Update app.py

app.py

@@ -1,58 +1,93 @@
 import gradio as gr
 import pandas as pd
 from io import BytesIO
+import chardet
 
-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
+def detect_encoding(file_bytes):
+    """Detect the encoding of a file using chardet"""
+    # Only use a sample of the file for detection to improve performance
+    result = chardet.detect(file_bytes[:10000])
+    return result['encoding']
+
+def convert_file(input_file, conversion_type, encoding_option):
     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
-        file_name = input_file
-        with open(file_name, "rb") as f:
-            file_bytes = f.read()
-    
-    file_extension = file_name.lower().split('.')[-1]
-    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.")
+        # Check if a file was uploaded
+        if input_file is None:
+            return None, "Please upload a file."
+        
+        # Determine if input_file is a file-like object or a file path string
+        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
+            file_name = input_file
+            with open(file_name, "rb") as f:
+                file_bytes = f.read()
+        
+        file_extension = file_name.lower().split('.')[-1]
+        df = None
+        output_file = None
+        converted_format = None
+        
+        # Handle encoding for CSV files
+        if encoding_option == "Auto-detect":
+            encoding = detect_encoding(file_bytes)
+        else:
+            encoding = encoding_option
+        
+        # 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."
+            
+            # Try with the selected/detected encoding
+            try:
+                df = pd.read_csv(BytesIO(file_bytes), encoding=encoding)
+            except UnicodeDecodeError:
+                # If auto-detection fails, try a few common encodings
+                common_encodings = ['latin1', 'iso-8859-1', 'cp1252']
+                for enc in common_encodings:
+                    try:
+                        df = pd.read_csv(BytesIO(file_bytes), encoding=enc)
+                        encoding = enc  # Update the successful encoding
+                        break
+                    except UnicodeDecodeError:
+                        continue
+                if df is None:
+                    return None, f"Failed to decode the CSV file. Auto-detected encoding was '{encoding}'. Please try selecting a specific encoding."
+            
+            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=encoding)
+            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"
+            f"Encoding used: {encoding}\n"
+            f"Total rows: {len(df)}\n"
+            f"Total columns: {len(df.columns)}\n\n"
+            f"Preview (Top 10 Rows):\n{preview}"
+        )
+        return output_file, info_message
     
-    # 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"
-        f"Total rows: {len(df)}\n"
-        f"Total columns: {len(df.columns)}\n\n"
-        f"Preview (Top 10 Rows):\n{preview}"
-    )
-    return output_file, info_message
+    except Exception as e:
+        return None, f"Error during conversion: {str(e)}"
 
 # Enhanced custom CSS for a more visually appealing interface
 custom_css = """
@@ -184,6 +219,24 @@ h2 {
     background: linear-gradient(to right, transparent, #ddd, transparent);
     margin: 25px 0;
 }
+
+.error-message {
+    color: #d93025;
+    background-color: #fce8e6;
+    padding: 10px;
+    border-radius: 8px;
+    margin-top: 10px;
+    font-size: 0.9rem;
+}
+
+.success-message {
+    color: #188038;
+    background-color: #e6f4ea;
+    padding: 10px;
+    border-radius: 8px;
+    margin-top: 10px;
+    font-size: 0.9rem;
+}
 """
 
 with gr.Blocks(css=custom_css, title="DataFormat Converter") as demo:
@@ -200,13 +253,14 @@ with gr.Blocks(css=custom_css, title="DataFormat Converter") as demo:
                 <h3>How It Works</h3>
                 <div class="instruction-step">1. Upload your CSV or Parquet file</div>
                 <div class="instruction-step">2. Select the conversion direction</div>
-                <div class="instruction-step">3. Click "Convert" and download your transformed file</div>
+                <div class="instruction-step">3. Choose encoding (or leave as auto-detect)</div>
+                <div class="instruction-step">4. Click "Convert" and download your transformed file</div>
             </div>
             
             <div class="info-section">
                 <div class="info-tag">Fast Conversion</div>
                 <div class="info-tag">Data Preview</div>
-                <div class="info-tag">No Size Limits</div>
+                <div class="info-tag">Multi-Encoding Support</div>
                 <div class="info-tag">Maintains Structure</div>
             </div>
             """)
@@ -240,6 +294,11 @@ with gr.Blocks(css=custom_css, title="DataFormat Converter") as demo:
                 value="CSV to Parquet",
                 elem_classes=["conversion-radio"]
             )
+            encoding_option = gr.Dropdown(
+                choices=["Auto-detect", "utf-8", "latin1", "iso-8859-1", "cp1252", "utf-16"],
+                value="Auto-detect",
+                label="Select CSV Encoding"
+            )
             convert_button = gr.Button("Convert Now", elem_classes=["convert-button"])
             gr.HTML('</div>')  # Close the file-box div
             
@@ -263,8 +322,22 @@ with gr.Blocks(css=custom_css, title="DataFormat Converter") as demo:
     
     convert_button.click(
         fn=convert_file, 
-        inputs=[input_file, conversion_type], 
+        inputs=[input_file, conversion_type, encoding_option], 
         outputs=[output_file, preview]
     )
 
-demo.launch()
+    # Add dependency handling to show/hide encoding options based on conversion type
+    def update_encoding_visibility(conversion_type):
+        if conversion_type == "CSV to Parquet":
+            return gr.update(visible=True)
+        else:
+            return gr.update(visible=False)
+    
+    conversion_type.change(
+        fn=update_encoding_visibility,
+        inputs=conversion_type,
+        outputs=encoding_option
+    )
+
+if __name__ == "__main__":
+    demo.launch()