Install Required Software or Run a Docker Image

This notebook installs the Pixi software environment required to run the other notebooks in the NISAR GCOV Cookbook.

Overview¶

Prerequisites
(Option 1) Install the environment with the Pixi pacakage manager
(Option 2) Run a Docker Image with the software environment already installed
Summary
Resources and references

1. Prerequisites¶

Prerequisite	Importance	Notes
Option 1: The Pixi package manager must be installed on the system running Jupyter Lab	Necessary	Pixi may not be supported on all Jupyter Hubs
Option 2: Docker must be installed and running on the system running Jupyter Lab	Necessary	Run the Docker image locally or in a Jupyter Hub
UW Scientific Software Engineering Center’s Pixi guide	Helpful	This is a great resource to get quickly get started with Pixi

Rough Notebook Time Estimate: 3 minutes

2. (Option 1) Install the environment with the Pixi pacakage manager¶

2a. Create a `work_dir` context manager¶

This is useful when you want to run a command in a working directory and then automatically change back to your original directory

import os
from contextlib import contextmanager
from pathlib import Path

@contextmanager
def work_dir(new_dir):
    old_dir = os.getcwd()
    os.chdir(new_dir)
    try:
        yield
    finally:
        os.chdir(old_dir)

2b. Install the `isce3` environment with Pixi¶

with work_dir(Path.cwd().parent):
    !pixi install -e isce3

2c. Register the `isce3` environment’s Python kernel with `ipykernel`¶

env_name = "isce3"
display_name = f'"{env_name} (Python)"'

!pixi run -e isce3 python -m ipykernel install \
  --user \
  --name $env_name \
  --display-name $display_name

2d. Ensure notebook shell commands run in the `isce3` environment¶

This ensures that shell commands executed from inside a notebook with ! run in the notebook kernel’s environment. This works by launching the entire Jupyter kernel process inside the Pixi environment, so the kernel’s PATH, environment variables, and Python executable all come from that Pixi environment, not from the parent JupyterLab environment.

from jupyter_client.kernelspec import KernelSpecManager
import json

ksm = KernelSpecManager()
spec = ksm.get_kernel_spec(env_name)
kernel_dir = Path(spec.resource_dir)
kernel_json = kernel_dir / "kernel.json"

data = json.loads(kernel_json.read_text())
orig_argv = data.get("argv", [])

new_argv = [
    "pixi",
    "run",
    "--manifest-path",
    str(Path.cwd().parent),
    "-e",
    env_name,
] + orig_argv

data["argv"] = new_argv
kernel_json.write_text(json.dumps(data, indent=2))
print(f"Updated kernel.json at {kernel_json} with Pixi wrapper.")
print("argv:", data["argv"])

2e. (Optional) Delete the environment and remove its `kernelspec`¶

# Uncomment and run the code below to delete your pixi environment

# with work_dir(Path.cwd().parent):
#     !pixi clean
#     !jupyter kernelspec remove isce3 -y

3. (Option 2) Run a Docker Image with the software environment already installed¶

Coming soon

4. Summary¶

Now that you have installed the software environment or have the Docker image running, make sure you have access to the data. You will then be ready to run the remaining notebooks in the NISAR GCOV Cookbook.

5. Resources and references¶

References¶

UW SSEC

Author: Alex Lewandowski