Add basic Pixi config in pyproject.toml and developer instructions

.gitattributes

@@ -7,3 +7,5 @@ sasview/test/* linguist-language=Other
 *.h text eol=lf
 *.cpp text eol=lf
 *.rst text eol=lf
+# SCM syntax highlighting & preventing 3-way merges
+pixi.lock merge=binary linguist-language=YAML linguist-generated=true -diff

.gitignore

@@ -81,3 +81,5 @@ tests.log
 /installers/dist
 *.exe
 installers/credits.html
+# pixi environments
+.pixi/*

INSTALL.md

@@ -3,7 +3,7 @@
 Whether you're installing SasView to use as a tool for your research or
 because you're wanting to work on the code, it is recommended that you
 work inside a Python virtual environment of some sort.
-A `venv` or a `conda` are both popular choices.
+A `venv` or a `Pixi` are both popular choices.
 
 ## Installing SasView as a User
 
@@ -18,39 +18,49 @@ environment
 
 ## Making a SasView Development Environment
 
-If you're familiar with working with developing in Python, then the very quick version is:
+### Prerequisite: Obtaining the source code
+
+Obtain the SasView source using `git`. This step is the same whether you are using `venv` or Pixi to create the development environment. You will likely need to coordinate
+updates to `sasdata` and `sasmodels`. The
+[`bumps`](https://github.com/bumps/bumps) and
+[`periodictable`](https://github.com/python-periodictable/periodictable)
+packages are far more loosely coupled, but depending on what you are
+doing you may also want them as development packages.
 
 ```shell
 # clone the repository
 git clone https://github.com/sasview/sasdata/
 git clone https://github.com/sasview/sasmodels/
 git clone https://github.com/sasview/sasview/
+```
+
+### Develop using `venv`
 
+If you're familiar with working with developing in Python, then the very quick version is:
+
+```shell
 cd sasview
 
-# create the virtual environment
+# Create the virtual environment
 python -m venv .venv
 # .venv\Scripts\activate & REM Windows: activate environment
 . .venv/bin/activate  # Linux/Mac: activate environment
 
-# install repositories in editable/developer mode in the venv
-# use "python -m ..." to ensure the venv's pip is used
+# Install repositories in editable/developer mode in the venv
+# Use "python -m ..." to ensure the venv's pip is used
 python -m pip install -e ../sasdata
 python -m pip install -e ../sasmodels
 python -m pip install -e .[dev,test]
 
-# test if sasview launches
+# Test if sasview launches
 python -m sas
+
+# To deactivate the virtual environment when finished developing
+deactivate
 ```
 
 Step by step, that is:
 
-1.  Obtain the SasView source using `git`. You will likely need to coordinate
-    updates to `sasdata` and `sasmodels`. The
-    [`bumps`](https://github.com/bumps/bumps) and
-    [`periodictable`](https://github.com/python-periodictable/periodictable)
-    packages are far more loosely coupled, but depending on what you are
-    doing you may also want them as development packages.
 1.  Create a Python virtual environment in the `.venv` directory.
 1.  Activate the `.venv` so that Python and its modules from the venv are used.
     Note that the particular syntax above works for the `bash` and `zsh` shells under Linux, Windows and macOS;
@@ -87,7 +97,46 @@ debugging software, e.g.:
 - [VS Code](https://code.visualstudio.com/docs/python/environments)
 - [PyCharm](https://www.jetbrains.com/help/pycharm/creating-virtual-environment.html)
 
-### Pre-Commit Hooks for Linting
+### Develop using Pixi
+
+An alternative to `venv` is the package management tool Pixi. If Pixi is not
+installed, follow the instructions [here](https://pixi.prefix.dev/latest/#installation).
+
+The very quick version for developing using Pixi is:
+
+```shell
+# Enter the developer environment
+# (This will create (or reuse) a local Pixi environment with all required dependencies.)
+cd sasview
+pixi shell
+
+# Start the GUI
+python -m sas
+
+# Run tests
+pixi run test
+
+# Exit the developer environment when finished developing
+exit
+```
+
+In more detail, the steps are:
+
+1. Create (or reuse) a local Pixi environment in `.pixi` and enter the shell
+   into the developer environment. The first time it will take a while to
+   download and unpack all dependencies.
+   Note that `sasdata`, `sasmodels` and `sasview` are installed in editable mode,
+   meaning that any code changes will be available the next time you run the
+   program.
+1. Run SasView! As an alternative to typing `python -m sas`, you can simply
+   type `sasview`.
+
+Almost all the modules that SasView needs are available as precompiled modules
+on PyPI, including numpy, scipy, h5py, pyside6. A handful of Python-only
+modules will be built into wheels on your local machine. Installing the
+dependencies should be a one-off task.
+
+## Pre-Commit Hooks for Linting
 
 The SasView, SasData and SasModels repositories include [pre-commit hooks](https://pre-commit.com/), which can be set up to enable linting to be run on the code. A linter is a tool that can detect programming errors, bugs, stylistic errors, etc. SasView uses the [Ruff](https://docs.astral.sh/ruff/) package for linting. Ruff is able to warn about a wide range of possible errors, and in some cases apply automatic fixes for them.
 

pyproject.toml

@@ -237,4 +237,19 @@ ignore = ["E501",  # line too long (leave to formatter)
 "sas" = ["sasdata", "sasmodels"]
 
 [tool.ruff.lint.isort]
-section-order = ["future", "standard-library", "third-party", "sas", "first-party", "local-folder"]
+section-order = ["future", "standard-library", "third-party", "sas", "first-party", "local-folder"]
+
+[tool.pixi.workspace]
+channels = ["conda-forge"]
+platforms = ["linux-64", "win-64", "osx-arm64"]
+
+[tool.pixi.pypi-dependencies]
+sasview = { path = ".", editable = true, extras = ["test", "dev"]}
+sasdata   = { path = "../sasdata", editable = true }
+sasmodels = { path = "../sasmodels", editable = true }
+# Uncomment if you want to install bumps and/or periodictable in editable mode:
+# bumps   = { path = "../bumps", editable = true }
+# periodictable = { path = "../periodictable", editable = true }
+
+[tool.pixi.tasks]
+test = { description = "Run the test suite", cmd = "python -m pytest -s test" }