docbank.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 os
import itertools
import numpy as np
from PIL import Image
from transformers import AutoTokenizer
import datasets

# TODO: Add BibTeX citation
# Find for instance the citation on arxiv or on the dataset repo/website
_CITATION = """\
@InProceedings{huggingface:dataset,
title = {A great new dataset},
author={huggingface, Inc.
},
year={2020}
}
"""

# TODO: Add description of the dataset here
# You can copy an official description
_DESCRIPTION = """\
This new dataset is designed to solve this great NLP task and is crafted with a lot of care.
"""

# TODO: Add a link to an official homepage for the dataset here
_HOMEPAGE = ""

# TODO: Add the licence for the dataset here if you can find it
_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 = {
    "sample": "http://hyperion.bbirke.de/data/docbank/sample.zip",
    "full": "",
}

_FEATURES = datasets.Features(
    {
        "id": datasets.Value("string"),
        "input_ids": datasets.Sequence(datasets.Value("int64")),
        "bboxes": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
        "RGBs": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
        "fonts": datasets.Sequence(datasets.Value("string")),
        "image": datasets.Array3D(shape=(3, 224, 224), dtype="uint8"),
        "original_image": datasets.features.Image(),
        "labels": datasets.Sequence(datasets.features.ClassLabel(
            names=['abstract', 'author', 'caption', 'date', 'equation', 'figure', 'footer', 'list', 'paragraph',
                   'reference', 'section', 'table', 'title']
        ))
        # These are the features of your dataset like images, labels ...
    }
)

_DEFUNCT_FILE_IDS = [
    '126.tar_1706.03360.gz_dispersion_v2_7', '119.tar_1606.07466.gz_20160819Draft_8',
    '167.tar_1412.4821.gz_IDM_TD_Paper_16', '17.tar_1701.07437.gz_muon-beam-dump_final_2',
    '31.tar_1702.04307.gz_held-karp_21', '7.tar_1401.4493.gz_ReversibleNoise_2'
]


def load_image(image_path, size=None):
    image = Image.open(image_path).convert("RGB")
    w, h = image.size
    if size is not None:
        # resize image
        image = image.resize((size, size))
        image = np.asarray(image)
        image = image[:, :, ::-1] # flip color channels from RGB to BGR
        image = image.transpose(2, 0, 1) # move channels to first dimension
    return image, (w, h)


def normalize_bbox(bbox, size):
    return [
        int(1000 * int(bbox[0]) / size[0]),
        int(1000 * int(bbox[1]) / size[1]),
        int(1000 * int(bbox[2]) / size[0]),
        int(1000 * int(bbox[3]) / size[1]),
    ]


def simplify_bbox(bbox):
    return [
        min(bbox[0::2]),
        min(bbox[1::2]),
        max(bbox[2::2]),
        max(bbox[3::2]),
    ]


def merge_bbox(bbox_list):
    x0, y0, x1, y1 = list(zip(*bbox_list))
    return [min(x0), min(y0), max(x1), max(y1)]


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

    CHUNK_SIZE = 512
    VERSION = datasets.Version("1.0.0")

    # This is an example of a dataset with multiple configurations.
    # If you don't want/need to define several sub-sets in your dataset,
    # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.

    # If you need to make complex sub-parts in the datasets with configurable options
    # You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
    # BUILDER_CONFIG_CLASS = MyBuilderConfig

    # You will be able to load one or the other configurations in the following list with
    # data = datasets.load_dataset('my_dataset', 'first_domain')
    # data = datasets.load_dataset('my_dataset', 'second_domain')
    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name="sample", version=VERSION,
                               description="This part of my dataset covers a first domain"),
        datasets.BuilderConfig(name="full", version=VERSION,
                               description="This part of my dataset covers a second domain"),
    ]

    DEFAULT_CONFIG_NAME = "small"  # It's not mandatory to have a default configuration. Just use one if it make sense.
    TOKENIZER = AutoTokenizer.from_pretrained("xlm-roberta-base")

    def _info(self):
        # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset

        return datasets.DatasetInfo(
            # This is the description that will appear on the datasets page.
            description=_DESCRIPTION,
            # This defines the different columns of the dataset and their types
            features=_FEATURES,  # Here we define them above because they are different between the two configurations
            # If there's a common (input, target) tuple from the features, uncomment supervised_keys line below and
            # specify them. They'll be used if as_supervised=True in builder.as_dataset.
            # supervised_keys=("sentence", "label"),
            # Homepage of the dataset for documentation
            homepage=_HOMEPAGE,
            # License for the dataset if available
            license=_LICENSE,
            # Citation for the dataset
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        # TODO: This method is tasked with downloading/extracting the data and defining the splits depending on the configuration
        # If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name

        # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLS
        # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
        # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
        urls = _URLS[self.config.name]
        data_dir = dl_manager.download_and_extract(urls)
        with open(os.path.join(data_dir, "train.csv")) as f:
            files_train = [{'id': row['id'], 'filepath_txt': os.path.join(data_dir, row['filepath_txt']),
                            'filepath_img': os.path.join(data_dir, row['filepath_img'])} for row in
                           csv.DictReader(f, skipinitialspace=True)]
        with open(os.path.join(data_dir, "test.csv")) as f:
            files_test = [{'id': row['id'], 'filepath_txt': os.path.join(data_dir, row['filepath_txt']),
                           'filepath_img': os.path.join(data_dir, row['filepath_img'])} for row in
                          csv.DictReader(f, skipinitialspace=True)]
        with open(os.path.join(data_dir, "validation.csv")) as f:
            files_validation = [{'id': row['id'], 'filepath_txt': os.path.join(data_dir, row['filepath_txt']),
                                 'filepath_img': os.path.join(data_dir, row['filepath_img'])} for row in
                                csv.DictReader(f, skipinitialspace=True)]
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": files_train,
                    "split": "train",
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": files_validation,
                    "split": "validation",
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": files_test,
                    "split": "test"
                },
            ),
        ]

    # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
    def _generate_examples(self, filepath, split):
        # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
        # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
        #print(filepath)
        key = 0
        for f in filepath:
            #print(f)
            f_id = f['id']
            f_fp_txt = f['filepath_txt']
            f_fp_img = f['filepath_img']
            tokens = []
            bboxes = []
            rgbs = []
            fonts = []
            labels = []

            image, size = load_image(f_fp_img, size=224)
            original_image, _ = load_image(f_fp_img)

            try:
                with open(f_fp_txt, newline='', encoding='utf-8') as csvfile:
                    reader = csv.reader(csvfile, delimiter='\t', quotechar=' ')
                    for row in reader:
                        #if f_id == '121.tar_1606.08710.gz_mutualEnergy_05_77':
                        tokenized_input = self.TOKENIZER(
                            row[0],
                            add_special_tokens=False,
                            return_offsets_mapping=False,
                            return_attention_mask=False,
                        )
                        tokens.append(tokenized_input['input_ids'][0] if len(tokenized_input['input_ids']) == 1 else self.TOKENIZER.unk_token_id)
                        bboxes.append(normalize_bbox(row[1:5], size))
                        rgbs.append(row[5:8])
                        fonts.append(row[8])
                        labels.append(row[9])
            except:
                continue

            for chunk_id, index in enumerate(range(0, len(tokens), self.CHUNK_SIZE)):

                split_tokens = tokens[index:index + self.CHUNK_SIZE]
                split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
                split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
                split_fonts = fonts[index:index + self.CHUNK_SIZE]
                split_labels = labels[index:index + self.CHUNK_SIZE]

                if len(split_tokens) > self.CHUNK_SIZE:
                    print('Err')
                    print(key)
                    print(f_id)
                    print(split_tokens)


                yield key, {
                    "id": f"{f_id}_{chunk_id}",
                    'input_ids': split_tokens,
                    "bboxes": split_bboxes,
                    "RGBs": split_rgbs,
                    "fonts": split_fonts,
                    "image": image,
                    "original_image": original_image,
                    "labels": split_labels
                }
                key += 1