Create app.py

app.py

@@ -0,0 +1,78 @@
+import os
+import langchain
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.document_loaders import TextLoader
+from langchain.vectorstores import FAISS
+from langchain.embeddings import HuggingFaceBgeEmbeddings
+from langchain.schema import Document
+import streamlit as st
+from langchain_groq import ChatGroq
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.prompts import PromptTemplate
+from langchain_core.output_parsers import StrOutputParser
+from langchain.schema.runnable import RunnablePassthrough
+text_splitter = RecursiveCharacterTextSplitter(chunk_size=75, chunk_overlap=0, separators=["\n",'.  '])
+loader = TextLoader('data.txt')
+docs = loader.load()
+split_docs = text_splitter.split_documents(docs)
+# Iterate over the Document objects and extract the content and metadata using the appropriate attributes
+split_docss = [Document(page_content=t.page_content, metadata=t.metadata) for t in split_docs]
+embeddings = HuggingFaceBgeEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+vectorstore = FAISS.from_documents(split_docss,embeddings)
+retriever=vectorstore.as_retriever()
+retriever.search_kwargs['k'] = 5
+os.environ['GROQ_API_KEY']='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'
+def preprocess(text):
+  return "\n".join(x.page_content for x in text)
+def format_docs(docs):
+  return "\n".join(doc.page_content for doc in docs)
+def printer(text):
+  print(text)
+  return text
+
+llm = ChatGroq(temperature=0, model_name="mixtral-8x7b-32768", streaming=True, verbose=False)
+# llama3-8b-8192
+prompt='''You are a coupon recommender chatbot.
+Understand the question with the category of the item and the location:
+{summary}
+and answer with only relevant coupons:
+{context}
+
+Instructions:
+List only the coupons that match the category of the item along with the location in the question.Use three sentences maximum and keep the answer concise.
+ '''
+template=PromptTemplate(template=prompt,input_variables=['summary','context'])
+rag_chain=({ "summary": RunnablePassthrough(), "context": retriever | format_docs ,
+             } | template|printer| llm| StrOutputParser())
+
+st.set_page_config(page_title="Lloyds Mobile Rewards Section", layout="wide")
+
+st.title("Mock Rewards Section")
+st.header("Welcome to Your Rewards")
+st.write("Find exclusive offers and discounts just for you!")
+
+st.sidebar.title("Check your eligibility to turn on the discover mode!")
+# category = st.sidebar.selectbox("Select Category", ["All", "Electronics", "Food", "Groceries"])
+if 'count' not in st.session_state:
+    st.session_state.count = 0
+
+enable_controls = st.session_state.count == 5
+
+def increment_count():
+    st.session_state.count += 1
+
+st.sidebar.title(f'Number of transactions : {st.session_state.count}')
+# st.write(f"Button has been pressed {st.session_state.count} times")
+button=st.sidebar.button("Mock Transaction")
+if button:
+    increment_count()
+response=None
+# Display the updated count
+if enable_controls:
+    st.sidebar.write("You are now eligible for discover mode ")
+    user_input = st.sidebar.text_input("Enter your request or choice of category", key='user_input')
+    response=retriever.get_relevant_documents(query=st.session_state.user_input)
+
+st.write(f'{response}')
+if not enable_controls:
+    st.sidebar.write(f"Keep going {20-st.session_state.count} times.")