drbx 2.0.0.dev0__py3-none-any.whl

drbx-2.0.0.dev0.dist-info/METADATA

@@ -0,0 +1,380 @@
+Metadata-Version: 2.4
+Name: drbx
+Version: 2.0.0.dev0
+Summary: JAX-first edge and scrape-off-layer plasma code with differentiable, hardware-agnostic solver infrastructure.
+License-Expression: MIT
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: jax
+Requires-Dist: scipy
+Requires-Dist: matplotlib
+Requires-Dist: netCDF4
+Requires-Dist: rich
+Requires-Dist: pillow
+Requires-Dist: solvax>=0.8.6
+Requires-Dist: tomli; python_version < "3.11"
+Provides-Extra: dev
+Requires-Dist: pytest; extra == "dev"
+Requires-Dist: pytest-cov; extra == "dev"
+Requires-Dist: pytest-xdist; extra == "dev"
+Provides-Extra: docs
+Requires-Dist: mkdocs; extra == "docs"
+Requires-Dist: pymdown-extensions; extra == "docs"
+Dynamic: license-file
+
+# DRBX
+
+[![Tests](https://github.com/uwplasma/drbx/actions/workflows/test.yml/badge.svg)](https://github.com/uwplasma/drbx/actions/workflows/test.yml)
+[![Docs](https://github.com/uwplasma/drbx/actions/workflows/docs.yml/badge.svg)](https://github.com/uwplasma/drbx/actions/workflows/docs.yml)
+[![Coverage](https://github.com/uwplasma/drbx/actions/workflows/coverage.yml/badge.svg)](https://github.com/uwplasma/drbx/actions/workflows/coverage.yml)
+[![PyPI](https://img.shields.io/pypi/v/drbx.svg)](https://pypi.org/project/drbx/)
+[![Python](https://img.shields.io/pypi/pyversions/drbx.svg)](https://pypi.org/project/drbx/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![JAX](https://img.shields.io/badge/JAX-enabled-0a9396.svg)](https://jax.readthedocs.io/)
+[![Read the Docs](https://readthedocs.org/projects/drbx/badge/?version=latest)](https://drbx.readthedocs.io/)
+
+**DRBX is a JAX-based, end-to-end differentiable drift-reduced Braginskii
+(DRB) code for edge and scrape-off-layer (SOL) plasma turbulence** — on both
+closed and open field lines, in axisymmetric (tokamak) and non-axisymmetric
+(stellarator) geometry via the flux-coordinate-independent (FCI) approach.
+
+Because the whole model is written in JAX, every simulation is `jit`-compiled,
+runs on CPU or GPU unchanged, and is differentiable: you can take gradients of
+any output (a saturated fluctuation energy, a transport level) with respect to
+any input (a density gradient, an adiabaticity, a diffusivity) through the
+solver. To our knowledge no other published DRB SOL turbulence code is
+differentiable, and none combines differentiability with FCI stellarator
+geometry.
+
+Documentation: [drbx.readthedocs.io](https://drbx.readthedocs.io/en/latest/).
+
+## Stellarator turbulence in three dimensions
+
+Four-field drift-reduced turbulence on a rotating-ellipse stellarator — a torus
+whose elliptical cross-section rotates with the toroidal angle. The cutaway
+shows the density fluctuations on a flux surface and through the interior;
+every frame is a `jit`-compiled, differentiable JAX step:
+
+![Stellarator turbulence in 3D](docs/media/stellarator_3d_turbulence.gif)
+
+*Reproduce with [`examples/stellarator/stellarator_3d_render.py`](examples/stellarator/stellarator_3d_render.py).*
+
+The same geometry supports closed and open field lines: core field lines (blue)
+stay on flux surfaces, while beyond a toroidal limiter the scrape-off-layer
+field lines (red) end on the limiter plate, where a Bohm sheath drains the
+plasma:
+
+![Closed and open field lines in 3D](docs/media/stellarator_3d_field_lines.png)
+
+*Reproduce with [`examples/stellarator/stellarator_3d_render.py`](examples/stellarator/stellarator_3d_render.py).*
+
+## Install
+
+```bash
+pip install drbx          # from PyPI
+# or, from source:
+git clone https://github.com/uwplasma/drbx && cd drbx && pip install -e .
+```
+
+Runtime dependencies are `jax`, `scipy`, `matplotlib`, `netCDF4`, `rich`,
+`pillow`, and [`solvax`](https://github.com/uwplasma/SOLVAX). Python 3.10-3.12.
+
+## Quick start
+
+Run a simulation from a TOML deck, or inspect one without running it:
+
+```bash
+drbx inspect examples/inputs/restartable_diffusion.toml   # resolve and print the plan
+drbx run     examples/inputs/restartable_diffusion.toml   # run and write artifacts
+```
+
+From Python, a differentiable turbulence run is a few lines:
+
+```python
+import jax.numpy as jnp
+import numpy as np
+from drbx.native.hasegawa_wakatani import HasegawaWakataniParameters, hw_grid, hw_run
+
+grid = hw_grid(64, 2 * jnp.pi * 8)
+params = HasegawaWakataniParameters(adiabaticity=1.0, gradient=1.0)
+rng = np.random.default_rng(0)
+zeta0 = jnp.fft.fft2(jnp.asarray(1e-2 * rng.standard_normal((64, 64))))
+n0 = jnp.fft.fft2(jnp.asarray(1e-2 * rng.standard_normal((64, 64))))
+zeta, n = hw_run(zeta0, n0, grid, params, dt=5e-3, steps=500)  # jit-compiled, differentiable
+```
+
+Every example below is a flat script: parameters at the top, run, plot.
+
+## Highlights
+
+**Turbulence on closed and open field lines.** The same multi-mode seed on the
+rotating-ellipse stellarator, with all field lines closed (top) and with a
+limiter opening the outer flux surfaces into a sheath-drained scrape-off layer
+(bottom). Four toroidal cross-sections; the mode pattern differs plane by plane
+because the flux surfaces rotate:
+
+![Stellarator turbulence, closed](docs/media/stellarator_turbulence_closed.gif)
+
+![Stellarator SOL turbulence, open](docs/media/stellarator_turbulence_open.gif)
+
+*Reproduce with [`examples/stellarator/stellarator_turbulence.py`](examples/stellarator/stellarator_turbulence.py).*
+
+**Island divertor.** A sheared rotational transform with resonant perturbations
+forms island chains and a stochastic edge. The open scrape-off layer emerges
+from the field itself: multi-transit field-line tracing marks the finite
+connection-length region, and the turbulence drains through it:
+
+![Island divertor](docs/media/island_divertor.png)
+
+*Reproduce with [`examples/stellarator/island_divertor.py`](examples/stellarator/island_divertor.py).*
+
+**Imported fields: real coils and VMEC equilibria.** The same closed/open
+field-line machinery runs on imported fields: the vacuum Biot-Savart field of
+the Landreman-Paul quasi-axisymmetric coil set (via ESSOS) shows nested closed
+surfaces inside a chaotic open edge, and a VMEC equilibrium (via vmec_jax)
+provides closed surfaces whose traced rotational transform matches the
+equilibrium's `iotaf` profile to ~1e-6 — with the open scrape-off-layer lines
+traced to their exit from the confined region:
+
+| Coil field (closed core, open edge) | VMEC equilibrium + coil-field SOL |
+|---|---|
+| ![Landreman-Paul coils Poincare](docs/media/closed_open_vacuum_poincare.png) | ![VMEC closed and open field lines](docs/media/vmec_jax_closed_open_field_lines.png) |
+
+*Reproduce with [`examples/geometry-3D/essos-field-lines/closed_open_vacuum_poincare.py`](examples/geometry-3D/essos-field-lines/closed_open_vacuum_poincare.py) (left) and [`examples/geometry-3D/vmec-jax/closed_open_field_lines.py`](examples/geometry-3D/vmec-jax/closed_open_field_lines.py) (right).*
+
+**Neutrals and detachment.** The open SOL flux tube reaches the two-point Bohm
+steady state; the hermes-3 neutral model (packaged AMJUEL atomic rates, target
+recycling) builds the neutral cushion; and with an evolved temperature the SOL
+detaches — the target cools through 1 eV and the target ion flux rolls over
+(the SD1D benchmark):
+
+| Open SOL flux tube | Recycling SOL with neutrals |
+|---|---|
+| ![Open SOL flux tube](docs/media/open_sol_flux_tube.png) | ![Recycling SOL](docs/media/recycling_sol.png) |
+
+*Reproduce with [`examples/sol/open_sol_flux_tube.py`](examples/sol/open_sol_flux_tube.py) (left) and [`examples/sol/recycling_sol.py`](examples/sol/recycling_sol.py) (right).*
+
+![Detachment rollover](docs/media/b6_detachment.png)
+
+*Reproduce with [`examples/benchmarks/b6_detachment_rollover.py`](examples/benchmarks/b6_detachment_rollover.py).*
+
+## Gradient-based optimization through the physics
+
+Because the solver is differentiable end to end, control and design problems
+become gradient computations. Example: **detachment control** — find the
+upstream density that places the divertor target exactly at the 1 eV
+detachment threshold (Dudson et al., PPCF 61, 065008; Body et al., NME 41,
+101819). The sensitivity `dTe_target/dn_up` is computed by forward-mode
+autodiff **through the entire 20,000-step stiff SOL solve**, and a trust-region
+Newton iteration walks down the detachment cliff to the threshold:
+
+![Detachment control](docs/media/detachment_control.png)
+
+*Reproduce with [`examples/autodiff/detachment_control.py`](examples/autodiff/detachment_control.py).*
+
+A second example: **turbulence optimization** — find the adiabaticity (the
+parallel electron conductivity) at which saturated drift-wave transport drops
+to a quarter of its hydrodynamic-regime level, the classic
+hydrodynamic-to-adiabatic transition of Hasegawa-Wakatani turbulence (Camargo,
+Biskamp & Scott, *Phys. Plasmas* 2, 48 (1995)). A damped Newton iteration on
+the adiabaticity, with forward-mode gradients through the saturated turbulence,
+converges in 7 iterations; an independent long run verifies a 3.96x flux
+reduction against the 4x target. Left column: the initial hydrodynamic state;
+right column: the optimized adiabatic state:
+
+![Hasegawa-Wakatani optimization](docs/media/hasegawa_wakatani_optimization.png)
+
+*Reproduce with [`examples/tokamak/hasegawa_wakatani_optimization.py`](examples/tokamak/hasegawa_wakatani_optimization.py).*
+
+The same machinery recovers a transport-drive parameter by gradient descent
+through nonlinear drift-wave turbulence
+([inverse design](examples/tokamak/drift_wave_inverse_design.py)), and the
+[differentiation-methods example](examples/autodiff/differentiation_methods.py)
+measures which method is cheapest (forward mode for a few parameters, ~2x a
+forward run; reverse mode for parameter fields; checkpointing when memory-bound).
+
+## Performance and parallel execution
+
+Single-CPU turbulence throughput is about 2 million cell-updates per second in
+float64, and one gradient through a 200-step rollout costs 2-3x a forward run.
+The full four-field FCI step — four RHS evaluations, each with a GMRES
+potential inversion — compiles as a single `jit` program with no host
+synchronization inside, which roughly halves the stellarator-turbulence
+step time on one CPU
+([details](docs/performance_and_differentiability.md)):
+
+| Turbulence performance | Cost of each differentiation method |
+|---|---|
+| ![Performance](docs/media/performance.png) | ![Differentiation methods](docs/media/differentiation_methods.png) |
+
+*Reproduce with [`examples/benchmarks/performance_benchmark.py`](examples/benchmarks/performance_benchmark.py) (left) and [`examples/autodiff/differentiation_methods.py`](examples/autodiff/differentiation_methods.py) (right).*
+
+The FCI stack runs across devices with `shard_map`: the sharded step is
+bit-exact against single-device execution, and on a 36-core host with one core
+per shard a 1.05M-cell step reaches a 7.4x speedup at 16 shards; the same step
+on one NVIDIA A4000 GPU runs ~96x faster than a single CPU shard
+([demo](examples/benchmarks/fci_sharded_strong_scaling.py)):
+
+![Strong scaling](docs/media/strong_scaling.png)
+
+*Reproduce with [`examples/benchmarks/fci_sharded_strong_scaling.py`](examples/benchmarks/fci_sharded_strong_scaling.py).*
+
+## Hasegawa-Wakatani benchmark
+
+The standard two-field drift-wave turbulence benchmark: grown from noise
+through the linear instability (growth rate verified against the analytic
+dispersion relation to ~1e-14) into nonlinear E×B transport. Detailed
+verification figures — dispersion scans, MMS convergence orders, and
+gradient-vs-finite-difference checks — are in the
+[documentation](https://drbx.readthedocs.io/en/latest/):
+
+![Drift-wave turbulence](docs/media/drift_wave_turbulence.gif)
+
+*Reproduce with [`examples/tokamak/drift_wave_turbulence.py`](examples/tokamak/drift_wave_turbulence.py).*
+
+## Reproducing the figures and movies
+
+Every figure and movie above states the script that generates it. Each is a
+flat pedagogical file in [`examples/`](examples/): all parameters at the top
+with comments, explicit model/geometry/boundary-condition setup through the
+public API, progress printed while it runs, plot written at the end. The
+embedded copies live compressed in `docs/media/`; the scripts regenerate
+full-quality versions under `output/`.
+
+## What it does
+
+| Capability | What ships |
+|---|---|
+| **Turbulence models** | Hasegawa-Wakatani drift-wave (pseudo-spectral, differentiable); FCI 2-field, 4-field interchange (density/vorticity/parallel flows), and electromagnetic drift-reduced stacks with curvature and vorticity/potential closures |
+| **Geometry** | Rotating-ellipse stellarator (closed core + optional limiter SOL), island-divertor field (emergent stochastic SOL), shifted-torus helical flux tube, open slab SOL, imported ESSOS coil / VMEC / hybrid equilibria — metrics by autodiff of analytic embeddings |
+| **Field-line topology** | Closed and open field lines; FCI traced field-line maps; multi-transit connection-length tracing; Bohm sheath + target recycling closure on open endpoints |
+| **Neutrals (hermes-3 model)** | Packaged AMJUEL ionization/recombination + charge-exchange rates (no external database); Galilean-invariant plasma-neutral coupling; recycling SOL and a self-consistent detaching SOL (implicit Spitzer conduction, self-limiting radiation, SD1D rollover) |
+| **Linear solver** | `drbx.linear` linearizes any model about an equilibrium; drift-wave, shear-Alfven, and interchange dispersion reproduced to machine precision |
+| **Differentiability** | `jit`/`grad`/`vmap` through every model — sensitivity, uncertainty propagation, inverse design, detachment control; forward/reverse/checkpointed methods measured and gated to agree |
+| **Parallelism** | Multi-device `shard_map` FCI stepping (bit-exact vs single device) with halo exchange; CPU strong scaling demonstrated, GPU-ready |
+| **Solvers** | Structured solves via [`solvax`](https://github.com/uwplasma/SOLVAX) (spectral Fourier-Helmholtz elliptic, tridiagonal, Krylov, preconditioners) |
+| **Runtime** | TOML-deck CLI (`drbx inspect` / `run`) and a small Python API; restartable runs; portable JSON/NPZ artifacts |
+
+## Validation
+
+`drbx` is validated against a ladder of literature-anchored benchmarks.
+Each rung has a test (or a documented gate) and an example that regenerates
+its figure.
+
+Verified today (each with a passing test):
+
+| Case | Anchor | What is checked |
+|------|--------|-----------------|
+| Method of manufactured solutions | Riva et al., *Phys. Plasmas* 21, 062301 (2014); Dudson et al. 23, 062303 (2016) | operator / 1D-fluid / FCI convergence order → 2 |
+| Resistive drift-wave dispersion | Dudson et al., *Comput. Phys. Commun.* 180, 1467 (2009) | growth rate and frequency vs analytic dispersion |
+| Shear-Alfvén wave dispersion | Stegmeir et al., *Phys. Plasmas* 26, 052517 (2019) | phase velocity vs analytic (with electron inertia) |
+| Interchange / Rayleigh-Taylor | curvature-driven flute dispersion | growth rate vs `√(gκ)·k_y/k` analytic |
+| FCI on non-axisymmetric geometry | Shanahan et al., *PPCF* 61, 025007 (2019, BSTING) | parallel-operator MMS; differentiable rollout (grad vs FD 6e-11) |
+| Rotating-ellipse (`l = 2`) FCI | Stegmeir et al., *Comput. Phys. Commun.* 198, 139 (2016, GRILLIX) | direct & traced-field-line parallel gradient converge at order 2 on a genuinely non-axisymmetric metric; shape-differentiable; a seeded four-field filament generates interchange vorticity on the rotating surfaces |
+| Island-divertor field (B8) | Shanahan et al., *J. Plasma Phys.* 90 (2024, BSTING); GBS island-divertor studies | sheared-iota island chains + stochastic edge; closed core and finite-connection-length open SOL emerge from multi-transit tracing; turbulence drains through the emergent divertor masks |
+| Open-field-line SOL flux tube | two-point / Bohm-sheath SOL theory (Stangeby, *The Plasma Boundary of Magnetic Fusion Devices*, 2000) | parallel flow reaches Mach 1 at the targets; target density = half upstream; exact Bohm particle balance and sheath-recycling accounting |
+| Neutrals and recycling (hermes-3 model) | hermes-3: Dudson et al., *Comput. Phys. Commun.* 296, 108991 (2024); AMJUEL atomic rates | physically-correct ionization/recombination/CX rates; exact plasma↔neutral particle & momentum conservation; neutrals conserve on the 3D closed rotating ellipse and recycle on the open slab |
+| SD1D detachment rollover (B6) | SD1D: Dudson et al., *Plasma Phys. Control. Fusion* 61, 065008 (2019) | self-consistent SOL (evolved temperature, implicit Spitzer conduction, self-limiting radiation): the target cools through 1 eV into the recombining regime and the target ion flux rolls over as upstream density rises; differentiable |
+| Differentiable inverse design | — | gradient descent through turbulence recovers a drive parameter |
+
+Planned rungs (seeded-blob inertial scaling and others) are
+tracked in the project planning notes; benchmark reports live under
+[docs/](docs/linear_dispersion_benchmark.md) and
+[docs/validation_gallery.md](docs/validation_gallery.md).
+
+## Examples
+
+Flagship simulations, by geometry:
+
+| | Turbulence flagship | Geometry |
+|---|---|---|
+| **Tokamak** | [drift-wave turbulence](examples/tokamak/drift_wave_turbulence.py) (Hasegawa-Wakatani; linear phase B2-verified, differentiable) + [inverse design](examples/tokamak/drift_wave_inverse_design.py) | periodic flux tube |
+| **Stellarator** | [turbulence on closed + open field lines](examples/stellarator/stellarator_turbulence.py) (four-field, limiter SOL, movies) + [3D renders](examples/stellarator/stellarator_3d_render.py) (cutaway turbulence movie, field-line topology) + [island divertor](examples/stellarator/island_divertor.py) (B8: Poincare, connection lengths, emergent open SOL) + [rotating-ellipse FCI](examples/stellarator/rotating_ellipse_fci.py) (parallel-operator convergence) + [seeded filament](examples/stellarator/rotating_ellipse_filament.py) + [differentiable FCI drift-reduced model](examples/stellarator/fci_differentiable.py) | rotating ellipse (closed core + limiter SOL) + shifted-torus helical + imported [ESSOS/VMEC](examples/geometry-3D/) |
+| **Coils (vacuum)** | [Landreman-Paul closed + open field lines](examples/geometry-3D/essos-field-lines/closed_open_vacuum_poincare.py) (ESSOS Biot-Savart, Poincare classification) | imported coil field |
+| **VMEC equilibria** | [closed field lines from a wout file](examples/geometry-3D/vmec-jax/closed_field_lines.py) (vmec_jax import; traced rotational transform matches the equilibrium `iotaf` profile to ~1e-6) + [closed + open field lines](examples/geometry-3D/vmec-jax/closed_open_field_lines.py) (coil field with the VMEC last closed flux surface overlaid) | imported VMEC equilibrium (Landreman-Paul precise QA) |
+| **SOL (open)** | [open SOL flux tube](examples/sol/open_sol_flux_tube.py) (parallel transport to Bohm-sheath targets; two-point steady state) + [recycling SOL](examples/sol/recycling_sol.py) (neutrals, ionization/recombination, detachment onset) | open slab flux tube |
+
+Open-field-line SOL:
+[open slab flux tube](examples/sol/open_sol_flux_tube.py) — parallel
+transport to Bohm-sheath-bounded targets, relaxing to the classic two-point
+steady state (Mach 1 at the targets, target density half the upstream density),
+with the FCI sheath/recycling closure on the target plates.
+
+Benchmarks, differentiable, and geometry examples:
+
+- Linear dispersion (B2/B3):
+  [examples/benchmarks/linear_dispersion.py](examples/benchmarks/linear_dispersion.py)
+  reproduces the drift-wave and shear-Alfvén dispersion relations from the
+  linear solver.
+- Autodiff: [gradient-based detachment control](examples/autodiff/detachment_control.py)
+  (forward-mode sensitivity through the stiff SOL solve, trust-region Newton onto
+  the 1 eV threshold), [inverse design through turbulence](examples/tokamak/drift_wave_inverse_design.py)
+  (recover a parameter by gradient descent through a nonlinear drift-wave run),
+  [choosing the most efficient differentiation method](examples/autodiff/differentiation_methods.py)
+  (forward vs reverse vs checkpointed reverse — same gradient, different cost),
+  plus [sensitivity](examples/autodiff_diffusion_sensitivity.py),
+  [uncertainty](examples/autodiff_diffusion_uncertainty.py), and reduced
+  [inverse design](examples/autodiff_diffusion_inverse_design.py).
+- Stellarator FCI and imported geometry:
+  [examples/geometry-3D/](examples/geometry-3D/).
+- Start with [examples/model_selection_guide.py](examples/model_selection_guide.py)
+  to choose a model family, dimension, and boundary conditions.
+
+The examples are self-contained — no external plasma code is needed to run
+them. Large figures and movies are hosted in GitHub releases so the checkout
+stays small.
+
+## Geometry and parallelization
+
+The FCI operator and domain-decomposition stack (`FciGeometry3D`,
+`fci_operators`, halo exchange) was contributed by **Aiken Xie** in
+[PR #3](https://github.com/uwplasma/drbx/pull/3) and is incorporated here.
+Built on it, the drift-reduced two-field step runs across multiple devices with
+`shard_map`: the domain is decomposed into halo-exchanged shards and the sharded
+RK4 step is **bit-exact** against the single-device step (checked to ~1e-16 for
+single-device and forced-four-device runs in
+[`tests/test_fci_sharded_2field.py`](tests/test_fci_sharded_2field.py)). On a
+36-core Linux host with one core bound per shard, a 1.05M-cell step reaches a
+**7.4x speedup at 16 shards**, and one NVIDIA A4000 GPU runs the same step
+~96x faster than a single CPU shard
+([strong-scaling script](examples/benchmarks/fci_sharded_strong_scaling.py),
+[docs](docs/performance_and_differentiability.md)).
+
+## Documentation
+
+- Physics and numerics: [physics_models.md](docs/physics_models.md),
+  [equation_to_code_map.md](docs/equation_to_code_map.md),
+  [code_structure.md](docs/code_structure.md).
+- Performance and differentiability:
+  [performance_and_differentiability.md](docs/performance_and_differentiability.md),
+  [profiling_runtime.md](docs/profiling_runtime.md).
+- Validation: [validation_gallery.md](docs/validation_gallery.md).
+- Testing policy: [testing_strategy.md](docs/testing_strategy.md).
+
+## Testing
+
+```bash
+pytest -q -m "not slow"                                   # full fast suite
+pytest -q -m "not slow" --cov=drbx --cov-branch        # with coverage
+```
+
+CI runs the full fast suite on Python 3.10–3.12.
+
+## Releases
+
+The current development series is described in
+[docs/release_notes_2_0_0_dev0.md](docs/release_notes_2_0_0_dev0.md).
+
+## Citing
+
+If you use DRBX in published work, please cite this repository
+(https://github.com/uwplasma/DRBX).
+
+## License
+
+MIT — see [LICENSE](LICENSE).

drbx-2.0.0.dev0.dist-info/RECORD

@@ -0,0 +1,106 @@
+drbx/__init__.py,sha256=yF8f2vTPl6VVZl7YgA98_wo71t7mhCv8pGWpAB7cVN8,585
+drbx/__main__.py,sha256=PSQ4rpL0dG6f-qH4N7H-gD9igQkdHzH4yVZDcW8lfZo,80
+drbx/cli.py,sha256=eyIPbYZ7kSLc1mCVErMi398nmUN0hBzp2vGD80wcbxA,21935
+drbx/config/__init__.py,sha256=LsgOLLlw0gU8xoXZrV3maheUn-DvBJXMOjPqVVJLHdQ,645
+drbx/config/boutinp.py,sha256=khEl_5yzR7F82T-xSGBbeD2iX01NUkD-hX4AQXC3UZI,20167
+drbx/config/model.py,sha256=Xg-yOrkj5Exw_0mQ8Lon_H0IjTP4iPDikGqhulHf1Tg,967
+drbx/config/normalization.py,sha256=dXjAFflh-Knpe1KlpMIR3ffSz0kfLUi6ei4xG1Gh4sQ,2074
+drbx/data/atomic_rates/__init__.py,sha256=Zu7gdLI2fry6Peo-_FHadbN_2OwHyVJkUSMInXUIGIc,74
+drbx/data/atomic_rates/iz_AMJUEL_H.x_2.1.5.json,sha256=phsv7xT5nsBi30B-Envr2F1J5SKsgKt_yb8VUOLh8eM,6027
+drbx/data/atomic_rates/iz_AMJUEL_H.x_2.3.9a.json,sha256=Gq85BB4_qssLR1WH0Yv6xuApwOw2yzToGbxcYMGnWcQ,6220
+drbx/data/atomic_rates/rec_AMJUEL_H.x_2.1.8.json,sha256=R5nYLJ7roKfE8C0QXNH8KN_5_DiAmlVQhJyIVIKHAII,6297
+drbx/data/atomic_rates/rec_AMJUEL_H.x_2.3.13a.json,sha256=y0N0wlfHjqdLoGvc3GcvN5LchpjQHG5q3J3tHZ0rux8,6381
+drbx/geometry/__init__.py,sha256=BcUiJeVit2B0vvoKs8zDhd6VssttCkEVIRJMYkkTvhA,6350
+drbx/geometry/embedding.py,sha256=0aFTJhvuLl9XcyLikPlcwo1li78zGPFrUazjOdlHhns,2001
+drbx/geometry/essos_import.py,sha256=nx0LEmr9OL9nuV37YPN9jaoafvYq1m-eO1falROKJ4U,56075
+drbx/geometry/fci_geometry.py,sha256=B_T_I_Q_4OnO7inedqnPYOKf2Zf6i3n6pqkm-DV5KEc,176396
+drbx/geometry/fci_maps.py,sha256=i0QPPzJjPk0gjyoVjo8sy-gt-8lzCl5wU4vsJAZ951Y,3193
+drbx/geometry/island_divertor.py,sha256=CzptKOwdwd91Tr1ECh7kFhxV98E16XPYGEG3ZMVY8WQ,12028
+drbx/geometry/metric_tensor.py,sha256=8o3hKfyyy35Vqk2q2NLy21nFYEGiyTEdGXdbDaYGxEE,2856
+drbx/geometry/open_slab.py,sha256=tSHAbq1ZnLNiRf1Y6rt4lWDxQfNilBPeQulziDUN_ck,5595
+drbx/geometry/rotating_ellipse.py,sha256=hA3JxdZD1WzH1oFfTaPWtwqDj5KclHMn0cNYHSe8YNQ,10481
+drbx/geometry/shifted_torus.py,sha256=MVoOGTDIGrR3n7lXrtLNKh7NTG3Q5pjzTLnnhpdnmqw,9234
+drbx/geometry/stellarator.py,sha256=ZtJ80ZM1o72nQTOcrjUBEQgF1WfcWgPyy4oON1Jb-tA,10683
+drbx/geometry/vmec_extender_import.py,sha256=L7n_yn3DnC1PycUzTyC6pjc29XNnMErwKYOfyByYNGU,18650
+drbx/geometry/vmec_jax_import.py,sha256=x5Nqt5yoWzwoeTkeIuIjHNSEDgCZCJp1X5jdwYQGXnI,12825
+drbx/linear/__init__.py,sha256=bfv1lXTg84ESa3GbThH07hqVamiEoCs1tbsuxdGOhxY,1016
+drbx/linear/dispersion.py,sha256=ue21B2iEFmslv2Zc85h4DHHHtzXLZRFMKsvYSutu-wI,5263
+drbx/linear/eigen.py,sha256=zQL8PNkBNFJDN6ZMzxJvQ9w7uVxyRx3tu2fWhbp3hsI,3004
+drbx/native/__init__.py,sha256=Ak5LhplfQHHGxw5rqQkRJ8Cdw8s2zF6b94SJ9CE1020,6758
+drbx/native/array_backend.py,sha256=Y5c8_pHk0hBT4Kwg286xA_j7o3fIdRyRFtyXZ4VF2n4,2138
+drbx/native/deck_runner.py,sha256=5r_OPIzRO3eOm2WxbOxE5A_fMKLTQvpO-GvIx_wQf14,30585
+drbx/native/electromagnetic.py,sha256=CaMkkc4hMMfGBeIjwwGmsakHTQjnrGy_OY7lHc-QPpQ,10944
+drbx/native/expression.py,sha256=eRGG65ovuPGIaExHaltypniwO6ZbHQ-QZh6hx2L5U7Q,6523
+drbx/native/fci.py,sha256=kwtMFQBxtGeTROKZR9zPVIRXER4uywLjYgJDnOkdSj4,11934
+drbx/native/fci_2_field_rhs.py,sha256=R4I-1pNy96lCyoSoAFe5A4WBbLdEMfR5XfbYnbkbjhQ,7013
+drbx/native/fci_4_field_rhs.py,sha256=YqPyFfJANFJhZLo0RWy-xDW6OHnXcCpUUYju8YukUsA,52050
+drbx/native/fci_boundaries.py,sha256=PFAoT4pljgNm9umTUVDdGxu6W3fVPD-5ebo7CAAgB_w,102407
+drbx/native/fci_differentiable_case.py,sha256=Z6kzR3sPWWpOxwN58F_8_5J5LJV6RZUKDS5rbWIOsmE,11134
+drbx/native/fci_drb_EB_rhs.py,sha256=S5Zh_ruXSQIBE3RlpUDeV_msJ3HU9IQPAE58DdbXqxw,46778
+drbx/native/fci_drb_rhs.py,sha256=uMhqFh13UrHY-MfnbAWQKtajxkyuNu_QCE4C3jKrKlA,6869
+drbx/native/fci_halo.py,sha256=GlhMSV7IQ_Rq0t-xo7zwEO_C6YHQ0lJzDnn8OHS0gfs,62615
+drbx/native/fci_helpers.py,sha256=07GmjguxhD6OP-EiZLb9bpmuUj8arKQ8IWUqIDkgfYQ,11740
+drbx/native/fci_model.py,sha256=8vMUEyDjffwReVNwhX3zwmOFsYISWfQAssOWYacnmRI,10193
+drbx/native/fci_neutral.py,sha256=B3xvcvwkqVdda9A1baEdYcWQ93vuG2hwhfHVwpOdsec,6422
+drbx/native/fci_operators.py,sha256=wRIGYpNsYnilkzvpd8rpU2QXJey0KFX6Zw5EAO7QlpE,167675
+drbx/native/fci_sharding.py,sha256=lBU0OuOAqAAZ4N7f_bfBXCRAbteQtGrwI99NnXKCfwY,21936
+drbx/native/fci_sheath_recycling.py,sha256=KvuIrqk1A0tWlP_5YWjCzy-fkd-L0lA-M6Xwk5Endqc,8254
+drbx/native/fci_time_integrator.py,sha256=FlAOrCjqYGN92ShB5bGp3hSBlTn_H1vGVU50ougFOY8,3399
+drbx/native/fci_vorticity.py,sha256=KnAp8YDbpF1hfHI2UsBjXDmjI8VrtIepXPZ3xjVWRsE,6805
+drbx/native/fluid_1d.py,sha256=K7ZO978m9VTYBDnilg-zFbaPyorJHTPRG16WVecnOYk,11684
+drbx/native/hasegawa_wakatani.py,sha256=6dbgS6xlhRkuZNhZPf6Nt_Q9eaRI6kkoFiJv19c1m_8,7477
+drbx/native/limiters.py,sha256=1bHUpopBSb9VDEDxys5vw-6r9FSxaQ5EwbPYiXBSZxU,2004
+drbx/native/mesh.py,sha256=P5Nct0yZ9WyekdZqNBlaSsxecOg51snzhgZByFRbME0,8153
+drbx/native/metrics.py,sha256=KTjUep5n93UYB0Z3scS28hWtFzfvv3c_VelHVhD57C0,8633
+drbx/native/sol_flux_tube.py,sha256=gNOdmR-SzJ30FU0g1ekmJcKO27k7-7X0GJGNt-N3U7c,5250
+drbx/native/stellarator_turbulence.py,sha256=wtdMLDFxilOF1VycxcIAnbZtUYMs0Iey6MgqacwsdX4,14312
+drbx/native/transport.py,sha256=mPrjJUpEz16gp9EAAN5eN_KbhRmQu3KQEzbjgWEZl4o,5390
+drbx/native/units.py,sha256=rk07gFL8nmNRinKfx75V8NznpfMp8PdUp4CQ_UFZvkQ,1133
+drbx/native/vorticity.py,sha256=CHxlZ24sIttFbFV1H8YO-vlzdwUnALuINYNY2li2ZSs,10036
+drbx/native/neutrals/__init__.py,sha256=m5BtTQEb5Ng9pc4qblbRS6a26hVIdGF6qTG0siuH2ws,1645
+drbx/native/neutrals/atomic_rates.py,sha256=9fTJjeQc2XqXTS5yjdCNHfizROxo_9Psm4YyWQ5LCUY,5405
+drbx/native/neutrals/detachment_sol_model.py,sha256=QtjR01pfK07FFRkbucKKYu9G_IEVP-RGwEnrRMN-XZc,10743
+drbx/native/neutrals/reactions.py,sha256=BFlOwWvXocQXpRremXuJYhII4eznw_2zoQsLwTOZUP0,6726
+drbx/native/neutrals/recycling_sol_model.py,sha256=tCQMeQrGHagoCR2pMyzELqCI8ARmSmgJNfjnaTHTYZc,9461
+drbx/runtime/__init__.py,sha256=FPZWM5MkvMRw6K1TCe_fTJbLoyoeh3FIT46hi7McnTM,1652
+drbx/runtime/artifacts.py,sha256=Bih-kovt2kj9Zg1QKFg-y1GHnuEEzrwV8q84inM1sBE,5211
+drbx/runtime/memory.py,sha256=5OjNXjrreTpUMDU6w_layARJxMY3ErK5MtPLS3K0gsk,4047
+drbx/runtime/output.py,sha256=jjbaw6hrqsxz0E49FBS6DHaeI1LhO-0pLqzeZ9UK_LI,15618
+drbx/runtime/paths.py,sha256=kWMwlkiPGAJYgR3ejD82-DK1DJaVHeliuKh9y8LaOdM,199
+drbx/runtime/performance.py,sha256=-24iA3tmypeKqWIXEIOr6H2HJv4Scg3hL2EpcIoDfM8,6125
+drbx/runtime/run_config.py,sha256=TrAOjiH0iM_5SnnJd0J-uhs-SH0cSP-oB630NpQ8Frw,6961
+drbx/runtime/scheduler.py,sha256=c0ZXB_ClfCD-bA94M4n8cC8MW_OXPGruQlBKZ9G9lyw,3178
+drbx/runtime/state.py,sha256=ZLgiSpJPX2rbTNzX950ZxXdl44ZibmEKhPo1lFyf8mw,1301
+drbx/validation/__init__.py,sha256=o8tXQyxmS__VmcjWKgyUYy_UeabIR8XIvkjxZxNaLLY,17257
+drbx/validation/autodiff_diffusion.py,sha256=nG1n47gqpVEwZHIpcrP0CrpdnKcyUXeT8kI-RiALcD4,9843
+drbx/validation/autodiff_diffusion_uncertainty.py,sha256=2pEwtNW0fYnZcHSSkzzQVRbjVl-4D_geUcNbo_UVIyI,9509
+drbx/validation/diverted_tokamak_movie.py,sha256=49nespLA3gfrIRuD2v_08pOEDb1dEXkQp2MlsvzZwzY,18698
+drbx/validation/essos_fieldline_import_campaign.py,sha256=EJlQT1g5BQ50ubII3qIer8-WMJDJK7zWDAwGo-lY6OY,6568
+drbx/validation/essos_imported_artifact_audit.py,sha256=vD34qC4WCd-qP71MTYV8UIsewI0YKhfuH60Cn1pt-Hg,19697
+drbx/validation/essos_imported_drb_movie_campaign.py,sha256=toXNllrajcquS7lGcay_3Cn6kL7_kl5Hcvx9i8CHNkA,115629
+drbx/validation/essos_imported_fci_campaign.py,sha256=mMwkOxWmn4uNc3EB-AWPAi5i78JfucNedK5RPqcQhQ0,226263
+drbx/validation/essos_imported_pytree_campaign.py,sha256=kuXORF4FF20-2GMbvf261ii6iy7IrdF8kNKrQXg9Ct4,16761
+drbx/validation/essos_vmec_closed_field_campaign.py,sha256=I3CMlX0Y285J3ufuzr4QFHnMUndBeTBpzdV3mqbIfSw,13863
+drbx/validation/essos_vmec_closed_field_transient_campaign.py,sha256=2ZrHz8HHbSqBDBYh-YoIt6hPr6YlBOBHQBQi9ovtybo,26601
+drbx/validation/essos_vmec_fieldline_surface_campaign.py,sha256=EonRCs13mKPIhgtNLMBRnWCsopeDMXnsa3oGwlZJAKo,26876
+drbx/validation/fluid_1d_mms_convergence.py,sha256=o1zuZbfYil8O_lJws7v6LuXhqGf3woAPH8gJx0EbxUs,10712
+drbx/validation/geometry_lineouts.py,sha256=5Sb-zCRd1Lgb03_m8KKQLCqhK5nIC9Aufj8qpUCPgfA,5217
+drbx/validation/geometry_slices.py,sha256=qeJTjbIzW-WEZg3r0iu4MIHySvtBu0uNHZeZNYsYDV0,6526
+drbx/validation/publication_plotting.py,sha256=xPtfpGrPrh5QMDh6GdyxKLQX6Gcb5Kd8Ft9gBhbCPcY,2517
+drbx/validation/stellarator_drb_pytree_campaign.py,sha256=GIOohVtHUh9ZLCRj0b2LAu8u3pC7NMz0i-rHB_4jxMA,25963
+drbx/validation/stellarator_fci_geometry_campaign.py,sha256=V5eJxp0jQfy1mece7nKlXZosOYPGfSe8TWezEBu2uno,8321
+drbx/validation/stellarator_fci_operator_campaign.py,sha256=J_oFiMxwm4oJ7WOqlPplPGl_BWnmZZtfUCOfNeOdIZQ,13253
+drbx/validation/stellarator_fci_suite_campaign.py,sha256=99RW0G4fluEWpWyNQjoXi41bpb6eNvGfnqnyb5jHakQ,9967
+drbx/validation/stellarator_metric_mms_campaign.py,sha256=zJ8cT9zzaSPhjs9QCISc-gOdzCSMOF4Qmeo5ZXTPxq4,11698
+drbx/validation/stellarator_neutral_physics_campaign.py,sha256=JKmiapvsrsLuW0OsJPJWJ3oKVXiNjOMufHDaGKRGFTs,10868
+drbx/validation/stellarator_sheath_recycling_campaign.py,sha256=1yu6Teqfp7dcMW89Z17qhwMnJYzy9f0jRB1pVG6OOM0,13898
+drbx/validation/stellarator_sol_showcase.py,sha256=Na6MU7MKDenn2Oo4aLfe7UYxa3xsOeL8FDyp058ZIhw,25611
+drbx/validation/stellarator_vorticity_campaign.py,sha256=c6WXJJG73wpBATdQnfv3Zg7gGmvU-4RXrSJP_3FSTbI,12792
+drbx/validation/vmec_extender_edge_field_campaign.py,sha256=y1wPBDuD6VXknFkwJUbrruGLLANbVt7UCc2bEvJZAII,10240
+drbx/validation/vmec_extender_sol_smoke_campaign.py,sha256=fUsU3PIav3MAtDLjVbpXrbEqXucCZtDvPe0mXi5tt9o,14644
+drbx-2.0.0.dev0.dist-info/licenses/LICENSE,sha256=prQlXlwFtThdCuEwKetGzhuBMlGzqfjHH9SluGKQ7vo,1070
+drbx-2.0.0.dev0.dist-info/METADATA,sha256=28Xys1K-QR-7HXgftRH8M22grLcQZ45h6-Xek9mDNWM,22896
+drbx-2.0.0.dev0.dist-info/WHEEL,sha256=K260EYznzXsJYBQGqmI8VTxEdiZYNvDZwW9cBh9-_MA,91
+drbx-2.0.0.dev0.dist-info/entry_points.txt,sha256=uhQAFugyxaqll7itbIQ74nHGwAq0ltMX5ZD2NM5i97s,39
+drbx-2.0.0.dev0.dist-info/top_level.txt,sha256=1zyUewDUDsUDeV7B78r8WornklzAfb9To1GvHit_gY0,5
+drbx-2.0.0.dev0.dist-info/RECORD,,

drbx-2.0.0.dev0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (83.0.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

drbx-2.0.0.dev0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+drbx = drbx.cli:main

drbx-2.0.0.dev0.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Rogerio Jorge
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

drbx-2.0.0.dev0.dist-info/top_level.txt

@@ -0,0 +1 @@
+drbx