data/dataset.py

# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
# SPDX-License-Identifier: MIT

import random
import torch
from torch.utils.data import Dataset
from torch.utils.data import sampler

# import lmdb
import torchvision.transforms as transforms
import six
import sys
from PIL import Image
import numpy as np
import os
import sys
import pickle
import numpy as np
from params import *
import glob, cv2
import torchvision.transforms as transforms

def get_transform(grayscale=False, convert=True):

    transform_list = []
    if grayscale:
        transform_list.append(transforms.Grayscale(1))

    if convert:
        transform_list += [transforms.ToTensor()]
        if grayscale:
            transform_list += [transforms.Normalize((0.5,), (0.5,))]
        else:
            transform_list += [transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]

    return transforms.Compose(transform_list)


class TextDataset:

    def __init__(self, base_path=DATASET_PATHS, num_examples=20, target_transform=None):

        self.NUM_EXAMPLES = num_examples

        # base_path = DATASET_PATHS
        file_to_store = open(base_path, "rb")
        self.IMG_DATA = pickle.load(file_to_store)["train"]
        self.IMG_DATA = dict(list(self.IMG_DATA.items()))  # [:NUM_WRITERS])
        if "None" in self.IMG_DATA.keys():
            del self.IMG_DATA["None"]
        self.author_id = list(self.IMG_DATA.keys())
        self.data = []
        for idx, (author_id, images) in enumerate(self.IMG_DATA.items()):
            for img_data in images:
                self.data.append(
                    {
                        "author_idx": idx,
                        "author_id": author_id,
                        "img": img_data["img"],
                        "label": img_data["label"],
                    }
                )
        self.transform = get_transform(grayscale=True)

        self.target_transform = target_transform

        self.collate_fn = TextCollator()

    def __len__(self):
        return len(self.data)

    def __getitem__(self, index):

        NUM_SAMPLES = self.NUM_EXAMPLES
        item_data = self.data[index]
        author_id = item_data["author_id"]
        img = item_data["img"]
        label = item_data["label"]
        author_idx = item_data["author_idx"]

        self.IMG_DATA_AUTHOR = self.IMG_DATA[author_id]
        random_idxs = np.random.choice(
            len(self.IMG_DATA_AUTHOR), NUM_SAMPLES, replace=True
        )

        rand_id_real = np.random.choice(len(self.IMG_DATA_AUTHOR))
        real_img = self.transform(Image.fromarray(np.array(img.convert("L"))))
        real_labels = label.encode()
        imgs = [
            np.array(self.IMG_DATA_AUTHOR[idx]["img"].convert("L"))
            for idx in random_idxs
        ]
        labels = [self.IMG_DATA_AUTHOR[idx]["label"].encode() for idx in random_idxs]

        max_width = 192  # [img.shape[1] for img in imgs]

        imgs_pad = []
        imgs_wids = []

        for img in imgs:

            img = 255 - img
            img_height, img_width = img.shape[0], img.shape[1]
            outImg = np.zeros((img_height, max_width), dtype="float32")
            outImg[:, :img_width] = img[:, :max_width]

            img = 255 - outImg

            imgs_pad.append(self.transform((Image.fromarray(img))))
            imgs_wids.append(img_width)

        imgs_pad = torch.cat(imgs_pad, 0)
        

        item = {
            "simg": imgs_pad,
            "swids": imgs_wids,
            "img": real_img,
            "label": real_labels,
            "img_path": "img_path",
            "idx": "indexes",
            "wcl": author_idx,
         
        }

        return item


class TextDatasetval:

    def __init__(self, base_path=DATASET_PATHS, num_examples=20, target_transform=None):

        self.NUM_EXAMPLES = num_examples
        # base_path = DATASET_PATHS
        file_to_store = open(base_path, "rb")
        self.IMG_DATA = pickle.load(file_to_store)["test"]
        self.IMG_DATA = dict(list(self.IMG_DATA.items()))  # [NUM_WRITERS:])
        if "None" in self.IMG_DATA.keys():
            del self.IMG_DATA["None"]
        self.author_id = list(self.IMG_DATA.keys())
        self.data = []
        for idx, (author_id, images) in enumerate(self.IMG_DATA.items()):
            for img_data in images:
                self.data.append(
                    {
                        "author_idx": idx,
                        "author_id": author_id,
                        "img": img_data["img"],
                        "label": img_data["label"],
                    }
                )
        self.transform = get_transform(grayscale=True)
        self.target_transform = target_transform

        self.collate_fn = TextCollator()

    def __len__(self):
        return len(self.data)

    def __getitem__(self, index):

        NUM_SAMPLES = self.NUM_EXAMPLES
        item_data = self.data[index]
        author_id = item_data["author_id"]
        img = item_data["img"]
        label = item_data["label"]
        author_idx = item_data["author_idx"]

        self.IMG_DATA_AUTHOR = self.IMG_DATA[author_id]
        random_idxs = np.random.choice(
            len(self.IMG_DATA_AUTHOR), NUM_SAMPLES, replace=True
        )

        rand_id_real = np.random.choice(len(self.IMG_DATA_AUTHOR))
        real_img = self.transform(Image.fromarray(np.array(img.convert("L"))))
        real_labels = label.encode()
        imgs = [
            np.array(self.IMG_DATA_AUTHOR[idx]["img"].convert("L"))
            for idx in random_idxs
        ]
        labels = [self.IMG_DATA_AUTHOR[idx]["label"].encode() for idx in random_idxs]

        max_width = 192  # [img.shape[1] for img in imgs]

        imgs_pad = []
        imgs_wids = []

        for img in imgs:

            img = 255 - img
            img_height, img_width = img.shape[0], img.shape[1]
            outImg = np.zeros((img_height, max_width), dtype="float32")
            outImg[:, :img_width] = img[:, :max_width]

            img = 255 - outImg

            imgs_pad.append(self.transform((Image.fromarray(img))))
            imgs_wids.append(img_width)

        imgs_pad = torch.cat(imgs_pad, 0)
        
        item = {
            "simg": imgs_pad,
            "swids": imgs_wids,
            "img": real_img,
            "label": real_labels,
            "img_path": "img_path",
            "idx": "indexes",
            "wcl": author_idx,
       
        }

        return item


class TextCollator(object):
    def __init__(self):
        self.resolution = resolution

    def __call__(self, batch):

        img_path = [item["img_path"] for item in batch]
        width = [item["img"].shape[2] for item in batch]
        indexes = [item["idx"] for item in batch]
        simgs = torch.stack([item["simg"] for item in batch], 0)
     
        wcls = torch.Tensor([item["wcl"] for item in batch])
        swids = torch.Tensor([item["swids"] for item in batch])
        imgs = torch.ones(
            [
                len(batch),
                batch[0]["img"].shape[0],
                batch[0]["img"].shape[1],
                max(width),
            ],
            dtype=torch.float32,
        )
        for idx, item in enumerate(batch):
            try:
                imgs[idx, :, :, 0 : item["img"].shape[2]] = item["img"]
            except:
                print(imgs.shape)
        item = {
            "img": imgs,
            "img_path": img_path,
            "idx": indexes,
            "simg": simgs,
            "swids": swids,
            "wcl": wcls,
       
        }
        if "label" in batch[0].keys():
            labels = [item["label"] for item in batch]
            item["label"] = labels
        if "z" in batch[0].keys():
            z = torch.stack([item["z"] for item in batch])
            item["z"] = z
        return item