Update app.py

app.py

@@ -1,5 +1,4 @@
 import gradio as gr
-from langchain.schema import AIMessage, HumanMessage
 
 
 import os
@@ -13,8 +12,8 @@ llm = HuggingFaceEndpoint(repo_id = repo_id, max_new_tokens = 128, temperature =
 from langchain_core.output_parsers import StrOutputParser
 from langchain_core.prompts import ChatPromptTemplate
 
-# prompt = ChatPromptTemplate.from_template("tell me a joke about {topic}")
-# chain = prompt | llm | StrOutputParser()
+prompt = ChatPromptTemplate.from_template("tell me a joke about {topic}")
+chain = prompt | llm | StrOutputParser()
 
 # from langchain.document_loaders.csv_loader import CSVLoader
 from langchain_community.document_loaders.csv_loader import CSVLoader
@@ -31,55 +30,129 @@ from langchain_huggingface.embeddings import HuggingFaceEndpointEmbeddings
 model = "BAAI/bge-m3"
 embeddings = HuggingFaceEndpointEmbeddings(model = model)
 
+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.
+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 rishi@aiotsmartlabs.com" Don't try to make up an answer.
+CONTEXT: {context}
+HISTORY: {history}
+QUESTION: {question}""")
+
+from langchain_core.runnables import RunnablePassthrough
+
+rag_chain = (
+    {"context": retriever, "history": RunnablePassthrough(), "question": RunnablePassthrough()}
+    | prompt
+    | llm
+    | StrOutputParser()
+)
 
 # Define the chat response function
 def chatresponse(message, history):
-    # history_langchain_format = []
-    # for human, ai in history:
-    #     history_langchain_format.append(HumanMessage(content=human))
-    #     history_langchain_format.append(AIMessage(content=ai))
-    # history_langchain_format.append(HumanMessage(content=message))
+    history_text = "\n".join([f"User: {h[0]}\nAssistant: {h[1]}" for h in history])
+    
+    inputs = {
+        "context": "",  # context will be retrieved by the retriever
+        "history": history_text,
+        "question": message
+    }
+    
+    output = rag_chain.invoke(inputs)
+    response = output.split('ANSWER: ')[-1].strip()
+    return response
+
+# Launch the Gradio chat interface
+gr.ChatInterface(chatresponse).launch()
+
+# import gradio as gr
+# from langchain.schema import AIMessage, HumanMessage
+
 
-    data_vectorstore = Chroma.from_documents(documents = data, embedding = embeddings)
-    # history_vectorstore = Chroma.from_documents(documents = history, embedding = embeddings)
-    # vectorstore = data_vectorstore + history_vectorstore
-    vectorstore = data_vectorstore
-    retriever = vectorstore.as_retriever()
+# import os
+# hftoken = os.environ["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)
+
+# from langchain_core.output_parsers import StrOutputParser
+# from langchain_core.prompts import ChatPromptTemplate
 
-    history_str = "\n".join([f"Human: {h[0]}\nAI: {h[1]}" for h in history])
+# # prompt = ChatPromptTemplate.from_template("tell me a joke about {topic}")
+# # chain = prompt | llm | StrOutputParser()
+
+# # from langchain.document_loaders.csv_loader import CSVLoader
+# from langchain_community.document_loaders.csv_loader import CSVLoader
+
+
+# loader = CSVLoader(file_path='aiotsmartlabs_faq.csv', source_column = 'prompt')
+# data = loader.load()
 
-    # from langchain.prompts import PromptTemplate
+# from langchain_huggingface import HuggingFaceEmbeddings
+# from langchain_chroma import Chroma
+# from langchain_huggingface.embeddings import HuggingFaceEndpointEmbeddings
+
+# # CHECK MTEB LEADERBOARD & FIND BEST EMBEDDING MODEL
+# model = "BAAI/bge-m3"
+# embeddings = HuggingFaceEndpointEmbeddings(model = model)
+
+
+
+# # Define the chat response function
+# def chatresponse(message, history):
+#     # history_langchain_format = []
+#     # for human, ai in history:
+#     #     history_langchain_format.append(HumanMessage(content=human))
+#     #     history_langchain_format.append(AIMessage(content=ai))
+#     # history_langchain_format.append(HumanMessage(content=message))
+
+#     data_vectorstore = Chroma.from_documents(documents = data, embedding = embeddings)
+#     # history_vectorstore = Chroma.from_documents(documents = history, embedding = embeddings)
+#     # vectorstore = data_vectorstore + history_vectorstore
+#     vectorstore = data_vectorstore
+#     retriever = vectorstore.as_retriever()
+
+#     history_str = "\n".join([f"Human: {h[0]}\nAI: {h[1]}" for h in history])
+
+#     # from langchain.prompts import PromptTemplate
     
-    from langchain_core.prompts import ChatPromptTemplate
+#     from langchain_core.prompts import ChatPromptTemplate
     
-    prompt = ChatPromptTemplate.from_template("""Given the following history, context and a question, generate an answer based on the context only.
+#     prompt = ChatPromptTemplate.from_template("""Given the following history, 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 rishi@aiotsmartlabs.com" Don't try to make up an answer.
+#     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 rishi@aiotsmartlabs.com" Don't try to make up an answer.
     
-    HISTORY: {history}
+#     HISTORY: {history}
     
-    CONTEXT: {context}
+#     CONTEXT: {context}
     
-    QUESTION: {question}""")
+#     QUESTION: {question}""")
     
-    from langchain_core.runnables import RunnablePassthrough
-    rag_chain = (
-    {"history": history_str, "context": retriever, "question": RunnablePassthrough()}
-    | prompt
-    | llm
-    | StrOutputParser()
-    )
+#     from langchain_core.runnables import RunnablePassthrough
+#     rag_chain = (
+#     {"history": history_str, "context": retriever, "question": RunnablePassthrough()}
+#     | prompt
+#     | llm
+#     | StrOutputParser()
+#     )
 
     
-    output = rag_chain.invoke(message)
-    response = output.split('ANSWER: ')[-1].strip()
-    return response
+#     output = rag_chain.invoke(message)
+#     response = output.split('ANSWER: ')[-1].strip()
+#     return response
 
-# Launch the Gradio chat interface
-gr.ChatInterface(chatresponse).launch()
+# # Launch the Gradio chat interface
+# gr.ChatInterface(chatresponse).launch()
 
 # import gradio as gr
 # from langchain.schema import AIMessage, HumanMessage