Update app.py

app.py

@@ -41,20 +41,25 @@ def convert_parquet_to_jsonl_polars(input_file, output_dir, override=False):
         print(f"Data written to {output_file_path}")
 
 def convert_parquet_to_jsonl(parquet_filename, jsonl_filename):
-    # Read the parquet file
-    df = pd.read_parquet(parquet_filename)
-
-    # Convert the dataframe to a JSON string and handle Unicode characters and forward slashes
-    json_str = df.to_json(orient='records', lines=True, force_ascii=False)
+    try:
+        # Read the parquet file
+        df = pd.read_parquet(parquet_filename)
+        logger.info(f"Read Parquet file {parquet_filename} successfully.")
 
-    # Replace escaped forward slashes if needed
-    json_str = json_str.replace('\\/', '/')
+        # Convert the dataframe to a JSON string and handle Unicode characters and forward slashes
+        json_str = df.to_json(orient='records', lines=True, force_ascii=False)
+        logger.info(f"Converted Parquet file to JSON string.")
 
-    # Write the modified JSON string to the JSONL file
-    with open(jsonl_filename, 'w', encoding='utf-8') as file:
-        file.write(json_str)
+        # Replace escaped forward slashes if needed
+        json_str = json_str.replace('\\/', '/')
 
-    print(f"Data saved to {jsonl_filename}")
+        # Write the modified JSON string to the JSONL file
+        with open(jsonl_filename, 'w', encoding='utf-8') as file:
+            file.write(json_str)
+        logger.info(f"Data saved to {jsonl_filename}")
+    except Exception as e:
+        logger.error(f"Failed to convert Parquet to JSONL: {e}")
+        raise
 
 # Function to count lines in a JSONL file
 def count_lines_in_jsonl(file_path):