More pyproject.toml updates + PR readiness.

setup.py now has the bare minimum required,
the dynamic version and compilation.
This also removes a duplicated/unused
no cython pyproject.toml,
and reverts feature actions to not run extra tests.
Notebook tests still not passing,
but that will be handled in a separate PR.

Author: sserita

.github/workflows/feature-branches.yml

@@ -18,7 +18,7 @@ jobs:
     strategy:
       # fail-fast: true is OK here
       matrix:
-        os: [macos-latest, ubuntu-latest, windows-latest] # TEMP: Try latest mac again # No Mac
+        os: [ubuntu-latest, windows-latest]
         python-version: [3.9, '3.12'] # Only extremal Python versions
     uses: ./.github/workflows/reuseable-main.yml
     name: Run pyGSTi tests
@@ -27,8 +27,8 @@ jobs:
       python-version: ${{ matrix.python-version }}
       use-cython: 'true' # Only test environment with Cython
       run-unit-tests: 'true'
-      run-extra-tests: 'true' # No integration tests # TEMP: Try integration tests
-      run-notebook-tests: 'true' # No notebook tests # TEMP: Try notebook tests
+      run-extra-tests: 'false' # No integration tests
+      run-notebook-tests: 'false' # No notebook tests
 
 
 

MANIFEST.in

@@ -1,5 +1,5 @@
 include README.md CONTRIBUTING.md LICENSE
-include pyproject.toml pyproject.toml.no_cython
+include pyproject.toml
 include install_locally.py
 include requirements.txt
 include optional-requirements.txt

pyproject.toml

@@ -6,8 +6,9 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pygsti"
-dynamic = ["version", "readme", "optional-dependencies"]
+dynamic = ["version"]
 description = 'A python implementation of Gate Set Tomography'
+readme = "README.md"
 authors = [
     {email = 'pygsti@sandia.gov'},
     {name = 'Erik Nielsen'},
@@ -48,6 +49,48 @@ classifiers=[
     "Operating System :: Unix"
 ]
 
+[project.optional-dependencies]
+diamond_norm = [
+    'cvxopt',
+    'cvxpy'
+]
+evolutionary_optimization = ['deap']
+extensions = ['cython']
+html_reports = ['jinja2', 'MarkupSafe']
+interpygate = ['csaps']
+linting = [
+    'autopep8',
+    'flake8'
+]
+memory_profiling = ['psutil']
+mongodb = ['pymongo']
+msgpack = ['msgpack']
+multiprocessor = ['mpi4py']
+notebooks = [
+    'ipython',
+    'notebook',
+    'jupyter_server'
+]
+pytorch = ['torch']
+report_pdf_figures = ['matplotlib']
+reports = ['pygsti[html_reports,report_pdf_figures]']
+testing_no_cython_mpi = [
+    'pytest',
+    'pytest-xdist',
+    'pytest-cov',
+    'nbval',
+    'packaging',
+    'zmq',
+    'seaborn',
+    'ply',
+    'cirq-core',
+    'pygsti[diamond_norm,interpygate,memory_profiling,msgpack,notebooks,pytorch,reports]'
+]
+testing_no_cython = ['pygsti[multiprocessor,testing_no_cython_mpi]']
+testing = ['pygsti[extensions,testing_no_cython]']
+no_mpi = ['pygsti[evolutionary_optimization,extensions,linting,mongodb,testing_no_cython_mpi]']
+complete = ['pygsti[multiprocessor,no_mpi]']
+
 [project.urls]
 Homepage = 'https://www.pygsti.info'
 Repository = 'https://github.com/sandialabs/pyGSTi'

pyproject.toml.no_cython

@@ -12,94 +12,6 @@
     from distutils.extension import Extension
     from distutils.command.build_ext import build_ext
 
-
-    
-descriptionTxt = """\
-Gate set tomography (GST) is a quantum tomography protocol that provides full characterization of a quantum logic device
-(e.g. a qubit).  GST estimates a set of quantum logic gates and (simultaneously) the associated state preparation and
-measurement (SPAM) operations.  GST is self-calibrating.  This eliminates a key limitation of traditional quantum state
-and process tomography, which characterize either states (assuming perfect processes) or processes (assuming perfect
-state preparation and measurement), but not both together.  Compared with benchmarking protocols such as randomized
-benchmarking, GST provides much more detailed and accurate information about the gates, but demands more data.  The
-primary downside of GST has been its complexity.  Whereas benchmarking and state/process tomography data can be analyzed
-with relatively simple algorithms, GST requires more complex algorithms and more fine-tuning (linear GST is an exception
-that can be implemented easily).  pyGSTi addresses and eliminates this obstacle by providing a fully-featured, publicly
-available implementation of GST in the Python programming language.
-
-The primary goals of the pyGSTi project are to:
-
-- provide efficient and robust implementations of Gate Set Tomography algorithms;
-- allow straightforward interoperability with other software;
-- provide a powerful high-level interface suited to inexperienced programmers, so that
-  common GST tasks can be performed using just one or two lines of code;
-- use modular design to make it easy for users to modify, customize, and extend GST functionality.
-"""
-
-# Extra requirements
-extras = {
-    'pytorch' : ['torch'],
-    'diamond_norm': [
-        'cvxopt',
-        'cvxpy'
-    ],
-    'memory_profiling': ['psutil'],
-    'multiprocessor': ['mpi4py'],
-    'evolutionary_optimization': ['deap'],
-    'report_pickling': ['pandas'],
-    'report_pdf_figures': ['matplotlib'],
-    'html_reports': ['jinja2', 'MarkupSafe'],
-    'notebooks': [
-        'ipython',
-        'notebook',
-        'jupyter_server'
-    ],
-    'mongodb': ['pymongo'],
-    'msgpack': ['msgpack'],
-    'extensions': ['cython'],
-    'linting': [
-        'autopep8',
-        'flake8'
-    ],
-    'interpygate': ['csaps'],
-    'testing': [
-        'pytest',
-        'pytest-xdist',
-        'pytest-cov',
-        'nbval',
-        'csaps',
-        'cvxopt',
-        'cvxpy',
-        'cython',
-        'matplotlib',
-        'mpi4py',
-        'msgpack',
-        'packaging',
-        'pandas',
-        'psutil',
-        'zmq',
-        'jinja2',
-        'seaborn',
-        'scipy',
-        'ply',
-        'cirq-core',
-        'notebook',
-        'ipython',
-        'jupyter_server',
-        'torch'
-    ]
-}
-
-# Add `complete' target, which will install all extras listed above
-extras['complete'] = list({pkg for req in extras.values() for pkg in req})
-
-# Add `no_mpi' target, identical to `complete' target but without mpi4py,
-# which is unavailable in some common environments.
-extras['no_mpi'] = [e for e in extras['complete'] if e != 'mpi4py']
-
-# Add testing_no_cython target, identical to `testing` but no cython
-extras['testing_no_cython'] = [e for e in extras['testing'] if e != 'cython']
-
-
 # Configure setuptools_scm to build a custom version (for more info,
 # see https://stackoverflow.com/a/78657279 and https://setuptools-scm.readthedocs.io/en/latest/extending)
 # If on a clean release, it uses no local scheme
@@ -150,10 +62,8 @@ def build_extensions(self):
 
 def setup_with_extensions(extensions=None):
     setup(
-        long_description=descriptionTxt,
         use_scm_version={'version_scheme': 'no-guess-dev', 'version_file': "pygsti/_version.py", 'local_scheme': custom_version},
         cmdclass={'build_ext': build_ext_compiler_check},
-        extras_require=extras,
         ext_modules=extensions or [],
         packages=find_packages(),
         package_data={