07: VSCode Tutorial

In this tutorial we will briefly explore using an integrated development environment (IDE) for python programming. For a more extensive look at VSCode, see the full demo in the bonus examples. IDEs offer many advantages over Jupyter Notebooks, particularly for developing production code that will be reused many times and shared with others. The advantages of IDEs include:

  • Linting to catch errors, dead code, and other code quality issues

  • automatic docstring generation and other auto completion

  • code navigation (jumping to function definitions, etc.)

  • interactive debugging (including plotting)

  • support for automated testing and version control

  • unlike Notebooks, code is always executed from start to finish (as it would be by another user), thereby increasing the chances for reproducibility.

VSCode offers the above advantages, and many other features through Extensions (plugins) that can be created by anyone. VSCode is also free and works well with large files (GB in size). The Live Share plugin allows for real-time code collaboration and debugging with multiple people.

Getting started

If you haven’t yet, install VS Code here.

Launching VSCode

  • Open a fresh Miniforge prompt from the Start menu and navigate to the root folder for the class (containing the AGENDA.md file).

  • Activate the class python environment (conda activate pyclass) if needed.

  • Then type code .

  • If prompted, click “Yes, I trust the authors”

Alternatively

  • Launch VSCode and from the initial screen or File menu choose Open. Navigate to the root folder for the class and select it

Getting started

See here for an overview of the VSCode user interface.

Once VSCode is launched, click on the Extensions icon on the Activity Bar on the left. Install the following extensions:

  • Python

  • Python indent

  • Pylance

  • Jupyter

  • Code spell checker

  • autoDocstring

  • Rainbow CSV

You may find other cool extensions that you want too. A key indicator of an extension’s quality is the number of downloads.

Linting

Now let’s open the script notebooks/part0_python_intro/bonus_examples/solutions/Theis-exercise-solution.py.

  • On the left toolbar, select the Explorer icon (that looks like a stack of papers). In the file tree below, navigate to the script and select it.

  • This script looks pretty clean, with no obvious issues except a lot of comments

  • Try adding import pandas as pd to the imports. Notice how the pd is grayed out. The linter is flagging this as “dead code” that is declared but never used.

  • Below the imports, type pd.DataFrame(stuff). The pd above is no longer grayed out, because now it is used, but there is a yellow squiggly line beneath stuff, because this variable is being used but wasn’t declared.

  • Linters are valuable tools that make it easy to write clean code!

To make the script easier to work with, you may want to also turn on Word Wrap (from the View menu) and fold all of the code (control + k + 0).

Debugging

  • Place a break point on line 99, by return s in the theis() function, by clicking to the left of a line number (you should see a red dot).

  • If needed, select the pyclass` environment as the Python interpreter. Go to View –> Command Palette, then type Python: Select Interpreter. Choose the option with (pyclass)`` from the dropdown menu.

  • With the Python script (e.g.Theis-exercise-solution.py) tab selected, you should see some version numbers followed by (pyclass) in the bottom right of the VS Code window. Note that you can also click here to change the Python environment.

  • Then go to either Run --> Start Debugging or click on the debug icon in the Activity Bar and choose Run and Debug. Choose Python File if prompted for a configuration. The debugger should run to the break point.

The debug working directory

Import pathlib and type pathlib.Path.cwd() in the debug console. Note that the current directory is the root folder for the class (where we launched VSCode). VSCode is structured around projects, which include everything in a folder that was opened (and any subfolders). By default, the working directory for debugging is set at the root level for the project. We can change this (and other debugging settings), by creating a configuration file called launch.json, which lives inside of a .vscode/ folder at the root level of the project.

  • After stopping the debugging session, make a default launch.json by

    • clicking on the debug icon in the Activity Bar and then

    • create a launch.json file.

    • Choose Python Debugger and then Python File if prompted for a configuration.

    • A new tab will open up with launch.json.

    • Add "cwd": "${fileDirname}" at the bottom (don’t forget the preceding comma!) so that the file looks like this:

      {
    // Use IntelliSense to learn about possible attributes.
    // Hover to view descriptions of existing attributes.
    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Current File",
            "type": "python",
            "request": "launch",
            "program": "${file}",
            "console": "integratedTerminal",
            "cwd": "${fileDirname}"
        }
    ]
}

  • simply making a new file at .vscode/launch.json (within the root folder for the class), and pasting in the above text also works.

  • run the debugger again. Path.cwd() should show the part0_python_intro folder.

The Run and Debug view

The Run and Debug view on the left (available via the Activity Bar) shows the current variables in the namespace, and an additional “Watch” panel where we can add specific variables that we want to observe as the script executes.

  • Add a watch for the s variable by clicking the + sign on the upper right part of the panel.

  • Now put a breakpoint at the return statement of the theis_xy function.

  • continue with the debugger (via the Run menu or debugging toolbar). Execution should stop again at the return statement where we just placed the breakpoint. Notice that s has changed in the Watch panel. There are also now two layers in the “Call Stack” box below– one for the theis_xy() function that we are in, and one for the enclosing main script.

  • step out to the main script by clicking the second item in the call stack (labeled <module>). Notice that s has changed again, reflecting the current state of s in that namespace. Notice also that the time-drawdown plot from Step 3 in the original notebook has popped up.

Plotting

During a debug session, as long as matplotlib has been imported, one can make plots via the Debug Console or by putting plotting code in the script.

If the plot isn’t showing up

  • Try making the plot by starting with fig, ax = plt.subplots() syntax (instead of plt.plot())

  • Then in the debug console enter fig.show() or plt.draw()

  • Alternativey, try plt.pause(1)

  • On Windows, look in the task bar for a matplotlib icon drawing (or press Alt + Tab to show the open windows)

Code navigation

In the main part of the script, right click on one of the function calls (e.g. theis) and choose Go to Definition. VSCode should take you to where the function is defined. Right clicking on theis after the def and then Go to References opens a “peek” window with a list of all of the times theis() is called. Clicking on one of them takes you to that location in the script.

Navigation works across modules and packages too. If you right click on np.meshgrid in the main part of the script and Go to Definition, VSCode takes you to the relevant code in numpy.