Datasets:
mariosasko
commited on
Commit
•
96b39ed
1
Parent(s):
077dc5b
Make dataset streamable
Browse files- illustrated_ads.py +38 -30
illustrated_ads.py
CHANGED
@@ -14,10 +14,10 @@
|
|
14 |
"""Dataset of illustrated and non illustrated 19th Century newspaper ads."""
|
15 |
|
16 |
import ast
|
|
|
17 |
import pandas as pd
|
18 |
import datasets
|
19 |
from PIL import Image
|
20 |
-
from pathlib import Path
|
21 |
|
22 |
# TODO: Add BibTeX citation
|
23 |
# Find for instance the citation on arxiv or on the dataset repo/website
|
@@ -46,6 +46,19 @@ _LICENSE = "Public Domain"
|
|
46 |
|
47 |
_URLS = "https://zenodo.org/record/5838410/files/images.zip?download=1"
|
48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
|
50 |
# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
|
51 |
class IllustratedAds(datasets.GeneratorBasedBuilder):
|
@@ -94,48 +107,43 @@ class IllustratedAds(datasets.GeneratorBasedBuilder):
|
|
94 |
)
|
95 |
|
96 |
def _split_generators(self, dl_manager):
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
),
|
105 |
-
]
|
106 |
-
|
107 |
-
def _generate_examples(self, data_dir):
|
108 |
-
dtypes = {
|
109 |
-
"page_seq_num": "int64",
|
110 |
-
"edition_seq_num": "int64",
|
111 |
-
"batch": "string",
|
112 |
-
"lccn": "string",
|
113 |
-
"score": "float64",
|
114 |
-
"place_of_publication": "string",
|
115 |
-
"name": "string",
|
116 |
-
"publisher": "string",
|
117 |
-
"url": "string",
|
118 |
-
"page_url": "string",
|
119 |
-
}
|
120 |
df_labels = pd.read_csv(
|
121 |
-
|
122 |
)
|
123 |
df_metadata = pd.read_csv(
|
124 |
-
|
125 |
index_col=0,
|
126 |
-
dtype=
|
127 |
)
|
128 |
df_metadata["file"] = df_metadata.filepath.str.replace("/", "_")
|
129 |
df_metadata = df_metadata.set_index("file", drop=True)
|
130 |
df = df_labels.join(df_metadata)
|
131 |
df = df.reset_index()
|
132 |
-
|
133 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
134 |
box = ast.literal_eval(row["box"])
|
135 |
row["box"] = box
|
136 |
row.pop("filepath")
|
137 |
ocr = " ".join(ast.literal_eval(row["ocr"]))
|
138 |
row["ocr"] = ocr
|
139 |
image = row["file"]
|
140 |
-
row["image"] =
|
141 |
yield id_, row
|
|
|
14 |
"""Dataset of illustrated and non illustrated 19th Century newspaper ads."""
|
15 |
|
16 |
import ast
|
17 |
+
import os
|
18 |
import pandas as pd
|
19 |
import datasets
|
20 |
from PIL import Image
|
|
|
21 |
|
22 |
# TODO: Add BibTeX citation
|
23 |
# Find for instance the citation on arxiv or on the dataset repo/website
|
|
|
46 |
|
47 |
_URLS = "https://zenodo.org/record/5838410/files/images.zip?download=1"
|
48 |
|
49 |
+
_DTYPES = {
|
50 |
+
"page_seq_num": "int64",
|
51 |
+
"edition_seq_num": "int64",
|
52 |
+
"batch": "string",
|
53 |
+
"lccn": "string",
|
54 |
+
"score": "float64",
|
55 |
+
"place_of_publication": "string",
|
56 |
+
"name": "string",
|
57 |
+
"publisher": "string",
|
58 |
+
"url": "string",
|
59 |
+
"page_url": "string",
|
60 |
+
}
|
61 |
+
|
62 |
|
63 |
# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
|
64 |
class IllustratedAds(datasets.GeneratorBasedBuilder):
|
|
|
107 |
)
|
108 |
|
109 |
def _split_generators(self, dl_manager):
|
110 |
+
images = dl_manager.download_and_extract(_URLS)
|
111 |
+
annotations = dl_manager.download(
|
112 |
+
[
|
113 |
+
"https://zenodo.org/record/5838410/files/ads.csv?download=1",
|
114 |
+
"https://zenodo.org/record/5838410/files/sample.csv?download=1"
|
115 |
+
]
|
116 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
117 |
df_labels = pd.read_csv(
|
118 |
+
annotations[0], index_col=0
|
119 |
)
|
120 |
df_metadata = pd.read_csv(
|
121 |
+
annotations[1],
|
122 |
index_col=0,
|
123 |
+
dtype=_DTYPES,
|
124 |
)
|
125 |
df_metadata["file"] = df_metadata.filepath.str.replace("/", "_")
|
126 |
df_metadata = df_metadata.set_index("file", drop=True)
|
127 |
df = df_labels.join(df_metadata)
|
128 |
df = df.reset_index()
|
129 |
+
annotations = df.to_dict(orient="records")
|
130 |
+
return [
|
131 |
+
datasets.SplitGenerator(
|
132 |
+
name=datasets.Split.TRAIN,
|
133 |
+
gen_kwargs={
|
134 |
+
"images": images,
|
135 |
+
"annotations": annotations,
|
136 |
+
},
|
137 |
+
),
|
138 |
+
]
|
139 |
+
|
140 |
+
def _generate_examples(self, images, annotations):
|
141 |
+
for id_, row in enumerate(annotations):
|
142 |
box = ast.literal_eval(row["box"])
|
143 |
row["box"] = box
|
144 |
row.pop("filepath")
|
145 |
ocr = " ".join(ast.literal_eval(row["ocr"]))
|
146 |
row["ocr"] = ocr
|
147 |
image = row["file"]
|
148 |
+
row["image"] = os.path.join(images, image)
|
149 |
yield id_, row
|