lmdb_dataset.py

from io import BytesIO

import lmdb
from PIL import Image
from torch.utils.data import Dataset


def default_flist_reader(flist):
    """
    flist format: impath label\nimpath label\n ...(same to caffe's filelist)
    """
    imlist = []
    with open(flist, 'r') as rf:
        for line in rf.readlines():
            impath = line.strip()
            imlist.append(impath)
        
    return imlist

class LMDBImageDataset(Dataset):
    def __init__(self, filelist_path, path, transform, resolution=256):
        self.env = lmdb.open(
            path,
            max_readers=32,
            readonly=True,
            lock=False,
            readahead=False,
            meminit=False,
        )

        if not self.env:
            raise IOError('Cannot open lmdb dataset', path)

        with self.env.begin(write=False) as txn:
            self.length = int(txn.get('length'.encode('utf-8')).decode('utf-8'))

        self.flist =  default_flist_reader(filelist_path)

        self.resolution = resolution
        self.transform = transform

    def __len__(self):
        return self.length

    def __getitem__(self, index):
        with self.env.begin(write=False) as txn:
            key = self.flist[index].encode('utf-8')
            img_bytes = txn.get(key)

        buffer = BytesIO(img_bytes)
        img = Image.open(buffer)
        img = self.transform(img)

        return img, 0