ui_refexp.py

# coding=utf-8
# 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.
"""Dataset script for UI Referring Expressions based on the UIBert RefExp dataset."""


import csv
import glob
import os
import tensorflow as tf

import datasets

import numpy as np

# Find for instance the citation on arxiv or on the dataset repo/website
_CITATION = """\
@misc{bai2021uibert,
      title={UIBert: Learning Generic Multimodal Representations for UI Understanding},
      author={Chongyang Bai and Xiaoxue Zang and Ying Xu and Srinivas Sunkara and Abhinav Rastogi and Jindong Chen and Blaise Aguera y Arcas},
      year={2021},
      eprint={2107.13731},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
"""

# TODO: Add description of the dataset here
# You can copy an official description
_DESCRIPTION = """\
This dataset is intended for UI understanding, referring expression and action automation model training. It's based on the UIBert RefExp dataset from Google Research, which is based on the RICO dataset.
"""

# TODO: Add a link to an official homepage for the dataset here
_HOMEPAGE = "https://github.com/google-research-datasets/uibert"

# TODO: Add the licence for the dataset here if you can find it
_LICENSE = "CC BY 4.0"

# Add link to the official dataset URLs here
# The HuggingFace dataset library don't host the datasets but only point to the original files
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
_DATA_URLs = {
    "ui_refexp": {
        "images": "https://huggingface.co/datasets/ncoop57/rico_captions/resolve/main/captions_hierarchies_images.zip",
    }
}

_METADATA_URLS = {
    "ui_refexp": {
        "train": "https://github.com/google-research-datasets/uibert/raw/main/ref_exp/train.tfrecord",
        "validation": "https://github.com/google-research-datasets/uibert/raw/main/ref_exp/dev.tfrecord",
        "test": "https://github.com/google-research-datasets/uibert/raw/main/ref_exp/test.tfrecord"
    }
}


def tfrecord2dict(raw_tfr_dataset: None):
  """Filter and convert refexp tfrecord file to dict object."""
  count = 0
  donut_refexp_dict = []
  for raw_record in raw_tfr_dataset:
    count += 1
    example = tf.train.Example()
    example.ParseFromString(raw_record.numpy())
    # print(f"total UI objects in this sample: {len(example.features.feature['image/object/bbox/xmin'].float_list.value)}")
    # print(f"feature keys: {example.features.feature.keys}")
    donut_refexp = {}
    image_id = example.features.feature['image/id'].bytes_list.value[0].decode()
    image_path=zipurl_template.format(image_id = image_id)
    donut_refexp["image_path"] = image_path
    donut_refexp["question"] = example.features.feature["image/ref_exp/text"].bytes_list.value[0].decode()
    object_idx = example.features.feature["image/ref_exp/label"].int64_list.value[0]
    object_idx = int(object_idx)
    # print(f"object_idx: {object_idx}")
    object_bb = {}
    # print(f"example.features.feature['image/object/bbox/xmin']: {example.features.feature['image/object/bbox/xmin'].float_list.value[object_idx]}")
    object_bb["xmin"] = example.features.feature['image/object/bbox/xmin'].float_list.value[object_idx]
    object_bb["ymin"] = example.features.feature['image/object/bbox/ymin'].float_list.value[object_idx]
    object_bb["xmax"] = example.features.feature['image/object/bbox/xmax'].float_list.value[object_idx]
    object_bb["ymax"] = example.features.feature['image/object/bbox/ymax'].float_list.value[object_idx]
    donut_refexp["answer"] = object_bb
    donut_refexp_dict.append(donut_refexp)
    if count != 3:
      continue
    print(f"Donut refexp: {donut_refexp}")
    # for key, feature in example.features.feature.items():
    #   if key in ['image/id', "image/ref_exp/text", "image/ref_exp/label", 'image/object/bbox/xmin', 'image/object/bbox/ymin', 'image/object/bbox/xmax', 'image/object/bbox/ymax']:
    #     print(key, feature)

  print(f"Total samples in the raw dataset: {count}")
  return donut_refexp_dict


class UIRefExp(datasets.GeneratorBasedBuilder):
    """Dataset with (image, question, answer) fields derive from UIBert RefExp."""

    VERSION = datasets.Version("1.1.0")

    # 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="ui_refexp",
             version=VERSION,
             description="Contains 66k+ unique UI screens. For each UI, we present a screenshot (JPG file) and the text shown on the screen that was extracted using an OCR model.",
         )
    #    ,
    #     # datasets.BuilderConfig(
    #     #     name="screenshots_captions_filtered",
    #     #     version=VERSION,
    #     #     description="Contains 25k unique UI screens. For each UI, we present a screenshot (JPG file) and the text shown on the screen that was extracted using an OCR model. Filtering was done as discussed in this paper: https://aclanthology.org/2020.acl-main.729.pdf",
    #     # ),
    # ]

    DEFAULT_CONFIG_NAME = "screenshots_captions_filtered"

    def _info(self):
        features = datasets.Features(
            {
                "screenshot": datasets.Image(),
                "prompt": datasets.Value("string"), # click the search button next to menu drawer at the top of the screen
                "target_bounding_box": dict, # [xmin, ymin, xmax, ymax], normalized screen reference values between 0 and 1
            }
        )

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            supervised_keys=("screenshot","prompt", "target_bounding_box"),
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""
        # 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
        image_urls = _DATA_URLs[self.config.name]
        image_archive = dl_manager.download(image_urls)
        # download and extract TFRecord labeling metadata
        local_tfrs = {}
        for split, tfrecord_url in _METADATA_URLS:
            local_tfr_file = dl_manager.download(tfrecord_url)
            local_tfrs[split] = local_tfr_file
        
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "root_dir": data_dir,
                    "metadata_file": local_tfrs["train"], 
                    "images": dl_manager.iter_archive(archive_path),
                    "split": "train",
                    
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "root_dir": data_dir,
                    "metadata_file": local_tfrs["validation"], 
                    "images": dl_manager.iter_archive(archive_path),
                    "split": "validation",
                },                
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "root_dir": data_dir,
                    "metadata_file": local_tfrs["test"], 
                    "images": dl_manager.iter_archive(archive_path),
                    "split": "test",
                },                
            )
        ]

    def _generate_examples(
        self,
        root_dir,
        metadata_file,
        images,
        split,  # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
    ):
        """Yields examples as (key, example) tuples."""
        # This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
        # The `key` is here for legacy reason (tfds) and is not important in itself.
        # filter tfrecord and convert to json
        
        with open(metadata_path, encoding="utf-8") as f:
            files_to_keep = set(f.read().split("\n"))
        for file_path, file_obj in images:
            if file_path.startswith(_IMAGES_DIR):
                if file_path[len(_IMAGES_DIR) : -len(".jpg")] in files_to_keep:
                    label = file_path.split("/")[2]
                    yield file_path, {
                        "image": {"path": file_path, "bytes": file_obj.read()},
                        "label": label,
                    }