import os
import re

def load_positions_mapping(positions_file_path):
    """
    Reads a report where entries look like:

      File: chesspgn_877.pgn
         Positions: 702430, Games: 9808

    and returns a dict { 'chesspgn_877.pgn': 702430, ... }.
    """
    mapping = {}
    file_pattern = re.compile(r'^\s*File:\s*(\S+)\s*$')
    pos_pattern  = re.compile(r'^\s*Positions:\s*(\d+)')
    with open(positions_file_path, 'r') as f:
        lines = f.readlines()

    i = 0
    while i < len(lines):
        m_file = file_pattern.match(lines[i])
        if m_file:
            fname = m_file.group(1)
            # look ahead up to 3 lines for Positions:
            found = False
            for j in range(i+1, min(i+4, len(lines))):
                m_pos = pos_pattern.match(lines[j])
                if m_pos:
                    mapping[fname] = int(m_pos.group(1))
                    found = True
                    break
            if not found:
                print(f"Warning: no Positions line found for {fname!r}")
        i += 1
    return mapping


def sum_positions_in_dir(mapping, split_dir):
    """
    Looks for all .pgn files in split_dir, sums up mapping[filename].
    Returns (count_of_files, total_positions).
    """
    pgns = [fn for fn in os.listdir(split_dir) if fn.endswith('.pgn')]
    total_positions = 0
    missing = []
    for fn in pgns:
        if fn in mapping:
            total_positions += mapping[fn]
        else:
            missing.append(fn)
    if missing:
        print(f"Warning: {len(missing)} files in {split_dir!r} not found in positions file:")
        for fn in missing:
            print("   ", fn)
    return len(pgns), total_positions

if __name__ == "__main__":
    # ←── adjust these paths to your environment ──→
    positions_txt = '/home/kage/chess_workspace/ChessBot-Battleground/dataset/ChessBot-Dataset/analysis_results.txt'
    input_dir     = '/home/kage/chess_workspace/ChessBot-Battleground/dataset/ChessBot-Dataset'    # contains subfolders 'train/' and 'test/'
    train_dir     = os.path.join(input_dir, 'train')
    test_dir      = os.path.join(input_dir, 'test')

    # load the per‐file positions
    mapping = load_positions_mapping(positions_txt)

    # compute for train
    n_train_files, train_positions = sum_positions_in_dir(mapping, train_dir)
    # compute for test
    n_test_files, test_positions   = sum_positions_in_dir(mapping, test_dir)

    print()
    print(f"TRAIN → {n_train_files} files, {train_positions:,} total positions")
    print(f" TEST → {n_test_files} files, {test_positions:,} total positions")