From 373db747d807c3f2597269ac9abf50291673764d Mon Sep 17 00:00:00 2001
From: Roman Solomatin <samoed.roman@gmail.com>
Date: Fri, 13 Dec 2024 15:42:01 +0500
Subject: [PATCH] fix: Eval langs not correctly passed to monolingual tasks
 (#1587)

* fix SouthAfricanLangClassification.py

* add check for langs

* lint
---
 .../multilingual/HinDialectClassification.py  | 46 +++++++++----------
 .../SouthAfricanLangClassification.py         | 26 +++++------
 tests/test_TaskMetadata.py                    | 13 ++++++
 3 files changed, 49 insertions(+), 36 deletions(-)

diff --git a/mteb/tasks/Classification/multilingual/HinDialectClassification.py b/mteb/tasks/Classification/multilingual/HinDialectClassification.py
index 6565d4b71..c9d6b3666 100644
--- a/mteb/tasks/Classification/multilingual/HinDialectClassification.py
+++ b/mteb/tasks/Classification/multilingual/HinDialectClassification.py
@@ -3,29 +3,29 @@
 from mteb.abstasks.AbsTaskClassification import AbsTaskClassification
 from mteb.abstasks.TaskMetadata import TaskMetadata
 
-_LANGUAGES = {
-    "pan": ["pan-Guru"],
-    "bgc": ["bgc-Deva"],
-    "mag": ["mag-Deva"],
-    "bns": ["bns-Deva"],
-    "kfq": ["kfg-Deva"],
-    "noe": ["noe-Deva"],
-    "bhb": ["bhb-Deva"],
-    "bho": ["bho-Deva"],
-    "gbm": ["gbm-Deva"],
-    "mup": ["mup-Deva"],
-    "anp": ["anp-Deva"],
-    "hne": ["hne-Deva"],
-    "bra": ["bra-Deva"],
-    "raj": ["raj-Deva"],
-    "awa": ["awa-Deva"],
-    "guj": ["guj-Gujr"],
-    "ben": ["ben-Beng"],
-    "bhd": ["bhd-Deva"],
-    "kfy": ["kfy-Deva"],
-    "mar": ["mar-Deva"],
-    "bjj": ["bjj-Deva"],
-}
+_LANGUAGES = [
+    "pan-Guru",
+    "bgc-Deva",
+    "mag-Deva",
+    "bns-Deva",
+    "kfg-Deva",
+    "noe-Deva",
+    "bhb-Deva",
+    "bho-Deva",
+    "gbm-Deva",
+    "mup-Deva",
+    "anp-Deva",
+    "hne-Deva",
+    "bra-Deva",
+    "raj-Deva",
+    "awa-Deva",
+    "guj-Gujr",
+    "ben-Beng",
+    "bhd-Deva",
+    "kfy-Deva",
+    "mar-Deva",
+    "bjj-Deva",
+]
 
 
 class HinDialectClassification(AbsTaskClassification):
diff --git a/mteb/tasks/Classification/multilingual/SouthAfricanLangClassification.py b/mteb/tasks/Classification/multilingual/SouthAfricanLangClassification.py
index 4cef2c060..217d300ec 100644
--- a/mteb/tasks/Classification/multilingual/SouthAfricanLangClassification.py
+++ b/mteb/tasks/Classification/multilingual/SouthAfricanLangClassification.py
@@ -3,19 +3,19 @@
 from mteb.abstasks.AbsTaskClassification import AbsTaskClassification
 from mteb.abstasks.TaskMetadata import TaskMetadata
 
-_LANGUAGES = {
-    "afr": ["afr-Latn"],
-    "eng": ["eng-Latn"],
-    "nbl": ["nbl-Latn"],
-    "nso": ["nso-Latn"],
-    "sot": ["sot-Latn"],
-    "ssw": ["ssw-Latn"],
-    "tsn": ["tsn-Latn"],
-    "tso": ["tso-Latn"],
-    "ven": ["ven-Latn"],
-    "xho": ["xho-Latn"],
-    "zul": ["zul-Latn"],
-}
+_LANGUAGES = [
+    "afr-Latn",
+    "eng-Latn",
+    "nbl-Latn",
+    "nso-Latn",
+    "sot-Latn",
+    "ssw-Latn",
+    "tsn-Latn",
+    "tso-Latn",
+    "ven-Latn",
+    "xho-Latn",
+    "zul-Latn",
+]
 
 
 class SouthAfricanLangClassification(AbsTaskClassification):
diff --git a/tests/test_TaskMetadata.py b/tests/test_TaskMetadata.py
index 91ef4aabe..2b606c2c1 100644
--- a/tests/test_TaskMetadata.py
+++ b/tests/test_TaskMetadata.py
@@ -4,6 +4,7 @@
 
 import pytest
 
+from mteb import AbsTask
 from mteb.abstasks.TaskMetadata import TaskMetadata
 from mteb.overview import get_tasks
 
@@ -1095,3 +1096,15 @@ def test_empy_descriptive_stat_in_new_datasets():
             assert (
                 task.metadata.name not in exceptions
             ), f"Dataset {task.metadata.name} should have descriptive stats"
+
+
+@pytest.mark.parametrize("task", get_tasks())
+def test_eval_langs_correctly_specified(task: AbsTask):
+    if task.is_multilingual:
+        assert isinstance(
+            task.metadata.eval_langs, dict
+        ), f"{task.metadata.name} should have eval_langs as a dict"
+    else:
+        assert isinstance(
+            task.metadata.eval_langs, list
+        ), f"{task.metadata.name} should have eval_langs as a list"