Starter LLM Inference Call

app/hybrid_rag.py

@@ -56,12 +56,36 @@ class HybridJiraRAG:
             allow_dangerous_deserialization=True
         )
         
+        # Create prompt
+        prompt = PromptTemplate(
+            template="Context: {context}\n\nQuestion: {question}\n\nAnswer:",
+            input_variables=["context", "question"]
+        )
+
+        # Format docs function
+        def format_docs(docs):
+            return "\n\n".join([doc.page_content for doc in docs])
+
+        # LCEL chain
+        retriever = self.vector_store.as_retriever(search_kwargs={"k": 5})
+
+        self.rag_chain = (
+            {
+                "context": retriever | format_docs,
+                "question": RunnablePassthrough()
+            }
+            | prompt
+            | self.llm
+            | StrOutputParser()
+        )
+
         # RAG chain
+        '''
         self.rag_chain = RetrievalQA.from_chain_type(
             llm=self.llm,
             retriever=self.vector_store.as_retriever(search_kwargs={"k": 5}),
             return_source_documents=True
-        )
+        )'''
     
     def _load_local_llm(self, model_name: str):
         """Load LLM locally to use GPU"""