Development Setup for `xradio` with `casatools` I/O Backend

This guide describes how to set up a local development environment for xradio, specifically configured to use the casatools I/O backend for interacting with CASA MeasurementSets (MS) and Images. This setup provides a unified development environment that can combine RADPS-targeted new libraries (e.g. xradio, astroviper) with legacy CASA/Pipeline codebase, facilitating experimentation, prototyping, and benchmarking.

Prerequisites:

Git
Conda, e.g. Miniforge

1. Clone Repositories

Clone the necessary repositories. If you only need to develop xradio itself, cloning it might be sufficient. However, for full local development, especially if modifying dependencies, clone the viper libraries as well.

# Clone xradio (required)
git clone https://github.com/casangi/xradio.git
cd xradio

# Clone viper libraries (optional, but recommended for full local dev)
# Ensure these are cloned in a location accessible for editable installs,
# e.g., alongside the xradio directory.
# git clone https://github.com/casangi/astroviper.git ../astroviper
# git clone https://github.com/casangi/toolviper.git ../toolviper
# git clone https://github.com/casangi/graphviper.git ../graphviper

(Adjust paths like ../astroviper based on your preferred directory structure)

2. Create and Activate Conda Environment

Create a dedicated Conda environment using the provided YAML configuration. This ensures all dependencies, including casatools, are installed correctly.

Create a file named xradio_casatools_dev.yaml (or similar) with the following content:

name: xradio_casatools_dev # Changed name slightly for clarity

channels:
  - conda-forge
  - defaults

dependencies:
  # Core environment tools
  - conda-ecosystem-user-package-isolation
  - python=3.12
  - ipython
  - pip

  # CASA dependencies via pip using NRAO's repository
  - pip:
    # Specify NRAO's pip index for casatools/casatasks
    - --extra-index-url https://casa-pip.nrao.edu/repository/pypi-group/simple
    - casatools == 6.7.0.* # Specify desired casatools version
    - casatasks == 6.7.0.* # Optional: if casatasks needed

    # Install xradio in editable mode from the current directory clone
    # Include optional dependencies for interactive use and documentation
    - -e .[interactive,docs]

    # --- Add these lines if you cloned the viper libraries for local dev ---
    # Install viper libraries in editable mode from their respective clones
    # Adjust paths relative to where you run the pip install command (e.g., from within the xradio dir)
    # - -e ../astroviper/.
    # - -e ../toolviper/.
    # - -e ../graphviper/.
    # --------------------------------------------------------------------

Create and activate the environment:

# Navigate to the directory containing the YAML file if needed
conda env create --file xradio_casatools_dev.yaml
conda activate xradio_casatools_dev

(Use conda env update –file xradio_casatools_dev.yaml –prune if the environment already exists and you want to update it)

3. Handle Potential `python-casacore` Conflicts

The casatools package includes its own casacore build and Python binding. If python-casacore (a separate package often installed from conda-forge or PyPI) is also present, it can lead to namespace conflicts. The current workaround for activiting the casatools-backend is to uninstall python-casacore in the Python development environment when using casatools.

# Run this after activating the environment
pip uninstall python-casacore

4. Running Unit Tests Locally

You can run unit tests using pytest. Execute these commands from the root directory of your xradio clone.

Run a specific test function:
```
python -m pytest -vvs tests/unit/test_image.py::TestCasacoreToXdsToCasacore::test_pixels_and_mask
```
(Note: Class name might differ slightly depending on test structure, adjust if needed)

Run all tests in a specific file:

python -m pytest -vvs tests/unit/test_image.py

Run all xradio tests (not fully working yet):
```
python -m pytest -vvs tests/
```

5. Optional: Setup Parallelization (Dask Cluster)

For performance testing or parallel processing workflows, you can set up a local Dask cluster using toolviper.

Important Note: casatools (and potentially python-casacore) are generally not thread-safe. When using Dask with xradio tasks involving these libraries, configure your Dask cluster to use processes instead of threads to avoid race conditions and crashes.

Start a local Dask cluster using processes:

# Example Python script or interactive session
from toolviper.dask.client import local_client

# Use processes=True or set n_workers > 0 and threads_per_worker=1
# serial_execution=False enables parallelism
viper_client = local_client(
    cores=8,             # Total number of cores (processes) for the cluster
    memory_limit="4GB",  # Memory limit per worker process
    serial_execution=False,
)

print(viper_client) # Display cluster information

Adjust logging verbosity (optional):

import toolviper.utils.logger as logger

viperlog = logger.get_logger()
viperlog.setLevel('DEBUG') # Set desired level (e.g., INFO, DEBUG, WARNING)

Development Setup for xradio with casatools I/O Backend