soybean_dataset.py

# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# TODO: Address all TODOs and remove all explanatory comments
"""TODO: Add a description here."""


import csv
import json
import os
from typing import List
import datasets
import logging

import numpy as np
from PIL import Image
import os
import io
import pandas as pd
import matplotlib.pyplot as plt
from numpy import asarray
import requests
from io import BytesIO
from numpy import asarray


# TODO: Add BibTeX citation
# Find for instance the citation on arxiv or on the dataset repo/website
_CITATION = """\
@article{chen2023dataset,
  title={A dataset of the quality of soybean harvested by mechanization for deep-learning-based monitoring and analysis},
  author={Chen, M and Jin, C and Ni, Y and Yang, T and Xu, J},
  journal={Data in Brief},
  volume={52},
  pages={109833},
  year={2023},
  publisher={Elsevier},
  doi={10.1016/j.dib.2023.109833}
}

"""

# TODO: Add description of the dataset here
# You can copy an official description
_DESCRIPTION = """\
This dataset contains images captured during the mechanized harvesting of soybeans, aimed at facilitating the development of machine vision and deep learning models for quality analysis. It contains information of original soybean pictures in different forms, labels of whether the soybean belongs to training, validation, or testing datasets, segmentation class of soybean pictures in one dataset.
"""

# TODO: Add a link to an official homepage for the dataset here
_HOMEPAGE = "https://huggingface.co/datasets/lisawen/soybean_dataset"

# TODO: Add the licence for the dataset here if you can find it
_LICENSE = "Under a Creative Commons license"

# TODO: Add link to the official dataset URLs here
# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)

_URLs = {
    "train" : "https://huggingface.co/datasets/lisawen/soybean_dataset/resolve/main/train.zip?download=true",
    "test": "https://huggingface.co/datasets/lisawen/soybean_dataset/resolve/main/test.zip?download=true",
    "valid": "https://huggingface.co/datasets/lisawen/soybean_dataset/resolve/main/valid.zip?download=true"
}

# TODO: Name of the dataset usually matches the script name with CamelCase instead of snake_case
class SoybeanDataset(datasets.GeneratorBasedBuilder):
    """TODO: Short description of my dataset."""

    _URLs = _URLs
    VERSION = datasets.Version("1.1.0")

    def _info(self):
      # raise ValueError('woops!')
      return datasets.DatasetInfo(
          description=_DESCRIPTION,
          features=datasets.Features(
              {
                  
                  "original_image": datasets.Image(),
                  "segmentation_image": datasets.Image(),
                  
              }
          ),
          # No default supervised_keys (as we have to pass both question
          # and context as input).
          supervised_keys=("original_image","segmentation_image"),
          homepage="https://github.com/lisawen0707/soybean/tree/main",
          citation=_CITATION,
      )

    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
        # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
        # Since the dataset is on Google Drive, you need to implement a way to download it using the Google Drive API.

        # The path to the dataset file in Google Drive
        urls_to_download = self._URLs
        downloaded_files = dl_manager.download_and_extract(urls_to_download)

        # Since we're using a local file, we don't need to download it, so we just return the path.
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
            datasets.SplitGenerator(
                name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),    
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["valid"]}),
        ]

    # def process_image(self,image_url):
    #         response = requests.get(image_url)
    #         response.raise_for_status()  # This will raise an exception if there is a download error

    #         # Open the image from the downloaded bytes and return the PIL Image
    #         img = Image.open(BytesIO(response.content))
    #         return img
        


    def _generate_examples(self, filepath):
        logging.info("Generating examples from = %s", filepath)
        
        # Initialize a counter for unique ID
        unique_id = 0
        
        # Get the list of files in the directory
        file_list = os.listdir(filepath)

        # Create pairs of original and segmentation images
        images_pairs = [(f, f.replace('_original.jpg', '_segmentation.png')) for f in file_list if '_original.jpg' in f]
        
        # Iterate over the pairs and yield examples
        for original_image_name, segmentation_image_name in images_pairs:
            # Construct the full path to the original and segmentation images
            original_image_path = os.path.join(filepath, original_image_name)
            segmentation_image_path = os.path.join(filepath, segmentation_image_name)
            
            # Open the original image
            original_image = Image.open(original_image_path)
            
            # Open the segmentation image
            segmentation_image = Image.open(segmentation_image_path)
            
            # Yield the current example
            yield unique_id, {
                "original_image": original_image,
                "segmentation_image": segmentation_image,
            }
            
            # Increment the unique ID for the next example
            unique_id += 1