Copyright and credits

This page states three things: the copyright and licence of this project, the provenance of each physics module, and the work that is owed acknowledgement. The per-file attribution of record is the repository NOTICE; this page is its readable summary, and every statement made here should be traceable to NOTICE or ACKNOWLEDGEMENTS.en.md.

Copyright and licence

fylite is developed as part of the work of the Integrated Modelling Discussion Group at the Institute of Plasma Physics, Chinese Academy of Sciences (ASIPP). Copyright is held by ASIPP and the fylite contributors, and it is released under the Apache License 2.0: free use, modification and redistribution (including commercially) are permitted, provided the copyright notice and NOTICE are retained and modifications are stated. The full licence text is in the repository LICENSE.

★This page — the browser demos under app/ — and the compute kernels it loads are under the same licence: the kernels are built from this repository's own sources, and each artifact's sha256 and provenance are recorded in the repository documentation, so they can be checked byte for byte.

The parent project: FyTok

fylite is the lightweight edition of FyTok, its parent project (also Apache-2.0, Institute of Plasma Physics, Chinese Academy of Sciences): a full tokamak integrated-modelling framework, with plugins, workflow scheduling, heterogeneous execution and provenance, aimed at HPC and the cloud. Both implement the fyo semantic contract; fylite takes the light end — the key solvers written afresh so they run in the browser with nothing to install — and doubles as FyTok's minimal functional verification: the two are coded independently against the same contract, so comparing them is a cross-check rather than a code confirming itself. The core-transport channel grammar is taken from FyTok's fytrans (see the table below), which is also the bit-for-bit anchor of the transport step.

Ports and provenance

The modules below are white-box translations of upstream code, not independent reimplementations from the literature. The distinction is a licence obligation and is therefore stated in the body of this page rather than in a footnote.

Statement of changes (Apache-2.0 §4b): the ports are translations, not copies — module state becomes explicit arguments, LAPACK / UMFPACK become in-tree dense and sparse routines, upstream STOP becomes an error code, and the result is re-entrant. Several points deliberately differ from upstream behaviour, mostly where the upstream silently overrides its own inputs; each is marked at its use site in the source.

Two GACODE revisions coexist, deliberately: GEO / NEO / TGYRO derive from 5efddfdf1 and TGLF from 6357db306, matching the two libraries the ports were checked against. UMFPACK was not translated — the sparse LU here is written from scratch.

Data, oracles and cross-code anchors

• GACODE regression decks and recorded outputs (the GA standard case, TGYRO treg01, libgeo/libneo/libtglf recordings) — the gold fixtures of the ports (General Atomics, Apache-2.0).
• EAST discharge data and the operational EFIT workflow — the EAST team at ASIPP; per-channel diagnostic geometry reaches fylite through fydata. None of it ships with the demos.
• The ITER device description and the IMAS Data Dictionary — ITER Organization, reaching fylite through fyo and fydata; the ITER configuration built into the demos comes from there.
• Cross-code anchors: 0-D against METIS (CEA/IRFM), ITER scenarios against FUSE, electromagnetics and vertical stability against TokSys (General Atomics). None of those codes is redistributed here.

What was deliberately not taken

An acknowledgement that lists only what was taken is incomplete. The items below were deliberately not taken, for differing reasons, but they belong to the same record:

• EFIT lineage: no EFIT-family source, Green’s-table generator or recorded output is present in any form; the equilibrium and inverse solvers are clean-room.
• GRAY: the EC ray-tracing port was stopped; its licence does not permit direct translation, and no GRAY physics source was read for fylite.
• UMFPACK: not ported — the sparse LU is written from scratch.

How this code was built

The physics modules of fylite were constructed from the public literature and open-source code listed above, with AI assistance (Anthropic Claude). Correctness is not claimed from authorship: it is established module by module against gold fixtures from the upstream implementations, with the tolerance and its error budget stated. fylite additionally exposes its own interfaces to AI hosts (MCP / JSON-RPC); that is a product feature and a separate matter from how it was written.

Acknowledgements

This project rests on physics, code and data that others made public. The debts are acknowledged item by item:

• The Integrated Modelling Discussion Group (ASIPP): Youwen SUN, Qilong REN, Xiang JIAN, Tianyang XIA, Zhi YU, Xiaojuan LIU, Yemin HU, Xiaotao XIAO and others.
• General Atomics and the GACODE contributors (J. Candy, E. Belli, G. M. Staebler and colleagues) — upstream of the neoclassical, turbulence and flux-surface-geometry lines; W. A. Houlberg (NCLASS, as carried by NEO).
• CEA/IRFM and the METIS authors (J. F. Artaud et al.) — source of the fast neutral-beam model and of the 0-D conventions, and the reference for the 0-D comparison.
• P. B. Snyder, O. Meneghini and the EPED / EPED1-NN authors (General Atomics) — source of the pedestal model: EPED1's two constraints (Phys. Plasmas 16 056118 (2009)) reach this project through the EPED1-NN surrogate (Nucl. Fusion 57 086034 (2017)), with the network weights taken from the open-source EPEDNN.jl (ProjectTorreyPines, Apache-2.0; license and checksum vendored in this repository).
• The authors of TokSys (General Atomics) — cross-code anchor for the electromagnetics and vertical stability.
• The EAST team at ASIPP — discharge data, the operational EFIT workflow and per-channel diagnostic geometry.
• J. P. Qian, G. Q. Li, Q. L. Ren, L. L. Lao et al. — for porting EFIT to EAST and building its kinetic reconstruction (KEFIT), the upstream reference baseline of the reconstruction line here.
• Youwen Sun — the HT-7 soft-X-ray tomography work (camera-geometry parameterisation and the chord-Green / weighted pseudo-inverse method skeleton).
• Ting Lan — for curating the EAST experimental data corpus; its per-channel diagnostic geometry reaches this project through fydata.
• Yao Huang and the EAST PCS group — the plasma-control-system interface note (magnetic-probe geometry and the equilibrium↔PCS variables).
• Xuemin WU, Hui SHENG — guidance on and clarification of the KEFIT operating procedure, and the benchmark cases they provided.
• The ITER Organization — the IMAS Data Dictionary and the ITER machine description, reaching this project through fyo and fydata; the ITER configuration built into the demos comes from there.
• The authors of the works listed in the following section — every formula used here names its source; none originates in this project.

★If you find work of yours used here and not named, please open an issue — the omission is an error, not a choice.

References

Every method used on this site is published method. ★This section previously sat on the entrance page and now lives here: provenance and acknowledgement are one subject and should not be maintained in two places.

Equilibrium and reconstruction

• The equilibrium itself is the Grad–Shafranov equation (Grad and Rubin 1958; Shafranov 1966).
• Fitting p′/FF′ polynomials to magnetic measurements and alternating with the equilibrium under the plasma-current constraint is the reconstruction framework of Lao et al.:
  L. L. Lao, H. St. John, R. D. Stambaugh, A. G. Kellman, W. Pfeiffer, Reconstruction of current profile parameters and plasma shapes in tokamaks, Nuclear Fusion 25 (1985) 1611.
• Free-boundary solving and the stabilisation of the vertical position:
  S. C. Jardin, Computational Methods in Plasma Physics, CRC Press, 2010.
• The fast direct solve on a regular grid:
  R. W. Hockney, A fast direct solution of Poisson's equation using Fourier analysis, Journal of the ACM 12 (1965) 95.
• Definitions of the safety factor, the shape quantities and the general tokamak background:
  J. Wesson, Tokamaks, 4th ed., Oxford University Press, 2011.

Transport and turbulence

• Analytic neoclassical conductivity and bootstrap current:
  O. Sauter, C. Angioni, Y. R. Lin-Liu, Neoclassical conductivity and bootstrap current formulas, Physics of Plasmas 6 (1999) 2834.
• The bootstrap recalibration, kept apart from the line above — the two readings differ by several per cent:
  A. Redl, C. Angioni, E. Belli, O. Sauter, A new set of analytical formulae for the computation of the bootstrap current, Physics of Plasmas 28 (2021) 022502.
• Analytic neoclassical variants and the trapped fraction:
  C. S. Chang, F. L. Hinton, Physics of Fluids 25 (1982) 1493; Y. R. Lin-Liu, R. L. Miller, Physics of Plasmas 2 (1995) 1666.
• The gyro-Landau-fluid quasilinear transport model (TGLF):
  G. M. Staebler, J. E. Kinsey, R. E. Waltz, A theory-based transport model with comprehensive physics, Physics of Plasmas 14 (2007) 055909.
• Direct solution of the drift-kinetic equation (NEO):
  E. A. Belli, J. Candy, Kinetic calculation of neoclassical transport including self-consistent electron and impurity dynamics, Plasma Physics and Controlled Fusion 50 (2008) 095010.

0-D modelling, sources and atomic data

• Energy-confinement scalings:
  ITER Physics Basis, Nuclear Fusion 39 (1999) 2175 (IPB98(y,2)); P. N. Yushmanov et al., Nuclear Fusion 30 (1990) 1999 (ITER89-P).
• D-T reactivity:
  H.-S. Bosch, G. M. Hale, Improved formulas for fusion cross-sections and thermal reactivities, Nuclear Fusion 32 (1992) 611.
• The L–H threshold power:
  Y. Martin et al., Power requirement for accessing the H-mode, Journal of Physics: Conference Series 123 (2008) 012033.
• The reference code for 0-D integrated modelling, and the source of the fast neutral-beam model:
  J. F. Artaud et al., Metis: a fast integrated tokamak modelling tool, Nuclear Fusion 58 (2018) 105001.
• Impurity radiation cooling curves:
  T. Pütterich et al., Nuclear Fusion 59 (2019) 056013; Open-ADAS, open.adas.ac.uk.
• Beam stopping and charge-exchange cross-sections:
  R. K. Janev, C. D. Boley, D. E. Post, Penetration of energetic neutral beams into fusion plasmas, Nuclear Fusion 29 (1989) 2125.

The full texts

This page is a summary. The binding, per-file statements are in the repository:

• LICENSE — the full Apache License 2.0
• NOTICE — per-file provenance, the statement of changes, and what is not included — it travels with the distribution, and this copy sits at the root of this site
• docs/ACKNOWLEDGEMENTS.md — the readable acknowledgement list: upstream code, transcribed formulae, data and fixtures, and what was deliberately not taken
• CONTRIBUTORS.md — maintainers and copyright

Repository: github.com/fusion-yun/fylite