unsplash-25k-photos / unsplash-25k-photos.py
predprey's picture
fix: Fix UnicodeDecodeError and pyarrow.lib.ArrowInvalid when loading
00024c9
raw
history blame
4.04 kB
import os
import datasets
_CITATION = """\
@InProceedings{huggingface:dataset,
title = {Unsplash Lite Dataset 1.2.0 Photos},
author={Unsplash},
year={2022}
}
"""
_DESCRIPTION = """\
This is a dataset that streams photos data from the Unsplash 25K servers.
"""
_HOMEPAGE = "https://github.com/unsplash/datasets/"
_LICENSE = ""
_URL = "https://unsplash.com/data/lite/latest"
class Unsplash(datasets.GeneratorBasedBuilder):
"""The Unsplash 25K dataset for photos"""
def _info(self):
return datasets.DatasetInfo(
description=_DESCRIPTION,
features=datasets.Features(
{
'photo_id': datasets.Value("string"),
'photo_url': datasets.Value("string"),
'photo_image_url': datasets.Value("string"),
'photo_submitted_at': datasets.Value("string"),
'photo_featured': datasets.Value("string"),
'photo_width': datasets.Value("int32"),
'photo_height': datasets.Value("int32"),
'photo_aspect_ratio': datasets.Value("float32"),
'photo_description': datasets.Value("string"),
'photographer_username': datasets.Value("string"),
'photographer_first_name': datasets.Value("string"),
'photographer_last_name': datasets.Value("string"),
'exif_camera_make': datasets.Value("string"),
'exif_camera_model': datasets.Value("string"),
'exif_iso': datasets.Value("string"),
'exif_aperture_value': datasets.Value("string"),
'exif_focal_length': datasets.Value("string"),
'exif_exposure_time': datasets.Value("string"),
'photo_location_name': datasets.Value("string"),
'photo_location_latitude': datasets.Value("string"),
'photo_location_longitude': datasets.Value("string"),
'photo_location_country': datasets.Value("string"),
'photo_location_city': datasets.Value("string"),
'stats_views': datasets.Value("uint32"),
'stats_downloads': datasets.Value("uint32"),
'ai_description': datasets.Value("string"),
'ai_primary_landmark_name': datasets.Value("string"),
'ai_primary_landmark_latitude': datasets.Value("string"),
'ai_primary_landmark_longitude': datasets.Value("string"),
'ai_primary_landmark_confidence': datasets.Value("string"),
'blur_hash': datasets.Value("string"),
}
),
supervised_keys=None,
homepage="https://github.com/unsplash/datasets/",
citation=_CITATION,
)
def _split_generators(self, dl_manager):
new_url = dl_manager.download_and_extract(_URL)
# remove extra files
for file in os.listdir(new_url):
if os.path.isfile(new_url+"/"+file):
if file != 'photos.tsv000':
os.remove(new_url+'/'+file)
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
gen_kwargs={"filepath": os.path.join(new_url, "photos.tsv000")}
),
]
def _generate_examples(self, filepath):
"""This function returns the examples in the raw (text) form."""
with open(filepath, "r", encoding="utf-8") as f:
id_ = 0
for line in f:
if id_ == 0:
cols = line.strip().split("\t")
id_ += 1
else:
values = line.strip().split("\t")
if len(values) != len(cols):
values.extend(['']*(len(cols)-len(values)))
yield id_, {cols[i]: values[i] for i in range(len(cols))}
id_ += 1