import os, re, json, h5py
from pathlib import Path
from datetime import datetime, timedelta
from tqdm import tqdm

DATA_ROOTS = [Path("/mnt/kingston"), Path("/mnt/data_sdb1")]
POSITIONS_CSV = Path("/mnt/kingston/Copie de SETE_AUV_DARFV4-Copier(1).csv")
OUTPUT_INDEX = Path("/mnt/kingston/seismic_webapp/data/index.json")
SAMPLE_RATE = 200

def load_pos():
    positions = {}
    if not POSITIONS_CSV.exists(): return {}
    with open(POSITIONS_CSV, 'r', encoding='utf-8', errors='replace') as f:
        lines = f.readlines()
        if len(lines) < 5: return {}
        headers = lines[3].strip().split(',')
        try:
            ni = headers.index('NodeCode')
            ei = headers.index('Aslaid Easting') if 'Aslaid Easting' in headers else headers.index('Preplot Easting')
            oi = headers.index('Aslaid Northing') if 'Aslaid Northing' in headers else headers.index('Preplot Northing')
        except: return {}
        for line in lines[4:]:
            parts = line.strip().split(',')
            try:
                nid = parts[ni].strip()
                positions[nid] = {
                    'easting': float(parts[ei]), 
                    'northing': float(parts[oi]), 
                    'depth': float(parts[headers.index('Aslaid Depth')]) if 'Aslaid Depth' in headers else 0.0
                }
            except: continue
    return positions

def scan():
    pos = load_pos()
    nodes = {}
    all_dates = set()
    file_count = 0
    
    print("🔍 Scanning ONLY 'data' H5 files (ignoring 'aux')...")
    all_h5_files = []
    for root in DATA_ROOTS:
        all_h5_files.extend(list(root.rglob("*.h5")))
    
    for h5_path in tqdm(all_h5_files):
        # FILTRE : Uniquement les fichiers contenant "data"
        if "_data_" not in h5_path.name.lower():
            continue
            
        try:
            match = re.search(r'auto_(\d+)_(\d{6})_b(\d+)_.*?_(\d{10})\.h5$', h5_path.name)
            if not match: continue
            
            julian_day = int(match.group(1))
            time_str = match.group(2)
            node_id = match.group(3)
            
            date_ref = datetime(2020, 1, 1) + timedelta(days=julian_day - 1)
            date_str = date_ref.strftime('%Y-%m-%d')
            
            h, m, s = int(time_str[:2]), int(time_str[2:4]), int(time_str[4:6])
            actual_start_ts = int(datetime(2020, 1, 1).timestamp() + (julian_day - 1) * 86400 + h * 3600 + m * 60 + s)
            
            with h5py.File(h5_path, 'r') as f:
                if 'adc_values' not in f: continue
                duration = f['adc_values'].shape[0] / SAMPLE_RATE
                actual_end_ts = actual_start_ts + duration
                
                all_dates.add(date_str)
                
                if node_id not in nodes:
                    nodes[node_id] = {
                        'id': node_id,
                        'position': pos.get(node_id),
                        'files': []
                    }
                
                # On extrait le canal du nom de fichier pour un matching plus précis
                channel_match = re.search(r'_ch(\d+)_', h5_path.name)
                channel = f"ch{channel_match.group(1)}" if channel_match else "ch0"
                
                nodes[node_id]['files'].append({
                    'path': str(h5_path),
                    'start': actual_start_ts,
                    'end': actual_end_ts,
                    'julian': julian_day,
                    'channel': channel # Canal spécifique au fichier
                })
                file_count += 1
        except: continue

    result = {
        'generated_at': datetime.now().isoformat(),
        'sample_rate_hz': SAMPLE_RATE,
        'nodes': nodes,
        'dates': sorted(list(all_dates))
    }
    
    with open(OUTPUT_INDEX, 'w') as f:
        json.dump(result, f, indent=2)
    
    print(f"✅ Index updated: {file_count} 'data' files, {len(nodes)} nodes.")

if __name__ == '__main__': scan()