Gradio 6.29; temp output dirs; theme to launch()

README.md

@@ -4,7 +4,7 @@ emoji: 📊
 colorFrom: purple
 colorTo: indigo
 sdk: gradio
-sdk_version: 5.30.0
+sdk_version: 6.29.0
 app_file: app.py
 pinned: false
 short_description: Support by Parquet, CSV, Jsonl, XLS

app.py

@@ -4,6 +4,8 @@ import json
 from io import BytesIO
 import requests
 import re
+import os
+import tempfile
 from openpyxl import Workbook
 
 def sanitize_value(val):
@@ -45,9 +47,9 @@ def dataset_converter(input_file, conversion_type, parquet_url):
     # Conversion: CSV to Parquet
     if conversion_type == "CSV to Parquet":
         if input_file is None or file_extension != "csv":
-            raise ValueError("For CSV to Parquet conversion, please upload a CSV file. 📄")
+            raise gr.Error("For CSV to Parquet conversion, please upload a CSV file. 📄")
         df = pd.read_csv(BytesIO(file_bytes))
-        output_file = "output.parquet"
+        output_file = os.path.join(tempfile.mkdtemp(), "output.parquet")
         df.to_parquet(output_file, index=False)
         converted_format = "Parquet"
         preview_str = df.head(10).to_string(index=False)
@@ -55,9 +57,9 @@ def dataset_converter(input_file, conversion_type, parquet_url):
     # Conversion: Parquet to CSV
     elif conversion_type == "Parquet to CSV":
         if input_file is None or file_extension != "parquet":
-            raise ValueError("For Parquet to CSV conversion, please upload a Parquet file. 📄")
+            raise gr.Error("For Parquet to CSV conversion, please upload a Parquet file. 📄")
         df = pd.read_parquet(BytesIO(file_bytes))
-        output_file = "output.csv"
+        output_file = os.path.join(tempfile.mkdtemp(), "output.csv")
         df.to_csv(output_file, index=False)
         converted_format = "CSV"
         preview_str = df.head(10).to_string(index=False)
@@ -65,9 +67,9 @@ def dataset_converter(input_file, conversion_type, parquet_url):
     # Conversion: CSV to JSONL
     elif conversion_type == "CSV to JSONL":
         if input_file is None or file_extension != "csv":
-            raise ValueError("For CSV to JSONL conversion, please upload a CSV file. 📄")
+            raise gr.Error("For CSV to JSONL conversion, please upload a CSV file. 📄")
         df = pd.read_csv(BytesIO(file_bytes), encoding='latin1')
-        output_file = "metadata.jsonl"
+        output_file = os.path.join(tempfile.mkdtemp(), "metadata.jsonl")
         total_data = []
         for index, row in df.iterrows():
             data = {}
@@ -89,14 +91,14 @@ def dataset_converter(input_file, conversion_type, parquet_url):
         if input_file is not None:
             df = pd.read_parquet(BytesIO(file_bytes))
         elif parquet_url:
-            response = requests.get(parquet_url)
+            response = requests.get(parquet_url, timeout=120)
             response.raise_for_status()
             df = pd.read_parquet(BytesIO(response.content))
             file_name = "from_url.parquet"
         else:
-            raise ValueError("For Parquet to JSONL conversion, please upload a file or provide a URL. 🌐")
+            raise gr.Error("For Parquet to JSONL conversion, please upload a file or provide a URL. 🌐")
         
-        output_file = "output.jsonl"
+        output_file = os.path.join(tempfile.mkdtemp(), "output.jsonl")
         def recursive_sanitize(val):
             if isinstance(val, bytes):
                 return val.decode("utf-8", errors="replace")
@@ -120,14 +122,14 @@ def dataset_converter(input_file, conversion_type, parquet_url):
         if input_file is not None:
             df = pd.read_parquet(BytesIO(file_bytes))
         elif parquet_url:
-            response = requests.get(parquet_url)
+            response = requests.get(parquet_url, timeout=120)
             response.raise_for_status()
             df = pd.read_parquet(BytesIO(response.content))
             file_name = "from_url.parquet"
         else:
-            raise ValueError("For Parquet to XLS conversion, please upload a file or provide a URL. 🌐")
+            raise gr.Error("For Parquet to XLS conversion, please upload a file or provide a URL. 🌐")
         
-        output_file = "output.xlsx"
+        output_file = os.path.join(tempfile.mkdtemp(), "output.xlsx")
         wb = Workbook(write_only=True)
         ws = wb.create_sheet()
         ws.append(list(df.columns))
@@ -139,7 +141,7 @@ def dataset_converter(input_file, conversion_type, parquet_url):
         preview_str = df.head(10).to_string(index=False)
     
     else:
-        raise ValueError("Invalid conversion type selected. ⚠️")
+        raise gr.Error("Invalid conversion type selected. ⚠️")
 
     info_message = (
         f"Input file: {file_name if file_name is not None else 'N/A'}\n"
@@ -148,7 +150,7 @@ def dataset_converter(input_file, conversion_type, parquet_url):
     )
     return output_file, info_message
 
-with gr.Blocks(title="Datasets Convertor", theme="Nymbo/Nymbo_Theme") as demo:
+with gr.Blocks(title="Datasets Convertor") as demo:
     gr.Markdown("# Datasets Convertor 🚀")
     gr.Markdown(
         "Upload a CSV or Parquet file (or provide a Parquet file URL for Parquet to JSONL/XLS conversion) "
@@ -178,4 +180,4 @@ with gr.Blocks(title="Datasets Convertor", theme="Nymbo/Nymbo_Theme") as demo:
         outputs=[output_file, preview]
     )
 
-demo.launch()
+demo.launch(theme="Nymbo/Nymbo_Theme")

requirements.txt

@@ -1,3 +1,4 @@
 pandas
-gradio
-openpyxl
+pyarrow
+openpyxl
+requests