urielplus 1.3__tar.gz → 1.3.2__tar.gz

{urielplus-1.3/urielplus.egg-info → urielplus-1.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: urielplus
-Version: 1.3
+Version: 1.3.2
 Summary: URIEL+: Knowledge base for natural language processing
 Home-page: https://github.com/Masonshipton25/URIELPlus
 Author: Mason Shipton
@@ -256,26 +256,43 @@ The core package is tested on Python 3.10, 3.11, and 3.12. It does not require T
 
 ## Database Integration Examples
 
++ NOTE: `integrate_databases()` integrates **all** available databases, including eWAVE and APiCS, which you may not want. eWAVE adds hundreds of features that only apply to English dialects, and APiCS adds hundreds of features but has data for only 76 languages. Both can add substantial sparsity and features you may not need. To exclude them, integrate only the databases you want, either individually or in one call (see "Integrating Only Some Databases" below).
+
++ Integrating All Databases:
+    ```python
+    u.integrate_databases()
+    ```
 + Integrating One Database:
     ```python
     u.integrate_{database}()
     ```
-+ Integrating Some Databases:
++ Integrating Only Some Databases (Excluding Others):
     ```python
     u.integrate_custom_databases({databases})
     ```
-+ Integrating All Databases:
+
++ Example: Integrating Every Database Except eWAVE and APiCS
     ```python
-    u.integrate_databases()
+    # Option 1: one call
+    u.integrate_custom_databases("UPDATED_SAPHON", "BDPROTO", "GRAMBANK", "GLOTTOLOG")
+
+    # Option 2: one database at a time
+    u.integrate_saphon()
+    u.integrate_bdproto()
+    u.integrate_grambank()
+    u.integrate_glottolog()
     ```
+
 + Set Language Codes to Glottocodes:
     ```python
     u.set_glottocodes()
     ```
+
 + Reset all changes:
     ```python
     u.reset()
     ```
+
 + NOTE: `reset()` only resets URIEL+'s in-memory attributes back to their initial state. It does not change or delete any files that were already written to disk when caching was enabled. If you want those cached files reset as well, you will need to remove or replace them manually.
 
 + Replace `{database}` with `saphon`, `bdproto`, `grambank`, `apics`, `ewave`, or `glottolog`.

{urielplus-1.3 → urielplus-1.3.2}/README.md

@@ -208,26 +208,43 @@ The core package is tested on Python 3.10, 3.11, and 3.12. It does not require T
 
 ## Database Integration Examples
 
++ NOTE: `integrate_databases()` integrates **all** available databases, including eWAVE and APiCS, which you may not want. eWAVE adds hundreds of features that only apply to English dialects, and APiCS adds hundreds of features but has data for only 76 languages. Both can add substantial sparsity and features you may not need. To exclude them, integrate only the databases you want, either individually or in one call (see "Integrating Only Some Databases" below).
+
++ Integrating All Databases:
+    ```python
+    u.integrate_databases()
+    ```
 + Integrating One Database:
     ```python
     u.integrate_{database}()
     ```
-+ Integrating Some Databases:
++ Integrating Only Some Databases (Excluding Others):
     ```python
     u.integrate_custom_databases({databases})
     ```
-+ Integrating All Databases:
+
++ Example: Integrating Every Database Except eWAVE and APiCS
     ```python
-    u.integrate_databases()
+    # Option 1: one call
+    u.integrate_custom_databases("UPDATED_SAPHON", "BDPROTO", "GRAMBANK", "GLOTTOLOG")
+
+    # Option 2: one database at a time
+    u.integrate_saphon()
+    u.integrate_bdproto()
+    u.integrate_grambank()
+    u.integrate_glottolog()
     ```
+
 + Set Language Codes to Glottocodes:
     ```python
     u.set_glottocodes()
     ```
+
 + Reset all changes:
     ```python
     u.reset()
     ```
+
 + NOTE: `reset()` only resets URIEL+'s in-memory attributes back to their initial state. It does not change or delete any files that were already written to disk when caching was enabled. If you want those cached files reset as well, you will need to remove or replace them manually.
 
 + Replace `{database}` with `saphon`, `bdproto`, `grambank`, `apics`, `ewave`, or `glottolog`.

{urielplus-1.3 → urielplus-1.3.2}/THIRD_PARTY_NOTICES.md

@@ -1,6 +1,6 @@
 # Third-party notices
 
-URIELPlus v1.3 is distributed under the Creative Commons Attribution-ShareAlike 4.0 International licence in `LICENSE.txt`. The source tables incorporated into the released database retain the separate licences stated below. The licences for those source tables are not replaced by the URIELPlus licence.
+URIELPlus v1.3.2 is distributed under the Creative Commons Attribution-ShareAlike 4.0 International licence in `LICENSE.txt`. The source tables incorporated into the released database retain the separate licences stated below. The licences for those source tables are not replaced by the URIELPlus licence.
 
 | Source material | Locked version | Files containing derived values | Attribution | Source licence | Primary licence record |
 | --- | --- | --- | --- | --- | --- |

{urielplus-1.3 → urielplus-1.3.2}/setup.py

@@ -5,7 +5,7 @@ with open("README.md", "r", encoding="utf-8") as fh:
 
 setup(
     name="urielplus",
-    version="1.3",
+    version="1.3.2",
     author="Mason Shipton",
     author_email="masonshipton25@gmail.com",
     description="URIEL+: Knowledge base for natural language processing",

{urielplus-1.3 → urielplus-1.3.2}/urielplus/urielplus_imputation.py

@@ -165,14 +165,14 @@ class URIELPlusImputation(BaseURIEL):
                                 child_idx = glottocode_to_idx.get(child_glottocode)
                                 if child_idx is not None and child_idx not in visited:
                                     child_features = aggregated_data[child_idx]
-                                    before = child_features.copy()
 
                                     fill_mask = (child_features == -1.0) & (parent_features > -1.0)
-                                    updated_child_features = np.where(fill_mask, parent_features, child_features)
 
-                                    changed_feats = np.where(updated_child_features != before)[0]
-                                    for feat_idx in changed_feats:
-                                        self.lineage_imputed_indices.add((child_idx, feat_idx))
+                                    if fill_mask.any():
+                                        aggregated_data[child_idx] = np.where(fill_mask, parent_features, child_features)
+
+                                        for feat_idx in np.where(fill_mask)[0]:
+                                            self.lineage_imputed_indices.add((child_idx, int(feat_idx)))
 
                                     visited.add(child_idx)
                                     queue.append(child_idx)

{urielplus-1.3 → urielplus-1.3.2/urielplus.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: urielplus
-Version: 1.3
+Version: 1.3.2
 Summary: URIEL+: Knowledge base for natural language processing
 Home-page: https://github.com/Masonshipton25/URIELPlus
 Author: Mason Shipton
@@ -256,26 +256,43 @@ The core package is tested on Python 3.10, 3.11, and 3.12. It does not require T
 
 ## Database Integration Examples
 
++ NOTE: `integrate_databases()` integrates **all** available databases, including eWAVE and APiCS, which you may not want. eWAVE adds hundreds of features that only apply to English dialects, and APiCS adds hundreds of features but has data for only 76 languages. Both can add substantial sparsity and features you may not need. To exclude them, integrate only the databases you want, either individually or in one call (see "Integrating Only Some Databases" below).
+
++ Integrating All Databases:
+    ```python
+    u.integrate_databases()
+    ```
 + Integrating One Database:
     ```python
     u.integrate_{database}()
     ```
-+ Integrating Some Databases:
++ Integrating Only Some Databases (Excluding Others):
     ```python
     u.integrate_custom_databases({databases})
     ```
-+ Integrating All Databases:
+
++ Example: Integrating Every Database Except eWAVE and APiCS
     ```python
-    u.integrate_databases()
+    # Option 1: one call
+    u.integrate_custom_databases("UPDATED_SAPHON", "BDPROTO", "GRAMBANK", "GLOTTOLOG")
+
+    # Option 2: one database at a time
+    u.integrate_saphon()
+    u.integrate_bdproto()
+    u.integrate_grambank()
+    u.integrate_glottolog()
     ```
+
 + Set Language Codes to Glottocodes:
     ```python
     u.set_glottocodes()
     ```
+
 + Reset all changes:
     ```python
     u.reset()
     ```
+
 + NOTE: `reset()` only resets URIEL+'s in-memory attributes back to their initial state. It does not change or delete any files that were already written to disk when caching was enabled. If you want those cached files reset as well, you will need to remove or replace them manually.
 
 + Replace `{database}` with `saphon`, `bdproto`, `grambank`, `apics`, `ewave`, or `glottolog`.

{urielplus-1.3 → urielplus-1.3.2}/urielplus/urielplus_databases.py

@@ -154,10 +154,10 @@ class URIELPlusDatabases(BaseURIEL):
         for row, language in enumerate(languages):
             if language not in selected.index:
                 continue
+            data[row, :, 0] = 0
             parts = self._lineage_parts(selected.at[language, "lineage"])
             if not parts:
                 continue
-            data[row, :, 0] = 0
             indices = [path_index[parts[:depth]] for depth in range(1, len(parts) + 1)]
             data[row, indices, 0] = 1