tutorial67

tutorial67/main.py

@@ -0,0 +1,14 @@
+from langchain_community.document_loaders import PyPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_chroma import Chroma
+from langchain_openai import OpenAIEmbeddings
+
+
+loader=PyPDFLoader("C:\\Users\\welcome\\OneDrive\\Documents\\GitHub\LLMtutorial\\tutorial67\\manual.pdf")
+docs=loader.load()
+print(docs)
+text_splitter=RecursiveCharacterTextSplitter(chunk_size=1000,chunk_overlap=50)
+chunked_documents=text_splitter.split_documents(docs)
+vectordb=Chroma.from_documents(
+    docs, OpenAIEmbeddings(), persist_directory="./chroma_db"
+)

tutorial67/prompt.txt

@@ -0,0 +1,6 @@
+"""
+You are an intelligent help desk agent specializing in printers. 
+Answer user queries {input} related to printers based strictly on the provided {context}. 
+Do not make assumptions or provide information not included in the context.
+If no information found in the {context} just reply sorry answer not known
+"""

tutorial67/qsn.txt

@@ -0,0 +1,5 @@
+how to use Epson Smart Panel Mobile App
+Using Epson RemotePrint
+i want to connect to a New Wi-Fi Router help me
+how to print photos
+i want to play football how to do this 

tutorial67/rag.py

@@ -0,0 +1,31 @@
+from langchain_community.vectorstores import Chroma 
+from langchain_openai import ChatOpenAI,OpenAIEmbeddings
+from langchain.chains import LLMChain
+from langchain.prompts import PromptTemplate
+import streamlit as st
+import os
+st.title("help desk agent for printers")
+st.write("printer help desk")
+
+vectordb=Chroma(persist_directory=os.getcwd()+"/chroma_db",embedding_function=OpenAIEmbeddings())
+prompt_template="""
+You are an intelligent help desk agent specializing in printers. 
+Answer user queries {input} related to printers based strictly on the provided {context}. 
+Do not make assumptions or provide information not included in the {context}.
+"""
+llm=ChatOpenAI(model="gpt-4o",max_tokens=1024)
+qa_chain=LLMChain(llm=llm,prompt=PromptTemplate.from_template(prompt_template))
+question=st.text_area("ask your question")
+
+if question is not None:
+    button=st.button("submit")
+    if button:
+        rd=vectordb.similarity_search(question,k=5)
+        context=""
+        for d in rd:
+            context +=d.page_content
+        print(context)
+        response=qa_chain.invoke({"input":question,"context":context})
+        result=response["text"]
+        print(result)
+        st.write(result)