Bugfix of evaluate function in llama_index.py (#1223)

Dict in dataset of the evaluate function specified. Does not work -> Dataset and then check of "ground_truth" in column_names.
explodinggradients · Aug 27, 2024 · 7f1073f · 7f1073f
1 parent 9dc26e2
commit 7f1073f
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/src/ragas/integrations/llama_index.py b/src/ragas/integrations/llama_index.py
@@ -43,7 +43,7 @@ def validate_dataset(dataset: dict, metrics: list[Metric]):
 
 def evaluate(
     query_engine,
-    dataset: dict,
+    dataset: Dataset,
     metrics: list[Metric],
     llm: t.Optional[LlamaindexLLM] = None,
     embeddings: t.Optional[LlamaIndexEmbeddings] = None,
@@ -98,7 +98,7 @@ def evaluate(
             "answer": answers,
         }
     )
-    if "ground_truth" in dataset:
+    if "ground_truth" in dataset.column_names:
         hf_dataset = hf_dataset.add_column(
             name="ground_truth",
             column=dataset["ground_truth"],