VariantEffect · EstelleDa · Mar 5, 2025 · Mar 4, 2025 · Mar 4, 2025
diff --git a/src/mavedb/routers/score_sets.py b/src/mavedb/routers/score_sets.py
@@ -648,10 +648,15 @@ async def upload_score_set_variant_data(
     assert_permission(user_data, item, Action.UPDATE)
     assert_permission(user_data, item, Action.SET_SCORES)
 
-    scores_df = csv_data_to_df(scores_file.file)
-    counts_df = None
-    if counts_file and counts_file.filename:
-        counts_df = csv_data_to_df(counts_file.file)
+    try:
+        scores_df = csv_data_to_df(scores_file.file)
+        counts_df = None
+        if counts_file and counts_file.filename:
+            counts_df = csv_data_to_df(counts_file.file)
+    # Handle non-utf8 file problem.
+    except UnicodeDecodeError as e:
+        raise HTTPException(status_code=400, detail=f"Error decoding file: {e}. Ensure the file has correct values.")
+
 
     if scores_file:
         # Although this is also updated within the variant creation job, update it here

diff --git a/tests/routers/counts.csv b/tests/routers/counts.csv
diff --git a/tests/routers/counts_with_different_variants.csv b/tests/routers/counts_with_different_variants.csv
diff --git a/tests/routers/counts_with_hgvs_nt_and_pro.csv b/tests/routers/counts_with_hgvs_nt_and_pro.csv
diff --git a/tests/routers/counts_with_score.csv b/tests/routers/counts_with_score.csv
diff --git a/tests/routers/data/scores_non_utf8_encoded.csv b/tests/routers/data/scores_non_utf8_encoded.csv
diff --git a/tests/routers/scores.csv b/tests/routers/scores.csv
diff --git a/tests/routers/scores_hgvs_nt_not_match_pro.csv b/tests/routers/scores_hgvs_nt_not_match_pro.csv
diff --git a/tests/routers/scores_hgvs_pro_has_same_values.csv b/tests/routers/scores_hgvs_pro_has_same_values.csv
diff --git a/tests/routers/scores_with_duplicate_columns.csv b/tests/routers/scores_with_duplicate_columns.csv
diff --git a/tests/routers/scores_with_hgvs_nt_and_pro.csv b/tests/routers/scores_with_hgvs_nt_and_pro.csv
diff --git a/tests/routers/scores_with_invalid_hgvs_nt_prefix.csv b/tests/routers/scores_with_invalid_hgvs_nt_prefix.csv
diff --git a/tests/routers/scores_with_invalid_hgvs_pro_prefix.csv b/tests/routers/scores_with_invalid_hgvs_pro_prefix.csv
diff --git a/tests/routers/scores_with_nan_column_name.csv b/tests/routers/scores_with_nan_column_name.csv
diff --git a/tests/routers/scores_with_string.csv b/tests/routers/scores_with_string.csv
diff --git a/tests/routers/scores_without_hgvs_column.csv b/tests/routers/scores_without_hgvs_column.csv
diff --git a/tests/routers/scores_without_score_column.csv b/tests/routers/scores_without_score_column.csv
diff --git a/tests/routers/test_score_set.py b/tests/routers/test_score_set.py
@@ -2089,6 +2089,26 @@ def test_score_set_not_found_for_non_existent_score_set_when_adding_score_calibr
     assert "score_calibrations" not in response_data
 
 
+########################################################################################################################
+# Score set upload files
+########################################################################################################################
+
+# Not sure why scores_non_utf8_encoded.csv file has a wrong encoding problem, but it's good for this test.
+def test_upload_a_non_utf8_file(session, client, setup_router_db, data_files):
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set(client, experiment["urn"])
+    scores_csv_path = data_files / "scores_non_utf8_encoded.csv"
+    with open(scores_csv_path, "rb") as scores_file:
+        response = client.post(
+            f"/api/v1/score-sets/{score_set['urn']}/variants/data",
+            files={"scores_file": (scores_csv_path.name, scores_file, "text/csv")},
+        )
+    assert response.status_code == 400
+    response_data = response.json()
+    assert "Error decoding file: 'utf-8' codec can't decode byte 0xdd in position 10: invalid continuation byte. " \
+           "Ensure the file has correct values." in response_data["detail"]
+
+
 ########################################################################################################################
 # Score set download files
 ########################################################################################################################