#!/usr/bin/env python3
"""
Script to randomly sample 1000 ELG galaxies from DESI DR1
"""

import pandas as pd
import numpy as np
import os
import json
from pathlib import Path

def sample_elg_galaxies():
    """Sample 1000 random ELG galaxies and prepare for spectrum download"""
    
    # Configuration
    elg_catalog_file = '/Users/sandyyuan/astro_mcp_data/desi/desi_search_dr1_types_galaxy_tracers_elg_20250811150055.csv'
    output_dir = '/Users/sandyyuan/astro_mcp_data/elg_1000_sample'
    n_samples = 1000
    
    print("=== ELG Galaxy Sample Preparation ===")
    print(f"Sampling {n_samples:,} ELG galaxies from DESI DR1")
    print(f"Source file: {elg_catalog_file}")
    print(f"Output directory: {output_dir}")
    
    # Create output directory
    Path(output_dir).mkdir(parents=True, exist_ok=True)
    
    # Load ELG catalog in chunks and sample randomly
    print(f"\nLoading ELG catalog...")
    chunk_size = 100000
    sampled_galaxies = []
    total_loaded = 0
    
    for chunk in pd.read_csv(elg_catalog_file, chunksize=chunk_size):
        total_loaded += len(chunk)
        print(f"  Loaded {total_loaded:,} galaxies so far...")
        
        # Sample from this chunk proportionally
        chunk_sample_size = min(n_samples, max(1, int(n_samples * len(chunk) / 100000)))
        if len(sampled_galaxies) < n_samples:
            remaining = n_samples - len(sampled_galaxies)
            sample_size = min(chunk_sample_size, remaining, len(chunk))
            
            if sample_size > 0:
                chunk_sample = chunk.sample(n=sample_size, random_state=42)
                sampled_galaxies.append(chunk_sample)
                print(f"    Sampled {sample_size} galaxies from this chunk")
        
        if len(sampled_galaxies) >= 10 or total_loaded >= 1000000:  # Stop after reasonable amount
            break
    
    # Combine all samples
    if sampled_galaxies:
        sample_df = pd.concat(sampled_galaxies, ignore_index=True)
        
        # If we have more than needed, randomly select final 1000
        if len(sample_df) > n_samples:
            sample_df = sample_df.sample(n=n_samples, random_state=123).reset_index(drop=True)
    else:
        print("Error: No galaxies sampled!")
        return
    
    print(f"\nFinal sample: {len(sample_df):,} ELG galaxies")
    
    # Analyze sample
    print(f"\n=== SAMPLE ANALYSIS ===")
    print(f"Redshift range: {sample_df['redshift'].min():.4f} - {sample_df['redshift'].max():.4f}")
    print(f"Mean redshift: {sample_df['redshift'].mean():.4f}")
    print(f"Median redshift: {sample_df['redshift'].median():.4f}")
    
    # Redshift distribution
    bins = [0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 2.0]
    labels = ['0.6-0.8', '0.8-1.0', '1.0-1.2', '1.2-1.4', '1.4-1.6', '1.6+']
    sample_df['z_bin'] = pd.cut(sample_df['redshift'], bins=bins, labels=labels, include_lowest=True)
    
    print(f"\nRedshift distribution:")
    z_counts = sample_df['z_bin'].value_counts().sort_index()
    for z_bin, count in z_counts.items():
        if pd.notna(z_bin):
            print(f"  {z_bin}: {count:,} galaxies ({count/len(sample_df)*100:.1f}%)")
    
    # Survey programs
    if 'program' in sample_df.columns:
        print(f"\nSurvey programs:")
        prog_counts = sample_df['program'].value_counts()
        for program, count in prog_counts.items():
            print(f"  {program}: {count:,} galaxies ({count/len(sample_df)*100:.1f}%)")
    
    # Save sample metadata
    sample_file = os.path.join(output_dir, 'elg_sample_metadata.csv')
    sample_df.to_csv(sample_file, index=False)
    print(f"\nSaved galaxy sample metadata to: {sample_file}")
    
    # Save target IDs
    target_ids = [str(tid) for tid in sample_df['targetid'].tolist()]
    target_ids_file = os.path.join(output_dir, 'target_ids.json')
    with open(target_ids_file, 'w') as f:
        json.dump(target_ids, f, indent=2)
    print(f"Saved {len(target_ids):,} target IDs to: {target_ids_file}")
    
    # Create summary
    summary = {
        'total_elg_galaxies_selected': len(sample_df),
        'redshift_range': [float(sample_df['redshift'].min()), float(sample_df['redshift'].max())],
        'mean_redshift': float(sample_df['redshift'].mean()),
        'target_ids_file': target_ids_file,
        'metadata_file': sample_file,
        'ready_for_download': True
    }
    
    summary_file = os.path.join(output_dir, 'elg_sample_summary.json')
    with open(summary_file, 'w') as f:
        json.dump(summary, f, indent=2, default=str)
    print(f"Saved summary to: {summary_file}")
    
    # Print first few target IDs for download
    print(f"\n=== READY FOR DOWNLOAD ===")
    print(f"Use MCP function 'mcp_astro-mcp_get_spectrum_by_targetid' for each target ID")
    print(f"First 10 target IDs:")
    for i, tid in enumerate(target_ids[:10]):
        print(f"  {i+1:2d}. {tid}")
    
    if len(target_ids) > 10:
        print(f"  ... and {len(target_ids)-10} more in {target_ids_file}")
    
    print(f"\nAll files saved to: {output_dir}")

if __name__ == "__main__":
    sample_elg_galaxies()