Create app.py

app.py

@@ -0,0 +1,79 @@
+# Generics
+import os
+import keyfile
+import warnings
+import streamlit as st
+from pydantic import BaseModel
+warnings.filterwarnings("ignore")
+
+# Langchain packages
+from langchain.document_loaders import TextLoader
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.embeddings import HuggingFaceEmbeddings
+
+loader = TextLoader("/content/drive/MyDrive/Intelli_GenAI/RAG/Machine Learning Operations.txt")
+documents = loader.load()
+text_splitter = CharacterTextSplitter(chunk_size = 1000, chunk_overlap = 4)
+docs = text_splitter.split_documents(documents)
+
+
+
+
+class AIMessage(BaseModel):
+    content: str
+    
+# initializing the sessionMessages
+if "sessionMessages" not in st.session_state:
+    st.session_state["sessionMessages"] = []
+# General Instruction
+if "sessionMessages" not in st.session_state:
+    st.session_state.sessionMessage = [
+         SystemMessage(content = "You are a medieval magical healer known for your peculiar sarcasm")
+    ]
+
+# Configuring the key
+os.environ["GOOGLE_API_KEY"] = keyfile.GOOGLEKEY
+
+# Create a model
+llm = ChatGoogleGenerativeAI(
+    model="gemini-1.5-pro",
+    temperature=0.7,
+    convert_system_message_to_human= True
+)
+
+
+# Response function
+def load_answer(question):
+    st.session_state.sessionMessages.append(HumanMessage(content=question))
+    assistant_response = llm.invoke(st.session_state.sessionMessages)
+    
+    # Assuming assistant_response is an object with a 'content' attribute
+    if hasattr(assistant_response, 'content') and isinstance(assistant_response.content, str):
+        processed_content = assistant_response.content
+        st.session_state.sessionMessages.append(AIMessage(content=processed_content))
+    else:
+        st.error("Invalid response received from AI.")
+        processed_content = "Sorry, I couldn't process your request."
+
+    return processed_content
+
+# def load_answer(question):
+#     st.session_state.sessionMessages.append(HumanMessage(content = question))
+#     assistant_answer = llm.invoke(st.session_state.sessionMessages)
+#     st.session_state.sessionMessages.append(AIMessage(content = assistant_answer))
+#     return assistant_answer.content
+
+# User message
+def get_text():
+    input_text = st.text_input("You: ", key = input)
+    return input_text
+
+
+# Implementation
+user_input = get_text()
+submit = st.button("Generate")
+
+if submit:
+    resp = load_answer(user_input)
+    st.subheader("Answer: ")
+    st.write(resp, key = 1)