Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,34 +1,27 @@
|
|
1 |
-
import requests
|
2 |
-
import pandas as pd
|
3 |
import gradio as gr
|
4 |
-
import
|
5 |
-
|
6 |
-
@spaces.GPU
|
7 |
-
def download_file(url):
|
8 |
-
response = requests.get(url)
|
9 |
-
return response.content
|
10 |
|
11 |
-
def convert_parquet_to_jsonl(parquet_file):
|
12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
jsonl_data = df.to_json(orient='records', lines=True)
|
14 |
-
return jsonl_data
|
15 |
-
|
16 |
-
def main(url):
|
17 |
-
parquet_file = download_file(url)
|
18 |
-
jsonl_data = convert_parquet_to_jsonl(parquet_file)
|
19 |
with open("output.jsonl", "w") as f:
|
20 |
f.write(jsonl_data)
|
21 |
return "output.jsonl"
|
22 |
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
)
|
31 |
-
demo.launch()
|
32 |
|
33 |
if __name__ == "__main__":
|
34 |
-
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
+
import pandas as pd
|
3 |
+
import requests
|
|
|
|
|
|
|
|
|
4 |
|
5 |
+
def convert_parquet_to_jsonl(parquet_file=None, parquet_url=None):
|
6 |
+
if parquet_file is not None:
|
7 |
+
df = pd.read_parquet(parquet_file.name)
|
8 |
+
elif parquet_url is not None:
|
9 |
+
response = requests.get(parquet_url)
|
10 |
+
df = pd.read_parquet(response.content)
|
11 |
+
else:
|
12 |
+
raise ValueError("Either parquet_file or parquet_url must be provided")
|
13 |
jsonl_data = df.to_json(orient='records', lines=True)
|
|
|
|
|
|
|
|
|
|
|
14 |
with open("output.jsonl", "w") as f:
|
15 |
f.write(jsonl_data)
|
16 |
return "output.jsonl"
|
17 |
|
18 |
+
demo = gr.Interface(
|
19 |
+
fn=convert_parquet_to_jsonl,
|
20 |
+
inputs=[gr.File(label="Parquet File"), gr.Textbox(label="Parquet File URL")],
|
21 |
+
outputs=[gr.File(label="JSONL Output")],
|
22 |
+
title="Parquet to JSONL Converter",
|
23 |
+
description="Convert a Parquet file to JSONL format from a downloadable link or file upload"
|
24 |
+
)
|
|
|
|
|
25 |
|
26 |
if __name__ == "__main__":
|
27 |
+
demo.launch()
|