Refactor class variables _LANGUAGES to global for MKQA SEACrowd#34

fhudi · Jan 31, 2024 · 76ca2e8 · 76ca2e8
1 parent 3d4808d
commit 76ca2e8
Showing 1 changed file with 9 additions and 10 deletions.
diff --git a/seacrowd/sea_datasets/mkqa/mkqa.py b/seacrowd/sea_datasets/mkqa/mkqa.py
@@ -64,6 +64,13 @@
 
 _SEACROWD_VERSION = "1.0.0"
 
+_LANGUAGES = [
+    "khm",
+    "msa",
+    "tha",
+    "vie",
+]  # follows the convention of 3-letter code as suggested since NusaCrowd.
+
 
 class MKQADataset(datasets.GeneratorBasedBuilder):
     """
@@ -76,14 +83,6 @@ class MKQADataset(datasets.GeneratorBasedBuilder):
     SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
     SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
 
-    LANGUAGES = [
-        "",
-        "khm",
-        "msa",
-        "tha",
-        "vie",
-    ]  # follows the convention of 3-letter code as suggested since NusaCrowd.
-
     _SOURCE_LANGUAGES = [
         "ar",
         "da",
@@ -129,7 +128,7 @@ class MKQADataset(datasets.GeneratorBasedBuilder):
                 schema="source",
                 subset_id=f"{_DATASETNAME}_{subset_lang}",
             )
-            for subset_lang in LANGUAGES
+            for subset_lang in ["", *_LANGUAGES]
         ],
         *[
             SEACrowdConfig(
@@ -139,7 +138,7 @@ class MKQADataset(datasets.GeneratorBasedBuilder):
                 schema="seacrowd_qa",
                 subset_id=f"{_DATASETNAME}_{subset_lang}",
             )
-            for subset_lang in LANGUAGES
+            for subset_lang in ["", *_LANGUAGES]
         ],
     ]