Refactor/stepup4

.gitignore

@@ -26,6 +26,7 @@ ENV/
 .envrc
 .env
 .venv
+uv.lock
 
 # Build artifacts, distribution / packaging
 src/stacie/_version.py

.pre-commit-config.yaml

@@ -29,11 +29,11 @@ repos:
     - id: remove-crlf
       exclude_types: [binary]
 - repo: https://github.com/python-jsonschema/check-jsonschema
-  rev: 0.37.2
+  rev: 0.37.4
   hooks:
     - id: check-github-workflows
 - repo: https://github.com/astral-sh/ruff-pre-commit
-  rev: v0.15.12
+  rev: v0.15.20
   hooks:
     - id: ruff-format
     - id: ruff-check
@@ -44,7 +44,7 @@ repos:
     - id: nb-clean
       args: ["--remove-empty-cells", "--preserve-cell-metadata", "tags", "--"]
 - repo: https://github.com/DavidAnson/markdownlint-cli2
-  rev: v0.22.1
+  rev: v0.23.0
   hooks:
     - id: markdownlint-cli2
       exclude: '^test'

docs/data/README.md

@@ -8,7 +8,7 @@ To rerun the MD simulations, you need files that are not included in the Zenodo
 such as the LAMMPS and OpenMM input files, and the force fields.
 These are available in the STACIE Git repository: <https://github.com/molmod/stacie>
 
-Each directory under `docs/data` contains a `plan.py` script that can be executed with `stepup boot`
+Each directory under `docs/data` contains a `plan.py` script that can be executed with `sb`
 to run all simulations in parallel.
 See the `README.md` file in each directory for more details.
 

docs/data/lammps_lj3d/README.md

@@ -37,7 +37,7 @@ but only to read the outputs.
 All LAMMPS simulations can be executed efficiently as follows:
 
 ```bash
-stepup boot -n 1.0
+sb -j 1.0
 ```
 
 This will create input files from the templates `template-init.lammps` and `template-ext.lammps`

docs/data/lammps_lj3d/job.sh

@@ -8,4 +8,4 @@
 cd ../../../../../
 . activate
 cd ${SLURM_SUBMIT_DIR}
-time stepup boot -n ${SLURM_CPUS_PER_TASK}
+time sb -j ${SLURM_CPUS_PER_TASK}

docs/data/lammps_lj3d/plan.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 from runlammps import runlammps
-from stepup.core.api import mkdir, render_jinja, static
+from stepup.core.api import render_jinja, static
 
 
 def plan_extension(ireplica: int, part: int, additional_steps: int):
@@ -17,7 +17,6 @@ def plan_extension(ireplica: int, part: int, additional_steps: int):
         Number of additional steps to run in this extension.
     """
     name = f"sims/replica_{ireplica:04d}_part_{part:02d}"
-    mkdir(f"{name}/")
     render_jinja(
         "template-ext.lammps",
         {
@@ -33,12 +32,10 @@ def plan_extension(ireplica: int, part: int, additional_steps: int):
 
 
 static("runlammps.py", "template-init.lammps", "template-ext.lammps")
-mkdir("sims/")
 nreplica = 100
 for ireplica in range(nreplica):
     # Initial production run
     name_i = f"sims/replica_{ireplica:04d}_part_00"
-    mkdir(f"{name_i}/")
     render_jinja("template-init.lammps", {"seed": ireplica + 1}, f"{name_i}/in.lammps")
     runlammps(f"{name_i}/")
 

docs/data/lammps_lj3d/runlammps.py

@@ -6,19 +6,19 @@
 """
 
 import argparse
-import subprocess
 
 from path import Path
-from stepup.core.api import amend, runpy
+from stepup.core.api import amend, run
+from stepup.core.extapi import run_subprocess
 
 __all__ = ("runlammps",)
 
 
 def runlammps(workdir: str, inp: list[str] = ()):
     workdir = Path(workdir)
-    runpy(
+    run(
         f"./runlammps.py {workdir}",
-        inp=["runlammps.py", workdir / "in.lammps", *inp],
+        inp=[workdir / "in.lammps", *inp],
         out=[workdir / "log.txt"],
     )
 
@@ -29,11 +29,15 @@ def main(argv: list[str] | None = None):
     for file in args.rundir.files():
         if file.name != "in.lammps":
             file.remove()
-    subprocess.run(
-        ["lmp", "-i", "in.lammps", "-l", "log.txt", "-sc", "none"],
-        cwd=args.rundir,
-        check=True,
+    run_subprocess(
+        "lmp -i in.lammps -l log.txt -sc none",
+        workdir=args.rundir,
     )
+    # This goes against good practices: ideally amend should come before run_subprocess.
+    # However, LAMMPS output files are not known in advance in general.
+    # This depends on the details of the input file.
+    # Here, we take the lazy approach of amending all files
+    # that are not the input file or the log file.
     extra_out = [path for path in args.rundir.files() if path.name not in ["in.lammps", "log.txt"]]
     amend(out=extra_out)
 

docs/data/openmm_salt/README.md

@@ -51,7 +51,7 @@ to avoid unit confusion in the worked example.
 All OpenMM simulations can be executed efficiently on a single compute node with StepUp:
 
 ```bash
-OPENMM_CPU_THREADS=1 stepup boot -n 1.0
+OPENMM_CPU_THREADS=1 sb -j 1.0
 ```
 
 This will run in parallel the Jupyter notebooks `initial.ipynb` and `extension.ipynb`

docs/data/openmm_salt/job.sh

@@ -10,4 +10,4 @@ cd ../../../../../
 cd ${SLURM_SUBMIT_DIR}
 export OPENMM_CPU_THREADS=1
 # export OPENMM_DEFAULT_PLATFORM=CPU
-time stepup boot -n ${SLURM_CPUS_PER_TASK}
+time sb -j ${SLURM_CPUS_PER_TASK}

docs/data/openmm_salt/plan.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-from stepup.core.api import mkdir, runpy, static
+from stepup.core.api import run, shq, static
 from stepup.reprep.api import execute_papermill
 
 
@@ -29,19 +29,19 @@ def plan_extension(seed: int, part: int, nstep: int):
 
 def plan_extract(seed: int, part: int, ensemble: str):
     """Extract essentials from OpenMM output (DCD and CSV) and store as NPZ."""
-    runpy(
-        "./${inp} ${out}",
-        inp=[
-            "extract.py",
-            f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.csv",
-            f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.dcd",
-            f"output/sim{seed:04d}_part{part:02d}_{ensemble}_last.pdb",
-        ],
-        out=f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.npz",
+    inp = [
+        f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.csv",
+        f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.dcd",
+        f"output/sim{seed:04d}_part{part:02d}_{ensemble}_last.pdb",
+    ]
+    out = f"output/sim{seed:04d}_part{part:02d}_{ensemble}_traj.npz"
+    run(
+        f"./extract.py {shq(inp)} {shq(out)}",
+        inp=inp,
+        out=out,
     )
 
 
-mkdir("output")
 static("bhmtf.py", "initial.ipynb", "extension.ipynb", "utils.py", "extract.py")
 for seed in range(100):
     # Initial production runs

docs/data/plan.py

@@ -2,12 +2,11 @@
 """StepUp Plan to create ZIP files for Zenodo after all simulations are done."""
 
 from path import Path
-from stepup.core.api import copy, glob, mkdir, runpy, runsh, static
+from stepup.core.api import copy, glob, run, static
 from stepup.reprep.api import compile_typst, make_inventory, sync_zenodo, zip_inventory
 
 # Convert JupyText files to Jupyter notebooks
-static("../source/", "../source/examples/", "../source/examples/matplotlibrc", "preprocess.py")
-mkdir("preprocessed/")
+static("preprocess.py", "../source/examples/matplotlibrc")
 copy("../source/examples/matplotlibrc", "./")
 paths_ipynb = ["./matplotlibrc"]
 for path_py in glob("../source/examples/*.py"):
@@ -16,25 +15,22 @@
         paths_ipynb.append(path_py.name)
     else:
         path_pre = Path("preprocessed") / path_py.name
-        runpy("./${inp} ${out}", inp=["preprocess.py", path_py], out=path_pre)
+        run(f"./preprocess.py {path_py} {path_pre}", inp=path_py, out=path_pre)
         path_ipynb = path_py.name.replace(".py", ".ipynb")
         paths_ipynb.append(path_ipynb)
-        runsh("jupytext --to notebook ${inp} --output ${out}", inp=path_pre, out=path_ipynb)
+        run(
+            f"jupytext --to notebook {path_pre} --output {path_ipynb}", inp=path_pre, out=path_ipynb
+        )
 
 # LAMMPS LJ3D example
-static("zenodo.yaml", "zenodo.md")
-static("lammps_lj3d/")
-static("lammps_lj3d/sims/")
-glob("lammps_lj3d/sims/replica_????_part_??/")
 lammps_paths_txt = glob("lammps_lj3d/sims/replica_????_part_??/nv?_*.txt")
 lammps_paths_yaml = glob("lammps_lj3d/sims/replica_????_part_??/info.yaml")
 
 # OpenMM Molten Salt example
-static("openmm_salt/", "openmm_salt/output/")
 openmm_paths_npz = glob("openmm_salt/output/*.npz")
 
 # Cloud cover example data
-static("cloud-cover/", "cloud-cover/download.sh")
+static("cloud-cover/download.sh")
 cloudcover_paths = glob("cloud-cover/*.csv")
 
 # Compile the README
@@ -53,5 +49,8 @@
     *cloudcover_paths,
 ]
 make_inventory(*paths, "inventory.txt")
+
+# Sync with Zenodo
+static("zenodo.yaml", "zenodo.md")
 zip_inventory("inventory.txt", "examples.zip")
 sync_zenodo("zenodo.yaml")

docs/source/development/setup.md

@@ -2,28 +2,41 @@
 
 ## Repository, Tests and Documentation Build
 
-It is assumed that you have previously installed Python, Git, pre-commit and direnv.
+It is assumed that you have previously installed
+[Python](https://www.python.org/),
+[uv](https://docs.astral.sh/uv/),
+[Git](https://git-scm.com/),
+[pre-commit](https://pre-commit.com/) and
+[direnv](https://direnv.net/).
 A local installation for testing and development can be installed as follows:
 
 ```bash
 git clone git@github.com:molmod/stacie.git
 cd stacie
+uv sync --extra docs,tests,dev
 pre-commit install
-python -m venv venv
-echo 'source venv/bin/activate' > .envrc
+echo 'source .venv/bin/activate' > .envrc
 direnv allow
-pip install -U pip
-pip install -e .[docs,tests]
+```
+
+Tests are implemented with [pytest](https://docs.pytest.org/).
+Run them as follows:
+
+```bash
 pytest -vv
+```
+
+Documentation is built with [Sphinx](https://www.sphinx-doc.org/).
+Rebuild the documentation as follows:
+
+```bash
 cd docs
 ./compile_html.sh
 ./compile_pdf.sh
 ```
 
 ## Documentation Live Preview
 
-The documentation is created using [Sphinx](https://www.sphinx-doc.org/).
-
 Edit the documentation Markdown files with a live preview
 by running the following command *in the root* of the repository:
 

pyproject.toml

@@ -63,15 +63,16 @@ docs = [
     "sphinx-tippy",
     "sphinxcontrib-bibtex",
     "sphinxcontrib-svg2pdfconverter",
-    "stepup>=3.1.2",
-    "stepup-reprep>=3.1.5",
+    "stepup>=4.0.0rc7; python_version>='3.11'",
+    "stepup-reprep>=4.0.0rc3; python_version>='3.11'",
 ]
 tests = [
     "pytest",
     "pytest-regtest",
     "pytest-xdist",
     "numdifftools",
 ]
+dev = ["pre-commit"]
 
 [project.urls]
 Documentation = "https://molmod.github.io/stacie/"

src/stacie/utils.py

@@ -159,7 +159,7 @@ def split(sequences: NDArray[float], nsplit: int) -> NDArray:
     """
     sequences = np.asarray(sequences)
     if sequences.ndim == 1:
-        sequences.shape = (1, -1)
+        sequences = sequences.reshape(1, -1)
     if not isinstance(nsplit, int) or nsplit <= 0 or nsplit > sequences.shape[-1] / 2:
         raise ValueError("nsplit must be a positive integer smaller than half the sequence length.")
     length = sequences.shape[1] // nsplit