Update app.py

app.py

@@ -1,19 +1,26 @@
 import gradio as gr
 import pandas as pd
+import chardet
 from io import BytesIO
 
+def detect_encoding(file_bytes):
+    """Detect the encoding of the file."""
+    # Use chardet to detect encoding
+    result = chardet.detect(file_bytes)
+    return result['encoding']
+
 def convert_file(input_file, conversion_type):
     # Check if a file was uploaded
     if input_file is None:
-        raise ValueError("Please upload a file.")
-
-    # Determine if input_file is a file-like object or a file path string.
+        return None, "Please upload a file."
+    
+    # Read the file content
     try:
         # Try reading from file-like object
         file_bytes = input_file.read()
         file_name = input_file.name
     except AttributeError:
-        # If there's an AttributeError, treat input_file as a file path.
+        # If there's an AttributeError, treat input_file as a file path
         file_name = input_file
         with open(file_name, "rb") as f:
             file_bytes = f.read()
@@ -22,34 +29,62 @@ def convert_file(input_file, conversion_type):
     df = None
     output_file = None
     converted_format = None
-
-    # Conversion: CSV to Parquet
-    if conversion_type == "CSV to Parquet":
-        if file_extension != "csv":
-            raise ValueError("For CSV to Parquet conversion, please upload a CSV file.")
-        df = pd.read_csv(BytesIO(file_bytes))
-        output_file = "output.parquet"
-        df.to_parquet(output_file, index=False)
-        converted_format = "Parquet"
-    # Conversion: Parquet to CSV
-    elif conversion_type == "Parquet to CSV":
-        if file_extension != "parquet":
-            raise ValueError("For Parquet to CSV conversion, please upload a Parquet file.")
-        df = pd.read_parquet(BytesIO(file_bytes))
-        output_file = "output.csv"
-        df.to_csv(output_file, index=False)
-        converted_format = "CSV"
-    else:
-        raise ValueError("Invalid conversion type selected.")
-
-    # Generate a preview of the top 10 rows
-    preview = df.head(10).to_string(index=False)
-    info_message = (
-        f"Input file: {file_name}\n"
-        f"Converted file format: {converted_format}\n\n"
-        f"Preview (Top 10 Rows):\n{preview}"
-    )
-    return output_file, info_message
+    
+    try:
+        # Conversion: CSV to Parquet
+        if conversion_type == "CSV to Parquet":
+            if file_extension != "csv":
+                return None, "For CSV to Parquet conversion, please upload a CSV file."
+            
+            # Detect the encoding of the CSV file
+            encoding = detect_encoding(file_bytes)
+            
+            # Try to read with detected encoding
+            try:
+                df = pd.read_csv(BytesIO(file_bytes), encoding=encoding)
+            except Exception as e:
+                # If that fails, try with other common encodings
+                for enc in ['utf-8', 'latin1', 'iso-8859-1', 'cp1252']:
+                    try:
+                        df = pd.read_csv(BytesIO(file_bytes), encoding=enc)
+                        encoding = enc
+                        break
+                    except:
+                        continue
+                if df is None:
+                    return None, f"Failed to read CSV with any encoding. Error: {str(e)}"
+            
+            output_file = "output.parquet"
+            df.to_parquet(output_file, index=False)
+            converted_format = "Parquet"
+            
+        # Conversion: Parquet to CSV
+        elif conversion_type == "Parquet to CSV":
+            if file_extension != "parquet":
+                return None, "For Parquet to CSV conversion, please upload a Parquet file."
+            
+            df = pd.read_parquet(BytesIO(file_bytes))
+            output_file = "output.csv"
+            df.to_csv(output_file, index=False, encoding='utf-8')
+            converted_format = "CSV"
+        else:
+            return None, "Invalid conversion type selected."
+        
+        # Generate a preview of the top 10 rows
+        preview = df.head(10).to_string(index=False)
+        info_message = (
+            f"Input file: {file_name}\n"
+            f"Converted file format: {converted_format}\n"
+        )
+        if conversion_type == "CSV to Parquet":
+            info_message += f"Detected encoding: {encoding}\n"
+        
+        info_message += f"\nPreview (Top 10 Rows):\n{preview}"
+        
+        return output_file, info_message
+    
+    except Exception as e:
+        return None, f"Error during conversion: {str(e)}"
 
 # Custom CSS for a modern and sleek look
 custom_css = """
@@ -93,7 +128,11 @@ with gr.Blocks(css=custom_css, title="CSV <-> Parquet Converter") as demo:
         with gr.Column(scale=1):
             input_file = gr.File(label="Upload CSV or Parquet File")
         with gr.Column(scale=1):
-            conversion_type = gr.Radio(choices=["CSV to Parquet", "Parquet to CSV"], label="Conversion Type")
+            conversion_type = gr.Radio(
+                choices=["CSV to Parquet", "Parquet to CSV"], 
+                label="Conversion Type",
+                value="CSV to Parquet"  # Set default value
+            )
     
     convert_button = gr.Button("Convert", elem_classes=["gradio-button"])
     
@@ -102,5 +141,13 @@ with gr.Blocks(css=custom_css, title="CSV <-> Parquet Converter") as demo:
         preview = gr.Textbox(label="Preview (Top 10 Rows)", lines=15)
     
     convert_button.click(fn=convert_file, inputs=[input_file, conversion_type], outputs=[output_file, preview])
+    
+    gr.Markdown("""
+    ### Notes:
+    - This converter can handle various CSV encodings
+    - Parquet files are always encoded in UTF-8
+    - The preview shows only the first 10 rows of data
+    """)
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()