fabtwin 0.1.0__tar.gz

fabtwin-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

fabtwin-0.1.0/PKG-INFO

@@ -0,0 +1,189 @@
+Metadata-Version: 2.4
+Name: fabtwin
+Version: 0.1.0
+Summary: Learned generative process twins for yield-aware inverse design of multilayer optics: exact TMM with a hand-derived adjoint, CVaR robustification through twin and physics at once
+Author: Tanvir Mahmud Mahim
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/TaN-MM-Org/fabtwin
+Project-URL: Paper, https://github.com/Tanvir-Mahmud-Mahim/Learned-generative-process-twins-for-yield-aware-inverse-design-of-multilayer-photonic-sensor
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Physics
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.26
+Requires-Dist: scipy>=1.11
+Provides-Extra: twin
+Requires-Dist: jax>=0.4.30; extra == "twin"
+Requires-Dist: optax>=0.2; extra == "twin"
+Provides-Extra: test
+Requires-Dist: pytest>=7; extra == "test"
+Requires-Dist: tmm>=0.1.8; extra == "test"
+Dynamic: license-file
+
+# fabtwin
+
+[![tests](https://github.com/TaN-MM-Org/fabtwin/actions/workflows/ci.yml/badge.svg)](https://github.com/TaN-MM-Org/fabtwin/actions)
+[![PyPI](https://img.shields.io/pypi/v/fabtwin)](https://pypi.org/project/fabtwin/)
+[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
+
+**Learned generative process twins for yield-aware inverse design of
+multilayer optics.** Real deposition errors are systematic,
+layer-correlated, design-conditional, skewed and heavy-tailed, and
+manufacturing yield is decided precisely in the tail that hand-specified
+Gaussian tolerance models represent worst. `fabtwin` is the generalized
+library form of the FabGAN-ID framework (Mahim, Islam, Rahman and
+Mohsin, *IEEE Sensors Journal*, 2026): learn the fabrication process
+from historical (recipe, outcome) traces, then robustify designs by
+ascending the yield-deciding lower-tail statistic **pathwise through the
+learned twin and the exact physics at once**.
+
+> Learn what cannot be simulated; differentiate what can.
+
+Where the paper's companion repository reproduces the paper (fixed
+20-layer SiNx notch platform, JAX required end to end), `fabtwin` is the
+reusable instrument: any layer count, any design box, any wavelength
+grid, any cited material system, any linear-in-T merit -- and the entire
+differentiable fabrication loop runs on **NumPy alone**, because the
+discrete adjoint of the transfer-matrix recursion is derived by hand
+rather than delegated to an autodiff framework. The learned conditional
+WGAN-GP twin is an optional extra (`pip install 'fabtwin[twin]'`, JAX +
+optax), and its autodiff gradients are asserted equal to the hand
+adjoint to machine precision -- two independent derivations of the same
+physics agreeing is the package's core cross-validation.
+
+## What is inside
+
+- `fabtwin.tmm` -- exact characteristic-matrix optics (Macleod) for
+  arbitrary stratified stacks: normal or oblique incidence (s/p),
+  absorbing layers (n + ik, gain refused), reflectance/transmittance,
+  and linear-in-T merit builders (fluorescence-rejection notch and
+  bandpass, per the paper's sensor front-ends). Cross-validated against
+  the open `tmm` reference (Byrnes, arXiv:1603.02720) to 1e-12 and
+  against closed forms: bare-interface Fresnel, the exact absentee
+  half-wave layer, the exact quarter-wave antireflection transform, the
+  analytic Brewster zero, and R + T = 1 at machine precision.
+- `fabtwin.adjoint` -- the hand-derived exact adjoint: `merit_and_grad`
+  returns dJ/d(thickness) and dJ/d(index) from two O(NL) sweeps of
+  prefix/suffix products. Matches central finite differences at the
+  paper's own accuracy figure and vanishes exactly at the closed-form
+  quarter-wave optimum.
+- `fabtwin.materials` -- a cited Sellmeier engine (Si3N4: Luke 2015;
+  fused silica: Malitson 1965; both CC0 via refractiveindex.info) with
+  validity-range refusal instead of silent extrapolation, and the
+  variable-index layer construction of the single-material SiNx
+  platform (Yesilyurt 2023).
+- `fabtwin.process` -- the six-mechanism virtual deposition process of
+  the FabGAN-ID benchmark as a parametric dataclass (systematic bias,
+  index drift, intermixing, design-conditional AR(1) thickness noise,
+  right-skewed index noise, rare particulates): every mechanism
+  re-scalable or removable, the paper's published magnitudes as cited
+  defaults, and an exact closed-form anchor when the randomness is off.
+- `fabtwin.twins` -- the common error parameterization
+  x = (t̃/t − 1, ñ − n) with exact round trip, and Gaussian twins
+  (diagonal/full) as differentiable parametric baselines.
+- `fabtwin.risk` -- CVaR (exact sorted-tail convention, Rockafellar &
+  Uryasev 2000), tail statistics, the hard pass/fail filter yield, and
+  Monte-Carlo scoring of a design under any process or twin.
+- `fabtwin.design` -- the probe-seeded projected-Adam adjoint engine
+  and its equal-budget random-search baseline; recovers the analytic
+  quarter-wave optimum in the tests.
+- `fabtwin.robust` -- pathwise CVaR (or mean − beta*sigma)
+  robustification in pure NumPy for affinely reparameterized twins,
+  with the frozen-latent gradient anchored against finite differences.
+- `fabtwin.twin_jax` (the `[twin]` extra) -- the conditional,
+  moment-matched WGAN-GP process twin, optional physics-in-the-loop
+  tail calibration (the paper's honest finding is reproduced in the
+  docstring: it helps only when traces are scarce), a JAX solver
+  asserted equal to the NumPy one, and `robustify_gan`: the paper's
+  Stage 2, CVaR ascent through generator and solver jointly.
+
+## Install
+
+```
+pip install fabtwin              # NumPy core: physics, adjoint, Gaussian twins, CVaR loop
+pip install 'fabtwin[twin]'      # + JAX/optax: the learned WGAN-GP twin
+```
+
+## The loop in five lines
+
+```python
+import numpy as np, fabtwin as ft
+
+lam = np.linspace(0.40, 0.80, 161)                     # um
+S = ft.dispersion_shape(lam, 0.550)                    # SiNx shape (Luke 2015)
+nsub = ft.SIO2_MALITSON1965.n(lam)                     # substrate (Malitson 1965)
+w, c0 = ft.notch_weights(lam, 0.532, 0.015, 0.030)     # 532-nm rejection merit
+box = ft.DesignBox(20, 0.020, 0.120, 1.6, 2.4)         # fabrication box
+
+t, n0, J, _ = ft.inverse_design(lam, S, w, c0, box, n_sub=nsub)   # nominal optimum
+
+rng = np.random.default_rng(0)
+recipes = box.sample(rng, 200)
+rt, rn, ftd, fnd = ft.PAPER_PROCESS.trace_dataset(*recipes, 2, rng)  # historical traces
+twin = ft.GaussianTwin(ft.errors_from_traces(rt, rn, ftd, fnd))      # or train the WGAN twin
+
+t_rob, n_rob, hist = ft.robustify(lam, t, n0, S, w, c0, twin, box,
+                                  alpha=0.05, n_sub=nsub)            # CVaR ascent
+report = ft.evaluate_under_process(
+    lambda a, b, K: ft.PAPER_PROCESS.ensemble(a, b, K, rng),
+    t_rob, n_rob, lam, S, w, c0, K=2000, n_sub=nsub)                 # score vs the process
+```
+
+The learned twin drops in through the `[twin]` extra:
+`fabtwin.twin_jax.train_wgan` on the same traces, then
+`robustify_gan` ascends the identical objective through the generator.
+
+## Status
+
+v0.1.0 (alpha). Implemented and tested (37 tests, Python 3.10-3.13;
+the JAX extra's tests skip cleanly without it): everything listed
+above, with every physics claim anchored to a closed form, an
+independent reference implementation, or an exact identity -- never to
+a stored number. The API may change before v1.0.
+
+Deliberate scope, designed out with reasons rather than overlooked:
+
+- The hand adjoint covers normal incidence and lossless (real-index)
+  designs -- the regime of the FabGAN-ID loop. The oblique and
+  absorbing *forward* solver is provided; its gradients are not faked.
+- No specification-conditioned correction policies (the paper's Stage
+  3): the paper itself scopes that study as exploratory, and a policy
+  layer would import its chain-GCN encoder wholesale rather than
+  generalize it. The pathwise machinery a policy trainer needs is all
+  here.
+- No neural forward surrogates, by *result*: the paper's protocol
+  study found the exact differentiable solver strictly dominates them
+  in this regime (faster, exact, with exact gradients), so shipping
+  one would package a documented mistake.
+- The benchmark data (300 designs / 48,300 samples / 400 traces) stays
+  with the paper's companion repository and its Zenodo archive; this
+  package regenerates equivalent data from `DepositionProcess` instead
+  of redistributing files.
+
+## Associated paper
+
+T. M. Mahim, M. N. Islam, M. M. Rahman, A. S. M. Mohsin, "FabGAN-ID:
+Learning the Fabrication Process for Yield-Aware Inverse Design of
+Multilayer Photonic Sensor Filters", *IEEE Sensors Journal* (2026).
+Companion repository:
+[Learned-generative-process-twins...](https://github.com/Tanvir-Mahmud-Mahim/Learned-generative-process-twins-for-yield-aware-inverse-design-of-multilayer-photonic-sensor)
+(benchmark archived on Zenodo, doi:10.5281/zenodo.21315793).
+
+## Support and governance
+
+The package is written and maintained by Tanvir Mahmud Mahim
+(Department of Electrical and Electronic Engineering, BRAC University),
+who reviews every change and takes the final decision on scope and
+releases. There is no separate governance body; design questions are
+discussed in the open in issues and pull requests, and the standing
+rule of [CONTRIBUTING.md](CONTRIBUTING.md) binds the maintainer exactly
+as it binds contributors: a change that touches physics arrives with a
+test, and a constant arrives with its source.
+
+## License and citation
+
+Apache-2.0. Cite via [CITATION.cff](CITATION.cff) and the associated
+paper above.

fabtwin-0.1.0/README.md

@@ -0,0 +1,164 @@
+# fabtwin
+
+[![tests](https://github.com/TaN-MM-Org/fabtwin/actions/workflows/ci.yml/badge.svg)](https://github.com/TaN-MM-Org/fabtwin/actions)
+[![PyPI](https://img.shields.io/pypi/v/fabtwin)](https://pypi.org/project/fabtwin/)
+[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
+
+**Learned generative process twins for yield-aware inverse design of
+multilayer optics.** Real deposition errors are systematic,
+layer-correlated, design-conditional, skewed and heavy-tailed, and
+manufacturing yield is decided precisely in the tail that hand-specified
+Gaussian tolerance models represent worst. `fabtwin` is the generalized
+library form of the FabGAN-ID framework (Mahim, Islam, Rahman and
+Mohsin, *IEEE Sensors Journal*, 2026): learn the fabrication process
+from historical (recipe, outcome) traces, then robustify designs by
+ascending the yield-deciding lower-tail statistic **pathwise through the
+learned twin and the exact physics at once**.
+
+> Learn what cannot be simulated; differentiate what can.
+
+Where the paper's companion repository reproduces the paper (fixed
+20-layer SiNx notch platform, JAX required end to end), `fabtwin` is the
+reusable instrument: any layer count, any design box, any wavelength
+grid, any cited material system, any linear-in-T merit -- and the entire
+differentiable fabrication loop runs on **NumPy alone**, because the
+discrete adjoint of the transfer-matrix recursion is derived by hand
+rather than delegated to an autodiff framework. The learned conditional
+WGAN-GP twin is an optional extra (`pip install 'fabtwin[twin]'`, JAX +
+optax), and its autodiff gradients are asserted equal to the hand
+adjoint to machine precision -- two independent derivations of the same
+physics agreeing is the package's core cross-validation.
+
+## What is inside
+
+- `fabtwin.tmm` -- exact characteristic-matrix optics (Macleod) for
+  arbitrary stratified stacks: normal or oblique incidence (s/p),
+  absorbing layers (n + ik, gain refused), reflectance/transmittance,
+  and linear-in-T merit builders (fluorescence-rejection notch and
+  bandpass, per the paper's sensor front-ends). Cross-validated against
+  the open `tmm` reference (Byrnes, arXiv:1603.02720) to 1e-12 and
+  against closed forms: bare-interface Fresnel, the exact absentee
+  half-wave layer, the exact quarter-wave antireflection transform, the
+  analytic Brewster zero, and R + T = 1 at machine precision.
+- `fabtwin.adjoint` -- the hand-derived exact adjoint: `merit_and_grad`
+  returns dJ/d(thickness) and dJ/d(index) from two O(NL) sweeps of
+  prefix/suffix products. Matches central finite differences at the
+  paper's own accuracy figure and vanishes exactly at the closed-form
+  quarter-wave optimum.
+- `fabtwin.materials` -- a cited Sellmeier engine (Si3N4: Luke 2015;
+  fused silica: Malitson 1965; both CC0 via refractiveindex.info) with
+  validity-range refusal instead of silent extrapolation, and the
+  variable-index layer construction of the single-material SiNx
+  platform (Yesilyurt 2023).
+- `fabtwin.process` -- the six-mechanism virtual deposition process of
+  the FabGAN-ID benchmark as a parametric dataclass (systematic bias,
+  index drift, intermixing, design-conditional AR(1) thickness noise,
+  right-skewed index noise, rare particulates): every mechanism
+  re-scalable or removable, the paper's published magnitudes as cited
+  defaults, and an exact closed-form anchor when the randomness is off.
+- `fabtwin.twins` -- the common error parameterization
+  x = (t̃/t − 1, ñ − n) with exact round trip, and Gaussian twins
+  (diagonal/full) as differentiable parametric baselines.
+- `fabtwin.risk` -- CVaR (exact sorted-tail convention, Rockafellar &
+  Uryasev 2000), tail statistics, the hard pass/fail filter yield, and
+  Monte-Carlo scoring of a design under any process or twin.
+- `fabtwin.design` -- the probe-seeded projected-Adam adjoint engine
+  and its equal-budget random-search baseline; recovers the analytic
+  quarter-wave optimum in the tests.
+- `fabtwin.robust` -- pathwise CVaR (or mean − beta*sigma)
+  robustification in pure NumPy for affinely reparameterized twins,
+  with the frozen-latent gradient anchored against finite differences.
+- `fabtwin.twin_jax` (the `[twin]` extra) -- the conditional,
+  moment-matched WGAN-GP process twin, optional physics-in-the-loop
+  tail calibration (the paper's honest finding is reproduced in the
+  docstring: it helps only when traces are scarce), a JAX solver
+  asserted equal to the NumPy one, and `robustify_gan`: the paper's
+  Stage 2, CVaR ascent through generator and solver jointly.
+
+## Install
+
+```
+pip install fabtwin              # NumPy core: physics, adjoint, Gaussian twins, CVaR loop
+pip install 'fabtwin[twin]'      # + JAX/optax: the learned WGAN-GP twin
+```
+
+## The loop in five lines
+
+```python
+import numpy as np, fabtwin as ft
+
+lam = np.linspace(0.40, 0.80, 161)                     # um
+S = ft.dispersion_shape(lam, 0.550)                    # SiNx shape (Luke 2015)
+nsub = ft.SIO2_MALITSON1965.n(lam)                     # substrate (Malitson 1965)
+w, c0 = ft.notch_weights(lam, 0.532, 0.015, 0.030)     # 532-nm rejection merit
+box = ft.DesignBox(20, 0.020, 0.120, 1.6, 2.4)         # fabrication box
+
+t, n0, J, _ = ft.inverse_design(lam, S, w, c0, box, n_sub=nsub)   # nominal optimum
+
+rng = np.random.default_rng(0)
+recipes = box.sample(rng, 200)
+rt, rn, ftd, fnd = ft.PAPER_PROCESS.trace_dataset(*recipes, 2, rng)  # historical traces
+twin = ft.GaussianTwin(ft.errors_from_traces(rt, rn, ftd, fnd))      # or train the WGAN twin
+
+t_rob, n_rob, hist = ft.robustify(lam, t, n0, S, w, c0, twin, box,
+                                  alpha=0.05, n_sub=nsub)            # CVaR ascent
+report = ft.evaluate_under_process(
+    lambda a, b, K: ft.PAPER_PROCESS.ensemble(a, b, K, rng),
+    t_rob, n_rob, lam, S, w, c0, K=2000, n_sub=nsub)                 # score vs the process
+```
+
+The learned twin drops in through the `[twin]` extra:
+`fabtwin.twin_jax.train_wgan` on the same traces, then
+`robustify_gan` ascends the identical objective through the generator.
+
+## Status
+
+v0.1.0 (alpha). Implemented and tested (37 tests, Python 3.10-3.13;
+the JAX extra's tests skip cleanly without it): everything listed
+above, with every physics claim anchored to a closed form, an
+independent reference implementation, or an exact identity -- never to
+a stored number. The API may change before v1.0.
+
+Deliberate scope, designed out with reasons rather than overlooked:
+
+- The hand adjoint covers normal incidence and lossless (real-index)
+  designs -- the regime of the FabGAN-ID loop. The oblique and
+  absorbing *forward* solver is provided; its gradients are not faked.
+- No specification-conditioned correction policies (the paper's Stage
+  3): the paper itself scopes that study as exploratory, and a policy
+  layer would import its chain-GCN encoder wholesale rather than
+  generalize it. The pathwise machinery a policy trainer needs is all
+  here.
+- No neural forward surrogates, by *result*: the paper's protocol
+  study found the exact differentiable solver strictly dominates them
+  in this regime (faster, exact, with exact gradients), so shipping
+  one would package a documented mistake.
+- The benchmark data (300 designs / 48,300 samples / 400 traces) stays
+  with the paper's companion repository and its Zenodo archive; this
+  package regenerates equivalent data from `DepositionProcess` instead
+  of redistributing files.
+
+## Associated paper
+
+T. M. Mahim, M. N. Islam, M. M. Rahman, A. S. M. Mohsin, "FabGAN-ID:
+Learning the Fabrication Process for Yield-Aware Inverse Design of
+Multilayer Photonic Sensor Filters", *IEEE Sensors Journal* (2026).
+Companion repository:
+[Learned-generative-process-twins...](https://github.com/Tanvir-Mahmud-Mahim/Learned-generative-process-twins-for-yield-aware-inverse-design-of-multilayer-photonic-sensor)
+(benchmark archived on Zenodo, doi:10.5281/zenodo.21315793).
+
+## Support and governance
+
+The package is written and maintained by Tanvir Mahmud Mahim
+(Department of Electrical and Electronic Engineering, BRAC University),
+who reviews every change and takes the final decision on scope and
+releases. There is no separate governance body; design questions are
+discussed in the open in issues and pull requests, and the standing
+rule of [CONTRIBUTING.md](CONTRIBUTING.md) binds the maintainer exactly
+as it binds contributors: a change that touches physics arrives with a
+test, and a constant arrives with its source.
+
+## License and citation
+
+Apache-2.0. Cite via [CITATION.cff](CITATION.cff) and the associated
+paper above.

fabtwin-0.1.0/pyproject.toml

@@ -0,0 +1,30 @@
+[build-system]
+requires = ["setuptools>=68"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "fabtwin"
+version = "0.1.0"
+description = "Learned generative process twins for yield-aware inverse design of multilayer optics: exact TMM with a hand-derived adjoint, CVaR robustification through twin and physics at once"
+readme = "README.md"
+requires-python = ">=3.10"
+license = {text = "Apache-2.0"}
+authors = [{name = "Tanvir Mahmud Mahim"}]
+dependencies = ["numpy>=1.26", "scipy>=1.11"]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Intended Audience :: Science/Research",
+    "Programming Language :: Python :: 3",
+    "Topic :: Scientific/Engineering :: Physics",
+]
+
+[project.urls]
+Homepage = "https://github.com/TaN-MM-Org/fabtwin"
+Paper = "https://github.com/Tanvir-Mahmud-Mahim/Learned-generative-process-twins-for-yield-aware-inverse-design-of-multilayer-photonic-sensor"
+
+[project.optional-dependencies]
+twin = ["jax>=0.4.30", "optax>=0.2"]
+test = ["pytest>=7", "tmm>=0.1.8"]
+
+[tool.setuptools.packages.find]
+where = ["src"]

fabtwin-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

fabtwin-0.1.0/src/fabtwin/__init__.py

@@ -0,0 +1,43 @@
+"""fabtwin: learned generative process twins for yield-aware inverse
+design of multilayer optics.
+
+Generalized library form of the FabGAN-ID framework (T. M. Mahim,
+M. N. Islam, M. M. Rahman, A. S. M. Mohsin, "FabGAN-ID: Learning the
+Fabrication Process for Yield-Aware Inverse Design of Multilayer
+Photonic Sensor Filters", IEEE Sensors Journal, 2026): exact
+transfer-matrix physics with a hand-derived discrete adjoint (no
+autodiff framework required), a parametric six-mechanism virtual
+deposition process, Gaussian and learned (conditional WGAN-GP, the
+optional [twin] JAX extra) process twins on a common error
+parameterization, probe-seeded adjoint inverse design, and pathwise
+CVaR robustification through twin and physics at once. The guiding
+principle of the paper is the guiding principle of the package:
+learn what cannot be simulated; differentiate what can.
+"""
+from .materials import (SI3N4_LUKE2015, SIO2_MALITSON1965,
+                        SellmeierMaterial, dispersion_shape, layer_index,
+                        sellmeier)
+from .tmm import (bandpass_weights, merit, notch_weights, reflectance,
+                  stack_rt, transmittance)
+from .adjoint import merit_and_grad, transmittance_and_grads
+from .process import PAPER_PROCESS, DepositionProcess
+from .twins import GaussianTwin, apply_errors, errors_from_traces
+from .risk import (cvar, evaluate_under_process, pass_fail,
+                   tail_statistics, yield_fraction)
+from .design import DesignBox, adam_ascent, inverse_design, random_search
+from .robust import cvar_objective_and_grad, robustify
+
+__version__ = "0.1.0"
+__all__ = [
+    "SellmeierMaterial", "SI3N4_LUKE2015", "SIO2_MALITSON1965",
+    "sellmeier", "dispersion_shape", "layer_index",
+    "transmittance", "reflectance", "stack_rt", "merit",
+    "notch_weights", "bandpass_weights",
+    "transmittance_and_grads", "merit_and_grad",
+    "DepositionProcess", "PAPER_PROCESS",
+    "GaussianTwin", "errors_from_traces", "apply_errors",
+    "cvar", "tail_statistics", "pass_fail", "yield_fraction",
+    "evaluate_under_process",
+    "DesignBox", "inverse_design", "random_search", "adam_ascent",
+    "cvar_objective_and_grad", "robustify",
+]