utils/data_insights.py

import pyvo as vo
import pandas as pd

def fetch_exoplanet_data():
    # Connect to NASA Exoplanet Archive TAP Service
    tap_service = vo.dal.TAPService("https://exoplanetarchive.ipac.caltech.edu/TAP")

    # Query to fetch all columns from the pscomppars table
    ex_query = """
        SELECT TOP 10 pl_name, hostname, sy_snum, sy_pnum, discoverymethod, disc_year, disc_facility, pl_controv_flag, pl_orbper, pl_orbsmax, pl_rade, pl_bmasse, pl_orbeccen, pl_eqt, st_spectype, st_teff, st_rad, st_mass, ra, dec, sy_vmag
        FROM pscomppars
    """
    # Execute the query
    qresult = tap_service.search(ex_query)

    # Convert to a Pandas DataFrame
    ptable = qresult.to_table()
    exoplanet_data = ptable.to_pandas()

    return exoplanet_data

def generate_data_insights(user_input, client, exoplanet_data, max_tokens=500, temperature=0.3):
    """
    Generate insights by passing the user's input along with the exoplanet data to GPT-4.
    """
    # Convert the dataframe to a readable format for GPT (e.g., CSV-style text)
    data_as_text = exoplanet_data.to_csv(index=False)  # CSV-style for better readability

    # Create a prompt with the user query and the data sample
    insights_prompt = (
        f"Analyze the following user query and provide relevant insights based on the provided exoplanet data.\n\n"
        f"User Query: {user_input}\n\n"
        f"Exoplanet Data:\n{data_as_text}\n\n"
        f"Please provide insights that are relevant to the user's query."
    )
    
    # Call GPT-4 to generate insights based on the data and user input
    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[
            {"role": "system", "content": "You are an expert in analyzing astronomical data and generating insights."},
            {"role": "user", "content": insights_prompt}
        ],
        max_tokens=max_tokens,
        temperature=temperature
    )
    
    # Extract and return GPT-4's insights
    insights_from_data = response.choices[0].message.content.strip()
    return insights_from_data