app.py

import streamlit as st
from gpt_researcher import GPTResearcher
import asyncio
import nest_asyncio

# Access secrets
openai_api_key = st.secrets["OPENAI_API_KEY"]
tavily_api_key = st.secrets["TAVILY_API_KEY"]

# Apply the asyncio patch from nest_asyncio if required
nest_asyncio.apply()

# Constants
REPORT_TYPE = "research_report"  # Assuming this remains constant; modify as needed


# Function to handle asynchronous calls
def run_async(coroutine):
    loop = asyncio.get_event_loop()
    return loop.run_until_complete(coroutine)


# Streamlit interface
st.title("GPT Research Report Generator")

# User inputs
query = st.text_input(
    "Enter your research query:",
    "Extract all the information about how the ranking for internal links works.",
)
report_type = st.selectbox(
    "Select report type:",
    options=["research_report", "summary", "detailed_analysis"],
    index=0,
)
sources = st.text_area("Enter source URLs (one per line if multiple):")

# Processing the sources input into a list
source_urls = [url.strip() for url in sources.split("\n") if url.strip()]

# Button to generate report
if st.button("Generate Report"):
    if not query:
        st.warning("Please enter a query to generate a report.")
    else:
        with st.spinner("Generating report..."):
            # Fetch the report asynchronously
            fetch_report_coroutine = fetch_report(
                query, report_type, source_urls if source_urls else None
            )
            report = run_async(fetch_report_coroutine)
            st.success("Report generated successfully!")
            st.write(report)