6 +48,8 @@
     """,
 ]
 
+regions_tuple: TypeAlias = Tuple[Region, ...]
+
 release: TypeAlias = Annotated[
     Union[str, Sequence[str]],
     "Release version identifier.",
@@ -65,6 +68,8 @@
     """,
 ]
 
+sample_sets_tuple: TypeAlias = Tuple[sample_set, ...]
+
 sample_query: TypeAlias = Annotated[
     str,
     """
@@ -94,6 +99,8 @@
     """,
 ]
 
+sample_indices_tuple: TypeAlias = Tuple[int, ...]
+
 sample: TypeAlias = Annotated[
     Union[str, int],
     "Sample identifier or index within sample set.",

malariagen_data/anoph/cnv_frq.py

@@ -632,6 +632,15 @@ def _gene_cnv_frequencies_advanced(
             columns=["gene_id", "gene_name", "cnv_type"],
         )
 
+        debug("sort variants for deterministic ordering")
+        sort_index = df_variants.sort_values(
+            ["contig", "start", "cnv_type"]
+        ).index.values
+        df_variants = df_variants.iloc[sort_index].reset_index(drop=True)
+        count = count[sort_index]
+        nobs = nobs[sort_index]
+        frequency = frequency[sort_index]
+
         debug("build the output dataset")
         ds_out = xr.Dataset()
 

malariagen_data/anoph/map_params.py

@@ -42,7 +42,7 @@ def get_basemap_abbrevs() -> dict:
         for key, provider_fn in _basemap_abbrev_candidates.items():
             try:
                 _basemap_abbrevs[key] = provider_fn()
-            except Exception:
+            except (ImportError, AttributeError):
                 warnings.warn(
                     f"Basemap provider {key!r} is not available and will be skipped.",
                     stacklevel=2,

malariagen_data/anoph/sample_metadata.py

@@ -1864,7 +1864,7 @@ def _locate_cohorts(*, cohorts, data, min_cohort_size):
         for coh, query in cohorts.items():
             try:
                 loc_coh = data.eval(query).values
-            except Exception as e:
+            except (KeyError, NameError, SyntaxError, TypeError, AttributeError) as e:
                 raise ValueError(
                     f"Invalid query for cohort {coh!r}: {query!r}. Error: {e}"
                 ) from e