Spaces:
Sleeping
Sleeping
Commit
•
4b1895a
1
Parent(s):
ec81263
add card data
Browse files
app.py
CHANGED
@@ -4,6 +4,7 @@ import gradio as gr
|
|
4 |
import pandas as pd
|
5 |
from datasets import ClassLabel, Dataset, Image
|
6 |
from httpx import Client
|
|
|
7 |
|
8 |
client = Client()
|
9 |
USER_DATA = {}
|
@@ -42,6 +43,31 @@ def check_user_data():
|
|
42 |
# return pd.read_csv(StringIO(resp.text)).columns.tolist()
|
43 |
|
44 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
45 |
def push_annotations_to_hub(project_id, input_column, input_column_type, label_column):
|
46 |
headers = {"Authorization": f'Token {USER_DATA["api_key"]}'}
|
47 |
resp = client.get(
|
@@ -56,6 +82,10 @@ def push_annotations_to_hub(project_id, input_column, input_column_type, label_c
|
|
56 |
if input_column_type == "image":
|
57 |
ds = ds.cast_column(input_column, Image())
|
58 |
ds.push_to_hub(USER_DATA["hub_dataset_id"], token=USER_DATA["hub_api_key"])
|
|
|
|
|
|
|
|
|
59 |
return ds.to_pandas().head(5)
|
60 |
|
61 |
|
|
|
4 |
import pandas as pd
|
5 |
from datasets import ClassLabel, Dataset, Image
|
6 |
from httpx import Client
|
7 |
+
from huggingface_hub import DatasetCard
|
8 |
|
9 |
client = Client()
|
10 |
USER_DATA = {}
|
|
|
43 |
# return pd.read_csv(StringIO(resp.text)).columns.tolist()
|
44 |
|
45 |
|
46 |
+
def convert_value(value: int) -> str:
|
47 |
+
if value < 1_000:
|
48 |
+
return "n<1K"
|
49 |
+
elif value < 10_000:
|
50 |
+
return "1K<n<10K"
|
51 |
+
elif value < 100_000:
|
52 |
+
return "10K<n<100K"
|
53 |
+
elif value < 1_000_000:
|
54 |
+
return "100K<n<1M"
|
55 |
+
elif value < 10_000_000:
|
56 |
+
return "1M<n<10M"
|
57 |
+
elif value < 100_000_000:
|
58 |
+
return "10M<n<100M"
|
59 |
+
elif value < 1_000_000_000:
|
60 |
+
return "100M<n<1B"
|
61 |
+
elif value < 10_000_000_000:
|
62 |
+
return "1B<n<10B"
|
63 |
+
elif value < 100_000_000_000:
|
64 |
+
return "10B<n<100B"
|
65 |
+
elif value < 1_000_000_000_000:
|
66 |
+
return "100B<n<1T"
|
67 |
+
else:
|
68 |
+
return "n>1T"
|
69 |
+
|
70 |
+
|
71 |
def push_annotations_to_hub(project_id, input_column, input_column_type, label_column):
|
72 |
headers = {"Authorization": f'Token {USER_DATA["api_key"]}'}
|
73 |
resp = client.get(
|
|
|
82 |
if input_column_type == "image":
|
83 |
ds = ds.cast_column(input_column, Image())
|
84 |
ds.push_to_hub(USER_DATA["hub_dataset_id"], token=USER_DATA["hub_api_key"])
|
85 |
+
card = DatasetCard.load(USER_DATA["hub_dataset_id"])
|
86 |
+
card.data.tags = ["label-studio-exported"]
|
87 |
+
card.data.size_categories = [convert_value(len(ds))]
|
88 |
+
card.push_to_hub(USER_DATA["hub_dataset_id"], repo_type="dataset")
|
89 |
return ds.to_pandas().head(5)
|
90 |
|
91 |
|