Update app.py

app.py

@@ -1,61 +1,126 @@
 import gradio as gr
-from langchain_community.document_loaders import CSVLoader  # Changed import
-from langchain_community.vectorstores import FAISS  # Changed import
-from langchain.prompts import PromptTemplate
-from langchain.chains import RetrievalQA
-from langchain.llms import HuggingFaceLLM  # Adjusted for correct instantiation
-import warnings
-from huggingface_hub import login
+
+
 import os
-from transformers import pipeline
+os.environ["hftoken"] = hftoken
+
+from langchain_huggingface import HuggingFaceEndpoint
+
+repo_id = "mistralai/Mistral-7B-Instruct-v0.3"
+llm = HuggingFaceEndpoint(repo_id = repo_id, max_new_tokens = 128, temperature = 0.7, huggingfacehub_api_token = hftoken)
 
-# Initialize the LLM using pipeline
-llm = pipeline("text-generation", model="meta-llama/Meta-Llama-3-8B-Instruct")  # Adjusted initialization
+from langchain_core.output_parsers import StrOutputParser
+from langchain_core.prompts import ChatPromptTemplate
 
-# Load CSV file
-loader = CSVLoader(file_path='aiotsmartlabs_faq.csv', source_column='prompt')
+prompt = ChatPromptTemplate.from_template("tell me a joke about {topic}")
+chain = prompt | llm | StrOutputParser()
+
+from langchain.document_loaders.csv_loader import CSVLoader
+
+loader = CSVLoader(file_path='aiotsmartlabs_faq.csv', source_column = 'prompt')
 data = loader.load()
 
-# Suppress warnings
-warnings.filterwarnings("ignore", category=UserWarning, message="TypedStorage is deprecated")
-warnings.filterwarnings("ignore", category=FutureWarning, message="`resume_download` is deprecated")
+from langchain_huggingface import HuggingFaceEmbeddings
+from langchain_chroma import Chroma
+from langchain_huggingface.embeddings import HuggingFaceEndpointEmbeddings
 
-# Embedding model
-model_name = "BAAI/bge-m3"
-instructor_embeddings = HuggingFaceLLM(model_name=model_name)  # Adjusted for correct instantiation
+# CHECK MTEB LEADERBOARD & FIND BEST EMBEDDING MODEL
+model = "BAAI/bge-m3"
+embeddings = HuggingFaceEndpointEmbeddings(model = model)
 
-# Create FAISS vector store from documents
-vectordb = FAISS.from_documents(documents=data, embedding=instructor_embeddings)
-retriever = vectordb.as_retriever()
+vectorstore = Chroma.from_documents(documents = data, embedding = embeddings)
+retriever = vectorstore.as_retriever()
+
+# from langchain.prompts import PromptTemplate
+
+from langchain_core.prompts import ChatPromptTemplate
+
+prompt = ChatPromptTemplate.from_template("""Given the following context and a question, generate an answer based on the context only.
 
-# Define the prompt template
-prompt_template = """Given the following context and a question, generate an answer based on the context only.
 In the answer try to provide as much text as possible from "response" section in the source document context without making much changes.
 If somebody asks "Who are you?" or a similar phrase, state "I am Rishi's assistant built using a Large Language Model!"
 If the answer is not found in the context, kindly state "I don't know. Please ask Rishi on Discord. Discord Invite Link: https://discord.gg/6ezpZGeCcM. Or email at [email protected]" Don't try to make up an answer.
+
 CONTEXT: {context}
-QUESTION: {question}"""
 
-PROMPT = PromptTemplate(
-    template=prompt_template, input_variables=["context", "question"]
-)
+QUESTION: {question}""")
+
+from langchain_core.runnables import RunnablePassthrough
 
-# Initialize the RetrievalQA chain
-chain = RetrievalQA.from_chain_type(llm=llm,  # Adjusted initialization
-                                    chain_type="stuff",
-                                    retriever=retriever,
-                                    input_key="query",
-                                    return_source_documents=True,
-                                    chain_type_kwargs={"prompt": PROMPT})
+rag_chain = (
+    {"context": retriever, "question": RunnablePassthrough()}
+    | prompt
+    | llm
+    | StrOutputParser()
+)
 
 # Define the chat response function
 def chatresponse(message, history):
-    output = chain(message)
-    return output['result']
+    output = rag_chain.invoke(message)
+    response = output.split('ANSWER: ')[-1].strip()
+    print(response)
 
 # Launch the Gradio chat interface
 gr.ChatInterface(chatresponse).launch()
 
+# import gradio as gr
+# from langchain_community.document_loaders import CSVLoader  # Changed import
+# from langchain_community.vectorstores import FAISS  # Changed import
+# from langchain.prompts import PromptTemplate
+# from langchain.chains import RetrievalQA
+# from langchain.llms import HuggingFaceLLM  # Adjusted for correct instantiation
+# import warnings
+# from huggingface_hub import login
+# import os
+# from transformers import pipeline
+
+# # Initialize the LLM using pipeline
+# llm = pipeline("text-generation", model="meta-llama/Meta-Llama-3-8B-Instruct")  # Adjusted initialization
+
+# # Load CSV file
+# loader = CSVLoader(file_path='aiotsmartlabs_faq.csv', source_column='prompt')
+# data = loader.load()
+
+# # Suppress warnings
+# warnings.filterwarnings("ignore", category=UserWarning, message="TypedStorage is deprecated")
+# warnings.filterwarnings("ignore", category=FutureWarning, message="`resume_download` is deprecated")
+
+# # Embedding model
+# model_name = "BAAI/bge-m3"
+# instructor_embeddings = HuggingFaceLLM(model_name=model_name)  # Adjusted for correct instantiation
+
+# # Create FAISS vector store from documents
+# vectordb = FAISS.from_documents(documents=data, embedding=instructor_embeddings)
+# retriever = vectordb.as_retriever()
+
+# # Define the prompt template
+# prompt_template = """Given the following context and a question, generate an answer based on the context only.
+# In the answer try to provide as much text as possible from "response" section in the source document context without making much changes.
+# If somebody asks "Who are you?" or a similar phrase, state "I am Rishi's assistant built using a Large Language Model!"
+# If the answer is not found in the context, kindly state "I don't know. Please ask Rishi on Discord. Discord Invite Link: https://discord.gg/6ezpZGeCcM. Or email at [email protected]" Don't try to make up an answer.
+# CONTEXT: {context}
+# QUESTION: {question}"""
+
+# PROMPT = PromptTemplate(
+#     template=prompt_template, input_variables=["context", "question"]
+# )
+
+# # Initialize the RetrievalQA chain
+# chain = RetrievalQA.from_chain_type(llm=llm,  # Adjusted initialization
+#                                     chain_type="stuff",
+#                                     retriever=retriever,
+#                                     input_key="query",
+#                                     return_source_documents=True,
+#                                     chain_type_kwargs={"prompt": PROMPT})
+
+# # Define the chat response function
+# def chatresponse(message, history):
+#     output = chain(message)
+#     return output['result']
+
+# # Launch the Gradio chat interface
+# gr.ChatInterface(chatresponse).launch()
+
 
 # import gradio as gr
 # # from langchain.llms import GooglePalm