From 79fb490c9152e5352b89ffa5d577b93400b0ad6a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Kara=C5=9B?= <mygnirut@gmail.com>
Date: Sun, 20 Apr 2025 19:39:02 +0200
Subject: [PATCH 1/2] Add metadata for GermanDPR and GermanQuAD

---
 mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py  | 14 +++++++-------
 mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py | 16 ++++++++--------
 pyproject.toml                                  |  2 +-
 tests/test_TaskMetadata.py                      |  2 +-
 4 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py b/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py
index a3118b8f73..a55ae4dffb 100644
--- a/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py
+++ b/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py
@@ -25,13 +25,13 @@ class GermanDPR(AbsTaskRetrieval):
         eval_splits=[_EVAL_SPLIT],
         eval_langs=["deu-Latn"],
         main_score="ndcg_at_10",
-        date=None,
-        domains=None,
-        task_subtypes=None,
-        license=None,
-        annotations_creators=None,
-        dialect=None,
-        sample_creation=None,
+        date=("2020-05-19", "2021-04-26"),
+        domains=["Written", "Non-fiction", "Web"],
+        task_subtypes=["Question answering"],
+        license="cc-by-4.0",
+        annotations_creators="human-annotated",
+        dialect=[],
+        sample_creation="found",
         bibtex_citation="""@misc{möller2021germanquad,
       title={GermanQuAD and GermanDPR: Improving Non-English Question Answering and Passage Retrieval},
       author={Timo Möller and Julian Risch and Malte Pietsch},
diff --git a/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py b/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py
index ba6a21e96e..2de3c2441b 100644
--- a/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py
+++ b/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py
@@ -31,7 +31,7 @@ class GermanQuADRetrieval(AbsTaskRetrieval):
     metadata = TaskMetadata(
         name="GermanQuAD-Retrieval",
         description="Context Retrieval for German Question Answering",
-        reference="https://www.kaggle.com/datasets/GermanQuAD",
+        reference="https://huggingface.co/datasets/deepset/germanquad",
         dataset={
             "path": "mteb/germanquad-retrieval",
             "revision": "f5c87ae5a2e7a5106606314eef45255f03151bb3",
@@ -42,13 +42,13 @@ class GermanQuADRetrieval(AbsTaskRetrieval):
         eval_splits=["test"],
         eval_langs=["deu-Latn"],
         main_score="mrr_at_5",
-        date=None,
-        domains=None,
-        task_subtypes=None,
-        license=None,
-        annotations_creators=None,
-        dialect=None,
-        sample_creation=None,
+        date=("2020-05-19", "2021-04-26"),
+        domains=["Written", "Non-fiction", "Web"],
+        task_subtypes=["Question answering"],
+        license="cc-by-4.0",
+        annotations_creators="human-annotated",
+        dialect=[],
+        sample_creation="found",
         bibtex_citation="""misc{möller2021germanquad,
       title={GermanQuAD and GermanDPR: Improving Non-English Question Answering and Passage Retrieval},
       author={Timo Möller and Julian Risch and Malte Pietsch},
diff --git a/pyproject.toml b/pyproject.toml
index 22357a45a2..c54c15473a 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mteb"
-version = "1.38.1"
+version = "1.38.2"
 description = "Massive Text Embedding Benchmark"
 readme = "README.md"
 authors = [
diff --git a/tests/test_TaskMetadata.py b/tests/test_TaskMetadata.py
index 75f3095a48..24cb80d1cb 100644
--- a/tests/test_TaskMetadata.py
+++ b/tests/test_TaskMetadata.py
@@ -53,7 +53,7 @@
     "TwitterHjerneRetrieval",
     "GerDaLIR",
     "GerDaLIRSmall",
-    "GermanDPR",
+    # "GermanDPR",
     "GermanQuAD-Retrieval",
     "LegalQuAD",
     "AILACasedocs",

From 9bfbef8962b98901b391df965391412bb3334d8f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20Kara=C5=9B?= <mygnirut@gmail.com>
Date: Sun, 20 Apr 2025 20:10:52 +0200
Subject: [PATCH 2/2] PR improvements

---
 pyproject.toml             | 2 +-
 tests/test_TaskMetadata.py | 2 --
 2 files changed, 1 insertion(+), 3 deletions(-)

diff --git a/pyproject.toml b/pyproject.toml
index c54c15473a..22357a45a2 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mteb"
-version = "1.38.2"
+version = "1.38.1"
 description = "Massive Text Embedding Benchmark"
 readme = "README.md"
 authors = [
diff --git a/tests/test_TaskMetadata.py b/tests/test_TaskMetadata.py
index 24cb80d1cb..3b5aa1e158 100644
--- a/tests/test_TaskMetadata.py
+++ b/tests/test_TaskMetadata.py
@@ -53,8 +53,6 @@
     "TwitterHjerneRetrieval",
     "GerDaLIR",
     "GerDaLIRSmall",
-    # "GermanDPR",
-    "GermanQuAD-Retrieval",
     "LegalQuAD",
     "AILACasedocs",
     "AILAStatutes",