Code Executor¶
Executes user-provided Python code against a single input in a controlled environment. Captures stdout/stderr to an execution log and returns the computed result along with success status. The environment restricts dangerous operations and only exposes a safe subset of built-ins and common utility modules.
Usage¶
Use this node when you need lightweight, custom transformation or logic that operates on a single piece of data without writing a full custom node. Typical workflows include quick data munging, conditional branching based on computed values, and in-line calculations on text, numbers, or objects passed from prior nodes.
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| python_code | True | STRING | The Python code to run. Access the incoming value via the variable 'input_data'. You may return a value explicitly with 'return', or assign to a variable named 'result' or 'output'. If none of these are used, the last evaluated expression is used as the result. | # Access input data data = input_data # Transform and return result = data * 2 if isinstance(data, (int, float)) else str(data).upper() return result |
| timeout_seconds | True | FLOAT | Suggested timeout duration for the execution. Note: current implementation does not enforce this at runtime. | 10.0 |
| input_data | False | * | The value to be processed by your Python code. Can be any data type (number, string, object, etc.). | 42 |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| result | * | The computed value from your code. Determined by an explicit return, or the variables 'result'/'output', or the last expression's value. | 84 |
| execution_log | STRING | A human-readable log combining captured stdout and warnings to help debug the code execution. | Execution completed successfully \| Output: processed |
| success | BOOLEAN | Indicates whether the code executed without errors. | true |
Important Notes¶
- Execution model: If your code contains 'return', it is wrapped and executed to honor the return value. Without 'return', the node returns 'result' or 'output' variables if set, otherwise the last evaluated expression.
- Restricted environment: Dangerous patterns are blocked (e.g., 'import os', 'import sys', 'import subprocess', 'import shutil', 'exec(', 'eval(', 'import', 'open('). Generic import statements are not supported because 'import' is not exposed.
- Available modules and built-ins: Safe built-ins and modules are provided: len, str, int, float, bool, list, dict, tuple, set, min, max, sum, abs, round, sorted, reversed, enumerate, zip, map, filter, range, print, type, isinstance, hasattr, getattr, setattr, and modules json, time, math, random, datetime.
- Timeout: The 'timeout_seconds' input is not currently enforced; long-running code can still run until completion or error.
- Stdout/stderr capture: print output and warnings are captured and included in 'execution_log'.
- Input flexibility: 'input_data' accepts any type; ensure your code defensively handles unexpected types.
Troubleshooting¶
- Invalid Python code or blocked operations: If 'success' is false with a message like 'Invalid Python code' or references a dangerous pattern, remove disallowed imports/functions and ensure your code parses syntactically.
- No result returned: If 'result' is null, ensure you either 'return' a value or assign to 'result' or 'output'. Without these, only the last expression value is used.
- Import errors: Standard 'import x' will fail due to missing 'import'. Use the pre-exposed modules (json, time, math, random, datetime) or avoid imports.
- Type errors: If your code assumes a specific type, add checks like 'isinstance(input_data, dict)' and handle alternatives to avoid runtime exceptions.
- Unexpected long runs: The 'timeout_seconds' setting is not enforced. Refactor code to avoid blocking calls or heavy loops.