app.py

from pandasai.llm import GoogleGemini
import streamlit as st
import os
import pandas as pd
from pandasai import SmartDataframe
from pandasai.responses.response_parser import  ResponseParser
from st_on_hover_tabs import on_hover_tabs
from ydata_profiling import ProfileReport
import google.generativeai as genai
import json

class StreamLitResponse(ResponseParser):
        def __init__(self,context) -> None:
              super().__init__(context)
        def format_dataframe(self,result):
               st.dataframe(result['value'])
               return
        def format_plot(self,result):
               st.image(result['value'])
               return
        def format_other(self, result):
               st.write(result['value'])
               return

gemini_api_key = os.environ['Gemini']

genai.configure(api_key=gemini_api_key)

generation_config = {
  "temperature": 0.2,
  "top_p": 0.95,
  "max_output_tokens": 5000,
}

model = genai.GenerativeModel(
  model_name="gemini-1.5-flash",
  generation_config=generation_config,
)


def get_pandas_profile(df):
    profile = ProfileReport(df, title="Profiling Report")
    json_profile = profile.to_json()
    dict_p = json.loads(json_profile)
    keys_to_keep = ['analysis', 'table', 'correlations', 'alerts', 'sample']

# Assuming your dictionary is named 'my_dict'
    filtered_dict = {key: dict_p[key] for key in keys_to_keep}
    return filtered_dict
    
def generateResponse(dataFrame,prompt):
        llm = GoogleGemini(api_key=gemini_api_key)
        pandas_agent = SmartDataframe(dataFrame,config={"llm":llm, "response_parser":StreamLitResponse})
        answer = pandas_agent.chat(prompt)
        return answer

st.write("# QuantBeta Insights")
st.markdown('<style>' + open('./style.css').read() + '</style>', unsafe_allow_html=True)
st.write("##### Engage in insightful conversations with your data")
with st.sidebar:
    st.title("QuantBeta Insights")
    st.sidebar.image("logoqb.jpeg", use_column_width=True)
    tabs = on_hover_tabs(tabName=['Chat', 'Reports'], 
                         iconName=['chat', 'dashboard'], default_choice=0)
        


uploaded_file = "64Qbeta.csv" 
#uploaded_file = "healthcare_dataset.csv"
if tabs =='Chat':
    df = pd.read_csv(uploaded_file)
    st.subheader("QuantBeta AI assistant")
    st.write("Get visualizations and analysis from our Gemini powered agent")
        
        # Read the CSV file
    #df = pd.read_csv(uploaded_file)

        # Display the data
    with st.expander("Preview"):
        st.write(df.head())

        # Plot the data
    user_input = st.text_input("Type your message here",placeholder="Ask me about your data")
    if user_input:
        answer = generateResponse(dataFrame=df,prompt=user_input)
        st.write(answer)

elif tabs == 'Reports':
    df = pd.read_csv(uploaded_file)
    

# Streamlit App
    st.subheader("Reports")
    st.write("Filter by Incubator to generate report")

# Display original 

# Filtering Interface
    st.write("Filtering Options")
    branch_names = df['Incubator Name'].unique().tolist()
    #product_names = df['Description'].unique().tolist()
    selected_branches = st.multiselect('Select incubator(s) Name(s)', branch_names, default=branch_names)
    #selected_products = st.multiselect('Select product(s) Name', product_names, default=product_names)

# Button to apply filters
    if st.button('Apply Filters and Generate report'):
        df = pd.read_csv(uploaded_file)
        filtered_df = df.copy()
    
    # Apply Branch Name Filter
        if selected_branches:
            filtered_df = filtered_df[filtered_df['Incubator Name'].isin(selected_branches)]
    
    # Apply Description Filter
    #if selected_products:
       # filtered_df = filtered_df[filtered_df['Product_Name'].isin(selected_products)]
    
    # Display filtered DataFrame
            st.write("Filtered DataFrame")
            with st.expander("Preview"):
                st.write(filtered_df.head())
            with st.spinner("Generating Report, Please Wait...."):
                prompt = """
You are an expert business analyst. Analyze the following data and generate a comprehensive and insightful business report, including appropriate key perfomance indicators and recommendations.

data:
"""  +  str(filtered_df.to_json(orient='records'))

                response = model.generate_content(prompt)
                response2 = generateResponse(filtered_df, "pie chart of number of hours  by incubator")
                report = response.text
                st.markdown(report)
                # Display the generated images
                st.success("Report Generated!")
        else:
            st.write("Filtered DataFrame")
            st.write("Click 'Apply Filters' to see the filtered data.")