Sandboxes
Python Sandbox
Parameters
- Script: The Python code to run within the sandbox
- Max Time: Max execution time in milliseconds
Returns
Returns the code output
The pythonSandbox allows you to execute Python code within the mAIstro environment. It is designed for running and testing dynamic Python scripts, especially those generated from prompts or templates. Ensure you use extractCode after an LLM call to handle valid code execution within the sandbox.
Example Usage 1
Let's create a script to calculate a falling object velocity:
<< name: timeInput, prompt: "Enter time in seconds:">>
{{ LLM | prompt: "Write Python to calculate the velocity of an object after free fall for a given time which is: << name: timeInput, prompt: false >>. Make sure to output just the code with no explanations" | cache: "true" }}=>{{ extractCode }}=>{{ variable | name: "code" }}
{{ pythonSandbox | script: "<< name: code, prompt: false >>" | maxTime: "1000" }}=>{{ variable | name: "codeResult" | mode: "" | value: "" }}
The velocity of an object after free fall for << name: timeInput, prompt: false >> seconds is << name: codeResult, prompt: false >>
Will yield something like:
Example Usage 2
Let's create a script to calculate circle area based on its radius:
<< name: radiusInput, prompt: "Enter the radius of the circle:">>
{{ LLM | prompt: "Write Python to calculate the area of a circle given the radius: << name: radiusInput, prompt: false >>. Make sure to output just the code with no explanations" | cache: "true" }}=>{{ extractCode }}=>{{ variable | name: "code" }}
{{ pythonSandbox | script: "<< name: code, prompt: false >>" | maxTime: "1000" }}=>{{ variable | name: "codeResult" | mode: "" | value: "" }}
The area of the circle with radius << name: radiusInput, prompt: false >> is << name: codeResult, prompt: false >>
Will yield something like:
JavaScript Sandbox
Parameters
- Script: The JavaScript code to run within the sandbox
- Max Time: Max execution time in milliseconds
Returns
Returns the code output
The javascriptSandbox enables the execution of JavaScript code within the mAIstro environment. It is optimized for testing dynamic JavaScript, especially code generated from prompts or templates. Leverage native JavaScript features like the fetch API for making asynchronous requests and handling data from external sources. As with the Python Sandbox, use extractCode after LLM calls to ensure the code is properly extracted and executed.
Example Usage 1
Let's fetch some data from an external API:
{{ LLM | prompt: "Write JavaScript to fetch data from the the Json Placeholder endpoint. https://jsonplaceholder.typicode.com/todos/1" | cache: "true" }}=>{{ extractCode }}=>{{ variable | name: "code" }}
{{ javascriptSandbox | script: "<< name: code, prompt: false >>" | maxTime: "1000" }}=>{{ variable | name: "codeResult" }}
<< name: codeResult, prompt: false >>
Will yield something like:
Example Usage 2
Let's use browser localStorage to store a user input value:
<< name: storageKey, prompt: "Enter a key to store in localStorage:">>
<< name: storageValue, prompt: "Enter a value to store in localStorage:">>
{{ LLM | prompt: "Write JavaScript to store << name: storageValue, prompt: false >> in localStorage under the key << name: storageKey, prompt: false >>, and then retrieve and log it to the console. Make sure to output just the code with no explanations" | cache: "true" }}=>{{ extractCode }}=>{{ variable | name: "code" }}
{{ javascriptSandbox | script: "<< name: code, prompt: false >>" | maxTime: "1000" }}=>{{ variable | name: "codeResult" }}
The value in '<< name: storageKey, prompt: false >>' localStorage key is '<< name: codeResult, prompt: false >>'