fix: linting issues (#138)

Author: jjmachan

docs/integrations/langchain.ipynb

@@ -51,7 +51,7 @@
     "index = VectorstoreIndexCreator().from_loaders([loader])\n",
     "\n",
     "\n",
-    "llm = ChatOpenAI(temperature= 0)\n",
+    "llm = ChatOpenAI(temperature=0)\n",
     "qa_chain = RetrievalQA.from_chain_type(\n",
     "    llm,\n",
     "    retriever=index.vectorstore.as_retriever(),\n",
@@ -373,6 +373,7 @@
    ],
    "source": [
     "from langchain.schema import Document\n",
+    "\n",
     "fake_result = result.copy()\n",
     "fake_result[\"source_documents\"] = [Document(page_content=\"I love christmas\")]\n",
     "eval_result = context_recall_chain(fake_result)\n",

src/ragas/metrics/answer_relevance.py

@@ -7,8 +7,9 @@
 from datasets import Dataset
 from langchain.callbacks.manager import trace_as_chain_group
 from langchain.embeddings import OpenAIEmbeddings
-from langchain.prompts import ChatPromptTemplate, HumanMessagePromptTemplate
 from langchain.embeddings.base import Embeddings
+from langchain.prompts import ChatPromptTemplate, HumanMessagePromptTemplate
+
 from ragas.metrics.base import EvaluationMode, MetricWithLLM
 from ragas.metrics.llms import generate
 
@@ -45,7 +46,7 @@ class AnswerRelevancy(MetricWithLLM):
         Here indicates the number questions generated per answer.
         Ideal range between 3 to 5.
     embeddings: Embedding
-        The langchain wrapper of Embedding object. 
+        The langchain wrapper of Embedding object.
         E.g. HuggingFaceEmbeddings('BAAI/bge-base-en')
     """
 

tests/benchmarks/benchmark_eval.py

@@ -2,7 +2,12 @@
 from torch.cuda import is_available
 
 from ragas import evaluate
-from ragas.metrics import answer_relevancy, context_relevancy, faithfulness
+from ragas.metrics import (
+    answer_relevancy,
+    context_recall,
+    context_relevancy,
+    faithfulness,
+)
 from ragas.metrics.critique import harmfulness
 
 DEVICE = "cuda" if is_available() else "cpu"
@@ -13,6 +18,12 @@
 if __name__ == "__main__":
     result = evaluate(
         ds.select(range(5)),
-        metrics=[answer_relevancy, context_relevancy, faithfulness, harmfulness],
+        metrics=[
+            answer_relevancy,
+            context_relevancy,
+            faithfulness,
+            harmfulness,
+            context_recall,
+        ],
     )
     print(result)