#!/usr/bin/env python3
"""Rebuild H5 metadata database - V3 (include expected positions from CSV)."""
import os
import re
import csv
import sqlite3
from datetime import datetime

H5_ROOTS = ['/mnt/data_sdb1', '/mnt/kingston']
CSV_PATH = '/mnt/kingston/Copie de SETE_AUV_DARFV4-Copier(1).csv'
DB_PATH = '/home/floppyrj45/docker/seismic-nodes-viewer/h5_data.db'

FILE_PATTERN = re.compile(r'b(\d+)')
CHANNEL_PATTERN = re.compile(r'ch(\d+)')

SCHEMA = [
    'CREATE TABLE IF NOT EXISTS positions (node_code INTEGER PRIMARY KEY, has_data BOOLEAN, has_aux BOOLEAN, sample_count INTEGER, last_seen TEXT, expected BOOLEAN)',
    'CREATE TABLE IF NOT EXISTS files (id INTEGER PRIMARY KEY AUTOINCREMENT, path TEXT, node_code INTEGER, channel INTEGER, dataset TEXT, size INTEGER, mtime INTEGER, FOREIGN KEY(node_code) REFERENCES positions(node_code))',
    'CREATE INDEX IF NOT EXISTS idx_files_node ON files(node_code)'
]

def rebuild_db():
    conn = sqlite3.connect(DB_PATH)
    cur = conn.cursor()

    for stmt in SCHEMA:
        cur.execute(stmt)

    cur.execute('DELETE FROM files')
    cur.execute('DELETE FROM positions')

    # 1. Charger les positions attendues depuis le CSV
    expected_nodes = set()
    try:
        with open(CSV_PATH, 'r', encoding='utf-8-sig') as f:
            reader = csv.DictReader(f)
            for row in reader:
                node_code = row.get('NodeCode', '').strip()
                if node_code and node_code.isdigit():
                    expected_nodes.add(int(node_code))
        print(f"✓ Loaded {len(expected_nodes)} expected positions from CSV")
    except Exception as e:
        print(f"⚠ CSV not found or error: {e}")
        print("  Continuing with file scan only...")

    # 2. Scanner les fichiers H5
    files_counter = 0
    found_nodes = {}

    for root in H5_ROOTS:
        for dirpath, _, filenames in os.walk(root):
            for filename in filenames:
                if not filename.endswith('.h5'):
                    continue
                
                filepath = os.path.join(dirpath, filename)
                
                node_match = FILE_PATTERN.search(filename)
                if not node_match:
                    continue
                
                node_code = int(node_match.group(1))
                channel_match = CHANNEL_PATTERN.search(filename)
                channel = int(channel_match.group(1)) if channel_match else -1
                dataset = 'aux' if 'aux' in filename else 'data'
                
                stat = os.stat(filepath)
                mtime = int(stat.st_mtime)
                size = stat.st_size
                
                found_nodes.setdefault(node_code, {'data': False, 'aux': False, 'count': 0, 'last': 0})
                found_nodes[node_code]['count'] += 1
                found_nodes[node_code]['last'] = max(found_nodes[node_code]['last'], mtime)
                if dataset == 'data':
                    found_nodes[node_code]['data'] = True
                else:
                    found_nodes[node_code]['aux'] = True
                
                cur.execute(
                    'INSERT INTO files (path, node_code, channel, dataset, size, mtime) VALUES (?, ?, ?, ?, ?, ?)',
                    (filepath, node_code, channel, dataset, size, mtime)
                )
                files_counter += 1

    print(f"✓ Indexed {files_counter} H5 files")
    print(f"✓ Found {len(found_nodes)} positions with data")

    # 3. Créer les entrées pour TOUTES les positions (attendues + trouvées)
    all_nodes = expected_nodes | set(found_nodes.keys())
    
    for node_code in all_nodes:
        is_expected = node_code in expected_nodes
        
        if node_code in found_nodes:
            stats = found_nodes[node_code]
            has_data = 1 if stats['data'] else 0
            has_aux = 1 if stats['aux'] else 0
            last_seen = datetime.fromtimestamp(stats['last']).isoformat()
            sample_count = stats['count']
        else:
            # Position attendue mais sans données
            has_data = 0
            has_aux = 0
            last_seen = None
            sample_count = 0
        
        cur.execute(
            'INSERT INTO positions (node_code, has_data, has_aux, sample_count, last_seen, expected) VALUES (?, ?, ?, ?, ?, ?)',
            (node_code, has_data, has_aux, sample_count, last_seen, 1 if is_expected else 0)
        )

    conn.commit()
    
    # Stats finales
    cur.execute('SELECT COUNT(*) FROM positions')
    total = cur.fetchone()[0]
    
    cur.execute('SELECT COUNT(*) FROM positions WHERE has_data = 1')
    with_data = cur.fetchone()[0]
    
    cur.execute('SELECT COUNT(*) FROM positions WHERE expected = 1')
    expected_count = cur.fetchone()[0]
    
    cur.execute('SELECT COUNT(*) FROM positions WHERE expected = 1 AND has_data = 0')
    missing = cur.fetchone()[0]
    
    print(f"\n📊 Database Summary:")
    print(f"  • Total positions in DB: {total}")
    print(f"  • Expected (from CSV): {expected_count}")
    print(f"  • With H5 data: {with_data}")
    print(f"  • Missing (expected but no data): {missing}")
    print(f"  • Coverage: {(with_data/expected_count*100 if expected_count else 0):.1f}%")
    
    conn.close()

if __name__ == '__main__':
    rebuild_db()