VisiData MCP Server

Instructions

Implementation Reference

• src/visidata_mcp/server.py:1012-1012 (registration)

  The parse_skills_column tool is registered with the MCP server using the @mcp.tool() decorator

  @mcp.tool()

• src/visidata_mcp/server.py:1013-1096 (handler)

  The parse_skills_column function implements the logic to parse comma-separated skills from a column and create one-hot encoding columns for each unique skill

  def parse_skills_column(file_path: str, skills_column: str, output_path: Optional[str] = None) -> str:
      """
      Parse comma-separated skills into individual skills and create one-hot encoding.
      
      Args:
          file_path: Path to the data file
          skills_column: Column name containing comma-separated skills
          output_path: Optional path to save the processed data
      
      Returns:
          Information about the parsed skills data
      """
      try:
          import pandas as pd
          from pathlib import Path
          
          # Load the data
          file_extension = Path(file_path).suffix.lower()
          if file_extension == '.csv':
              df = pd.read_csv(file_path)
          elif file_extension == '.json':
              df = pd.read_json(file_path)
          elif file_extension in ['.xlsx', '.xls']:
              df = pd.read_excel(file_path)
          elif file_extension == '.tsv':
              df = pd.read_csv(file_path, sep='\t')
          else:
              df = pd.read_csv(file_path)
          
          if skills_column not in df.columns:
              return f"Error: Column '{skills_column}' not found in data"
          
          # Parse skills and create one-hot encoding
          all_skills = set()
          
          # Extract all unique skills
          for skills_str in df[skills_column].dropna():
              if pd.isna(skills_str):
                  continue
              skills = [skill.strip() for skill in str(skills_str).split(',') if skill.strip()]
              all_skills.update(skills)
          
          all_skills = sorted(list(all_skills))
          
          # Create one-hot encoding for each skill
          skills_df = df.copy()
          for skill in all_skills:
              skills_df[f"skill_{skill.replace(' ', '_').replace('-', '_').lower()}"] = 0
          
          # Fill in the one-hot encoding
          for idx, skills_str in enumerate(df[skills_column]):
              if pd.isna(skills_str):
                  continue
              skills = [skill.strip() for skill in str(skills_str).split(',') if skill.strip()]
              for skill in skills:
                  col_name = f"skill_{skill.replace(' ', '_').replace('-', '_').lower()}"
                  if col_name in skills_df.columns:
                      skills_df.loc[idx, col_name] = 1
          
          # Save processed data if output path provided
          if output_path:
              if output_path.endswith('.csv'):
                  skills_df.to_csv(output_path, index=False)
              elif output_path.endswith('.json'):
                  skills_df.to_json(output_path, orient='records', indent=2)
              elif output_path.endswith(('.xlsx', '.xls')):
                  skills_df.to_excel(output_path, index=False)
              else:
                  skills_df.to_csv(output_path, index=False)
          
          result = {
              "skills_parsed": True,
              "original_column": skills_column,
              "unique_skills_count": len(all_skills),
              "unique_skills": all_skills[:20],  # First 20 skills for preview
              "rows_processed": len(df),
              "new_columns_added": len(all_skills),
              "output_file": output_path if output_path else None
          }
          
          return json.dumps(result, indent=2)
          
      except Exception as e:
          return f"Error parsing skills: {str(e)}\n{traceback.format_exc()}"