added csv-agent-default in next.js

main.py

@@ -93,7 +93,7 @@ async def execute_code(
         code_request: Dictionary containing 'code' key with the Python code to execute
         
     Returns:
-        dict: Dictionary containing execution results with all plots, output, variables, any errors, and Excel files
+        dict: Dictionary containing execution results with all plots, output, variables, and any errors
     """
     if not token_valid:
         raise HTTPException(status_code=401, detail="Not authenticated")
@@ -118,9 +118,9 @@ async def execute_code(
         response_content = {
             "success": result['error'] is None,
             "output": result['output'],
+            # "variables": result['variables'],
             "plots": [],
-            "html_charts": result.get('html_charts', []),
-            "excel_files": []
+            "html_charts": result.get('html_charts', [])
         }
         
         # Add all plots if they exist
@@ -134,23 +134,10 @@ async def execute_code(
                 for index, plot_data in enumerate(result['plots'])
             ]
         
-        # Add Excel files if they exist
-        if result.get('excel_files'):
-            response_content["excel_files"] = [
-                {
-                    "filename": file['filename'],
-                    "content": file['content'],
-                    "content_type": file['content_type']
-                }
-                for file in result['excel_files']
-            ]
-        
         # Add error information if exists
         if result['error']:
             response_content["error"] = result['error']
         
-        logger.info(f"Response content: {response_content}")
-        
         return response_content
         
     except Exception as e:

python_code_interpreter_service.py

@@ -6,7 +6,6 @@ import os
 import base64
 from pathlib import Path
 import uuid
-import time
 import numpy as np
 import pandas as pd
 import matplotlib
@@ -26,8 +25,6 @@ import warnings
 import plotly.express as px
 import plotly.graph_objects as go
 from plotly.io import to_html
-import openpyxl
-
 
 
 def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
@@ -54,7 +51,6 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
     plot_base64 = []
     variables = {}
     html_charts = []
-    excel_files = []
     
     # Monkey patch plt.show() to save figures
     original_show = plt.show
@@ -62,6 +58,7 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
     def custom_show():
         for i, fig in enumerate(plt.get_fignums()):
             figure = plt.figure(fig)
+            # Save plot to bytes buffer instead of file
             buf = io.BytesIO()
             figure.savefig(buf, format='png', bbox_inches='tight')
             buf.seek(0)
@@ -72,51 +69,27 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
     original_plotly_show = go.Figure.show
     
     def custom_plotly_show(fig, *args, **kwargs):
+        # Generate unique filename
         chart_id = str(uuid.uuid4())
         filename = f"chart_{chart_id}.html"
         filepath = charts_dir / filename
+        
+        # Save as HTML
         html = to_html(fig, include_plotlyjs='cdn')
         with open(filepath, 'w', encoding='utf-8') as f:
             f.write(html)
-        html_charts.append(filename)
-        fig._grid_ref = None
-        return None
-    
-    # Monkey patch pd.ExcelWriter to capture Excel files
-    original_ExcelWriter = pd.ExcelWriter
-    
-    def custom_ExcelWriter(*args, **kwargs):
-        # Force openpyxl engine if no engine specified
-        if 'engine' not in kwargs:
-            kwargs['engine'] = 'openpyxl'
-        # Create in-memory file
-        excel_buffer = io.BytesIO()
-        kwargs['path'] = excel_buffer
-        writer = original_ExcelWriter(*args, **kwargs)
         
-        # Add cleanup and capture logic
-        def save():
-            writer.close()
-            excel_buffer.seek(0)
-            excel_content = base64.b64encode(excel_buffer.read()).decode('utf-8')
-            filename = args[0] if len(args) > 0 else kwargs.get('path', 'output.xlsx')
-            if isinstance(filename, Path):
-                filename = filename.name
-            excel_files.append({
-                'filename': filename,
-                'content': excel_content,
-                'content_type': 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet'
-            })
+        # Add to html_charts list
+        html_charts.append(filename)
         
-        writer.save = save
-        return writer
+        # Close the figure to free memory
+        fig._grid_ref = None  # Help with memory cleanup
+        return None
     
     try:
-        # Patch ExcelWriter before execution
-        pd.ExcelWriter = custom_ExcelWriter
-        
-        # Create execution context
+        # Create a comprehensive execution context with all common data science libraries
         exec_globals = {
+            # Core libraries
             'np': np,
             'pd': pd,
             'plt': plt,
@@ -125,25 +98,33 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
             'stats': stats,
             'sklearn': sklearn,
             'tabulate': tabulate,
+            
+            # Plotly libraries
             'px': px,
             'go': go,
+            
+            # Date/time libraries
             'datetime': datetime,
             'parser': parser,
             'pytz': pytz,
+            
+            # Utility
             'os': os,
             'sys': sys,
             'warnings': warnings,
             'json': json,
-            'pd.ExcelWriter': pd.ExcelWriter,
-            'time': time,
-            'openpyxl': openpyxl,
+            
+            # File paths
             'DATA_DIR': data_dir,
             'CHARTS_DIR': charts_dir,
+            
+            # Provided DataFrame
             'df': df,
+            
             '__builtins__': __builtins__,
         }
         
-        # Add sklearn components
+        # Add common sklearn components
         from sklearn import (
             datasets, preprocessing, model_selection, 
             linear_model, ensemble, metrics, svm,
@@ -162,15 +143,18 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
             'feature_selection': feature_selection,
         })
         
-        # Replace show methods
+        # Replace plt.show with custom implementation
         plt.show = custom_show
+        
+        # Replace plotly figure's show method
         go.Figure.show = custom_plotly_show
         
-        # Execute code
+        # Execute code and capture output
         with contextlib.redirect_stdout(stdout):
+            # First execute to get variables
             exec(code, exec_globals)
             
-            # Capture variables
+            # Capture all variables that were created
             for name, value in exec_globals.items():
                 if not name.startswith('_') and name not in [
                     'np', 'pd', 'plt', 'sns', 'sm', 'stats', 'sklearn',
@@ -178,8 +162,7 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
                     'os', 'sys', 'warnings', 'json', 'DATA_DIR', 'CHARTS_DIR',
                     'datasets', 'preprocessing', 'model_selection', 'linear_model',
                     'ensemble', 'metrics', 'svm', 'decomposition', 'cluster',
-                    'feature_selection', 'df', '__builtins__', 'pd.ExcelWriter',
-                    'time', 'openpyxl'
+                    'feature_selection', 'df'  # Exclude our parameter from variables
                 ]:
                     variables[name] = value
             
@@ -191,12 +174,12 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
             "traceback": traceback.format_exc()
         }
     finally:
-        # Restore original functions
+        # Restore original plt.show
         plt.show = original_show
+        # Restore original plotly show
         go.Figure.show = original_plotly_show
-        pd.ExcelWriter = original_ExcelWriter
         
-    # Convert variables to serializable formats
+    # Convert various objects to serializable formats
     def convert_objects(obj):
         if isinstance(obj, (np.ndarray, np.generic)):
             return obj.tolist() if obj.size > 1 else obj.item()
@@ -227,15 +210,16 @@ def execute_python_code(code: str, df: pd.DataFrame = None) -> Dict[str, Any]:
             return f"<function {obj.__name__}>"
         return obj
     
-    processed_vars = {
-        k: convert_objects(v) 
-        for k, v in variables.items()
-    }
+    processed_vars = {}
+    for k, v in variables.items():
+        try:
+            processed_vars[k] = convert_objects(v)
+        except Exception as e:
+            processed_vars[k] = f"<Unable to serialize: {str(e)}>"
     
     return {
         'output': output,
         'error': error,
         'plots': plot_base64,
-        'html_charts': html_charts,
-        'excel_files': excel_files,
+        'html_charts': html_charts
     }