File size: 6,532 Bytes
7c96b72 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 |
import os
import pandas as pd
import datasets
from os.path import join
# convert these to features
#id,latitude,longitude,thumb_original_url,country,sequence,captured_at,lon_bin,lat_bin,cell,region,sub-region,city,land_cover,road_index,drive_side,climate,soil,dist_sea,quadtree_10_5000,quadtree_10_25000,quadtree_10_1000,quadtree_10_50000,quadtree_10_12500,quadtree_10_500,quadtree_10_2500,unique_region,unique_sub-region,unique_city,unique_country,creator_username,creator_id
#3859149887465501,-43.804769384023,-176.61409250805,,8,"(0, 8)",Chatham Islands,,Waitangi,4,4.661764145,1,15,3,0.0068841379890803,0,0,0,0,0,0,0,Chatham Islands_NZ,,Waitangi_NaN_Chatham Islands_NZ,NZ,roadroid,111336221091714.0
class OSV5M(datasets.GeneratorBasedBuilder):
def __init__(self, *args, **kwargs):
self.full = kwargs.pop('full', False)
super().__init__(*args, **kwargs)
print('OSV5M', self.__dict__)
def _info(self):
if self.full:
return datasets.DatasetInfo(
features=datasets.Features(
{
"image": datasets.Image(),
"latitude": datasets.Value("float32"),
"longitude": datasets.Value("float32"),
"thumb_original_url": datasets.Value("string"),
"country": datasets.Value("string"),
"sequence": datasets.Value("string"),
"captured_at": datasets.Value("string"),
"lon_bin": datasets.Value("float32"),
"lat_bin": datasets.Value("float32"),
"cell": datasets.Value("string"),
"region": datasets.Value("string"),
"sub-region": datasets.Value("string"),
"city": datasets.Value("string"),
"land_cover": datasets.Value("float32"),
"road_index": datasets.Value("float32"),
"drive_side": datasets.Value("float32"),
"climate": datasets.Value("float32"),
"soil": datasets.Value("float32"),
"dist_sea": datasets.Value("float32"),
"quadtree_10_5000": datasets.Value("int32"),
"quadtree_10_25000": datasets.Value("int32"),
"quadtree_10_1000": datasets.Value("int32"),
"quadtree_10_50000": datasets.Value("int32"),
"quadtree_10_12500": datasets.Value("int32"),
"quadtree_10_500": datasets.Value("int32"),
"quadtree_10_2500": datasets.Value("int32"),
"unique_region": datasets.Value("string"),
"unique_sub-region": datasets.Value("string"),
"unique_city": datasets.Value("string"),
"unique_country": datasets.Value("string"),
"creator_username": datasets.Value("string"),
"creator_id": datasets.Value("string"),
}
)
)
else:
return datasets.DatasetInfo(
features=datasets.Features(
{
"image": datasets.Image(),
"latitude": datasets.Value("float32"),
"longitude": datasets.Value("float32"),
"country": datasets.Value("string"),
"region": datasets.Value("string"),
"sub-region": datasets.Value("string"),
"city": datasets.Value("string"),
}
)
)
def df(self, annotation_path):
if not hasattr(self, 'df_'):
self.df_ = {}
if annotation_path not in self.df_:
df = pd.read_csv(annotation_path, dtype={
'id': str, 'creator_id': str, 'creator_username': str,
'unique_country': str, 'unique_city': str, 'unique_sub-region': str, 'unique_region': str,
'quadtree_10_2500': int, 'quadtree_10_500': int, 'quadtree_10_12500': int, 'quadtree_10_50000': int, 'quadtree_10_1000': int, 'quadtree_10_25000': int, 'quadtree_10_5000': int,
'dist_sea': float, 'soil': float, 'climate': float, 'drive_side': float, 'road_index': float, 'land_cover': float, 'city': str, 'sub-region': str, 'region': str, 'cell': str, 'lat_bin': float, 'lon_bin': float, 'captured_at': str, 'sequence': str, 'country': str, 'thumb_original_url': str, 'longitude': float, 'latitude': float
})
if not self.full:
df = df[['id', 'latitude', 'longitude', 'country', 'region', 'sub-region', 'city']]
df = df.set_index('id')
self.df_[annotation_path] = df.to_dict('index')
return self.df_[annotation_path]
def _split_generators(self, dl_manager):
_URLS = {
"train": [join('images', 'train', str(i).zfill(2) + '.zip') for i in range(98)],
"test": [join('images', 'test', str(i).zfill(2) + '.zip') for i in range(5)],
"train_meta": "train.csv",
"test_meta": "test.csv",
}
data_files = dl_manager.download_and_extract(_URLS)
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
gen_kwargs={
"image_paths": dl_manager.iter_files(data_files["train"]),
"annotation_path": data_files["train_meta"],
},
),
datasets.SplitGenerator(
name=datasets.Split.TEST,
gen_kwargs={
"image_paths": dl_manager.iter_files(data_files["test"]),
"annotation_path": data_files["test_meta"],
},
),
]
def _generate_examples(self, image_paths, annotation_path):
"""Generate examples."""
df = self.df(annotation_path)
for idx, image_path in enumerate(image_paths):
info_id = os.path.splitext(os.path.split(image_path)[-1])[0]
try:
example = {
"image": image_path,
} | df[info_id]
except Exception as e:
print('Exception ' + str(e), info_id, idx, image_path, sep='\n')
continue
yield idx, example |