Update model_pipelineV2.py

model_pipelineV2.py

@@ -13,24 +13,53 @@ from langchain_core.runnables import RunnableLambda, RunnablePassthrough
 from operator import itemgetter
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 
+class VectorStoreSingleton:
+    _instance = None
+
+    @classmethod
+    def get_instance(cls):
+        if cls._instance is None:
+            cls._instance = create_vectorstore()  # Your existing function to create the vectorstore
+        return cls._instance
+
+class LanguageModelSingleton:
+    _instance = None
+
+    @classmethod
+    def get_instance(cls):
+        if cls._instance is None:
+            cls._instance = load_llm()  # Your existing function to load the LLM
+        return cls._instance
+
 
 class ModelPipeLine:
     DEFAULT_DOCUMENT_PROMPT = PromptTemplate.from_template(template="{page_content}")
     def __init__(self):
         self.curr_dir = os.path.dirname(__file__)
-        self.knowledge_dir = "knowledge"
-        print("Knowledge Directory:", self.knowledge_dir)
+        self.knowledge_dir = 'knowledge'
         self.prompt_dir = 'prompts'
         self.child_splitter = RecursiveCharacterTextSplitter(chunk_size=200)
         self.parent_splitter = RecursiveCharacterTextSplitter(chunk_size=500)
-        self.documents = process_pdf_document([os.path.join(self.knowledge_dir, 'depression_1.pdf'), os.path.join(self.knowledge_dir, 'depression_2.pdf')])
-        self.vectorstore, self.store = create_vectorstore()
-        self.retriever = rag_retriever(self.vectorstore, self.store, self.documents, self.parent_splitter, self.child_splitter) # Create the retriever
-        self.llm = load_llm() # Load the LLM model
-        self.memory = ConversationBufferMemory(return_messages=True, 
-                                               output_key="answer", 
-                                               input_key="question") # Instantiate ConversationBufferMemory
-
+        self._documents = None  # Initialize as None for lazy loading
+        self.vectorstore, self.store = VectorStoreSingleton.get_instance()
+        self._retriever = None  # Corrected: Initialize _retriever as None for lazy loading
+        self.llm = LanguageModelSingleton.get_instance()
+        self.memory = ConversationBufferMemory(return_messages=True, output_key="answer", input_key="question")
+
+    @property
+    def documents(self):
+        if self._documents is None:
+            self._documents = process_pdf_document([
+                os.path.join(self.knowledge_dir, 'depression_1.pdf'),
+                os.path.join(self.knowledge_dir, 'depression_2.pdf')
+            ])
+        return self._documents
+
+    @property
+    def retriever(self):
+        if self._retriever is None:
+            self._retriever = rag_retriever(self.vectorstore, self.store, self.documents, self.parent_splitter, self.child_splitter)
+        return self._retriever
     
     def get_prompts(self, system_file_path='system_prompt_template.txt', 
                     condense_file_path='condense_question_prompt_template.txt'):
@@ -95,15 +124,12 @@ class ModelPipeLine:
     def call_conversational_rag(self,question, chain):
         """
         Calls a conversational RAG (Retrieval-Augmented Generation) model to generate an answer to a given question.
-
         This function sends a question to the RAG model, retrieves the answer, and stores the question-answer pair in memory 
         for context in future interactions.
-
         Parameters:
         question (str): The question to be answered by the RAG model.
         chain (LangChain object): An instance of LangChain which encapsulates the RAG model and its functionality.
         memory (Memory object): An object used for storing the context of the conversation.
-
         Returns:
         dict: A dictionary containing the generated answer from the RAG model.
         """