Spaces:
Sleeping
Sleeping
| from langchain.chains import LLMChain | |
| from langchain_community.llms import OpenAI | |
| from langchain_core.prompts import PromptTemplate | |
| import streamlit as st | |
| # Set the page to wide mode | |
| st.set_page_config(layout="wide") | |
| mini_template = "You are an expert researcher. You\'ve talked to hundreds of {Target Audience}. \ | |
| Each person in the niche of {Target Audience} has certain struggles that make it easier to sell {My Course}. \ | |
| These are called Pain Points. There's a recipe for getting to the core of the Pain Points of {Target Audience}. \ | |
| Namely, answer each of these Questions 3 times, each getting deeper in the issues of {Target Audience}, \ | |
| appealing to their Emotions and uncertainties related to {My Course}. \ | |
| The Questions (answer each QUESTION 3 times in listicle format according to the instructions):\ | |
| 1. What keeps them awake at night?\ | |
| 2. What are they afraid of?\ | |
| 3. What are they angry about?\ | |
| " | |
| st.title("Saas Application") | |
| prompt = PromptTemplate( | |
| input_variables = ["Target Audience", "My Course"], | |
| template=mini_template, | |
| ) | |
| chain = LLMChain(llm=OpenAI(), prompt=prompt) | |
| #target_audience = "professionals looking for course on Power BI" | |
| #my_course = "Zero to Hero in PowerBI" | |
| # Use the sidebar for input | |
| target_audience = st.sidebar.text_input("Enter your target audience") | |
| my_course = st.sidebar.text_input("Enter your course name") | |
| if st.sidebar.button("Get response"): | |
| if target_audience and my_course: | |
| with st.spinner("Generating response..."): | |
| with st.expander("Show prompt", expanded=False): | |
| st.info(prompt.template) | |
| answer = chain.run({"Target Audience": target_audience, "My Course": my_course}) | |
| # Split the 'answer' into sections based on the questions | |
| sections = [section.strip() for section in answer.split("\n\n") if section.strip() != ""] | |
| # Assuming there are exactly three sections based on your output structure | |
| if len(sections) == 3: | |
| # Extract titles for tabs | |
| titles = [section.split('\n')[0] for section in sections] | |
| # Extract content for each section, removing the title | |
| contents = [section.split('\n')[1:] for section in sections] | |
| # Create tabs for each category | |
| tabs = st.tabs(titles) | |
| for i, tab in enumerate(tabs): | |
| with tab: | |
| st.header(titles[i]) | |
| for content in contents[i]: | |
| st.markdown(content) | |
| else: | |
| st.error("The answer format does not match the expected structure.") | |
| st.success("Hope you like the response.❤") | |
| elif target_audience: | |
| st.error("Enter your course name.") | |
| elif my_course: | |
| st.error("Enter your target audience.") | |
| else: | |
| st.error("No input detected, Please provide the desired information.") | |