Skip to content

TPI: copilot tools for the python envs ext #338

@eleanorjboyd

Description

@eleanorjboyd

Refs: #280

Complexity: 4

Author: @eleanorjboyd

Create Issue

Create Issue


✅ Test Plan Item: Manual Verification of python_environment_tool and python_install_package_tool

🧪 Objective

Validate that python_environment_tool and python_install_package_tool provide accurate and useful information for a Python environment in VS Code. Ensure they support a variety of common developer tasks and return expected results.


⚙️ Setup

  • Install the Python Environments extension in VS Code.
  • Open a folder with a Python environment (virtual environment, Conda environment, or system Python). Use existing or ask copilot to create one for you and see how it goes!

🧪 Test Steps

ask questions to copilot in agent mode to test the tools and tool calling!

[ ] Prompt the model in ways to use the python_environment_tool. Ideas include:

  • Get environment details (type, Python version, installed packages, run command).
  • Ask for the Python version specifically.
  • Ask how to run a file or script with Python and check the returned run command.
  • Use it in a notebook and confirm it detects the notebook’s environment correctly.
  • Try in a data science context and verify it still returns appropriate information.

[ ] Prompt the model in ways to use the python_install_package_tool. Ideas include:

  • Install a new package (e.g., requests).
  • Install a specific version of a package (e.g., requests==2.31.0).
  • Update an existing package.
  • Install packages needed for a project (simulate by asking to install a list like numpy, pandas, scikit-learn).

✅ Expected Outcome

  • python_environment_tool correctly provides environment type, Python version, run command, and installed packages.
  • python_install_package_tool successfully installs or updates the specified packages.
  • Agent responses summarize environment info clearly and suggest relevant actions.
  • Outputs are correct and usable in typical Python and data science workflows.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions