Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gd
|
2 |
+
import polars as pl
|
3 |
+
|
4 |
+
|
5 |
+
data = pl.read_parquet("hf://datasets/polinaeterna/text_unnested/data/*.parquet")
|
6 |
+
min_min = data["min"].min()
|
7 |
+
min_max = data["max"].max()
|
8 |
+
|
9 |
+
|
10 |
+
def filter(min_value: min_min, max_value: min_max):
|
11 |
+
df = data.select((pl.col("min") >= min_value) & (pl.col("min") <= max_value)).to_pandas()
|
12 |
+
if df.shape[0] > 100:
|
13 |
+
return df.head(100)
|
14 |
+
return df
|
15 |
+
|
16 |
+
|
17 |
+
with gr.Blocks() as demo:
|
18 |
+
gr.Markdown("# 💫 Filter text datasets by string lengths distribution 💫")
|
19 |
+
min_value = gr.Slider(min_min, min_max, 0, step=1, label="Min min value")
|
20 |
+
max_value = gr.Slider(min_min, min_max, 0, step=1, label="Max min value")
|
21 |
+
btn = gr.Button("Get datasets ")
|
22 |
+
datasets = gr.DataFrame()
|
23 |
+
btn.click(filter, inputs=[min_value, max_value], outputs=[datasets])
|
24 |
+
|
25 |
+
demo.launch(debug=True)
|