pyturb 0.2.0__tar.gz

pyturb-0.2.0/.github/workflows/ci.yml

@@ -0,0 +1,45 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/ruff-action@v3
+        with:
+          args: check
+
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.9", "3.11", "3.12", "3.13"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - run: pip install -e ".[test,fits]"
+      - run: pytest -q --cov=pyturb --cov-report=xml
+      - name: Upload coverage
+        if: matrix.python-version == '3.12'
+        uses: codecov/codecov-action@v4
+        with:
+          files: coverage.xml
+        continue-on-error: true
+
+  docs:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install -e ".[docs]"
+      - run: mkdocs build --strict

pyturb-0.2.0/.github/workflows/docs.yml

@@ -0,0 +1,41 @@
+name: Docs
+
+# Build the mkdocs site and publish it to GitHub Pages on every push to main.
+# Enable Pages once at Settings -> Pages -> Source: GitHub Actions.
+
+on:
+  push:
+    branches: [main]
+
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+concurrency:
+  group: pages
+  cancel-in-progress: true
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install -e ".[docs]"
+      - run: mkdocs build --strict --site-dir _site
+      - uses: actions/upload-pages-artifact@v3
+        with:
+          path: _site
+
+  deploy:
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - id: deployment
+        uses: actions/deploy-pages@v4

pyturb-0.2.0/.github/workflows/gpu.yml

@@ -0,0 +1,33 @@
+name: GPU
+
+# The CuPy path is not exercised by the CPU-only `CI` workflow. This job runs
+# the GPU-marked tests (pytest --run-gpu) on a self-hosted runner that has a
+# CUDA GPU + CuPy. It is manual (workflow_dispatch) and, optionally, scheduled;
+# it never blocks a PR, so a repo without a GPU runner is unaffected.
+#
+# To use it: register a self-hosted runner with the labels [self-hosted, gpu]
+# on a CUDA machine, then trigger this workflow from the Actions tab. Locally,
+# the same check is just `pytest --run-gpu` in an env with cupy installed.
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly, Mondays 06:00 UTC — a low-frequency regression guard for the
+    # CuPy path. Harmless (a no-op) if no self-hosted GPU runner is online.
+    - cron: "0 6 * * 1"
+
+jobs:
+  gpu-test:
+    runs-on: [self-hosted, gpu]
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install with CUDA 12 extras
+        run: pip install -e ".[test,fits,cuda12]"
+      - name: Show GPU
+        run: |
+          python -c "import cupy; print('CuPy', cupy.__version__); \
+            print(cupy.cuda.runtime.getDeviceProperties(0)['name'])"
+      - name: Run GPU-marked tests
+        run: pytest -q --run-gpu -m gpu
+      - name: Run full suite on the GPU host (CPU + GPU)
+        run: pytest -q --run-gpu

pyturb-0.2.0/.github/workflows/release.yml

@@ -0,0 +1,37 @@
+name: Release
+
+# Build and publish to PyPI on a version tag (e.g. v0.2.0), using PyPI
+# trusted publishing (OIDC) — no API token needed. Configure the publisher
+# once at https://pypi.org/manage/project/pyturb/settings/publishing/.
+
+on:
+  push:
+    tags: ["v*"]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install build
+      - run: python -m build
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist
+          path: dist/
+
+  publish:
+    needs: build
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write        # required for trusted publishing
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          name: dist
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@release/v1

pyturb-0.2.0/.gitignore

@@ -0,0 +1,223 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#   Usually these files are written by a python script from a template
+#   before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+# Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+# uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+# poetry.lock
+# poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+# pdm.lock
+# pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+# pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#   JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#   be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#   and can be added to the global gitignore or merged into this file.  For a more nuclear
+#   option (not recommended) you can uncomment the following to ignore the entire idea folder.
+# .idea/
+
+# Abstra
+#   Abstra is an AI-powered process automation framework.
+#   Ignore directories containing user credentials, local state, and settings.
+#   Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#   Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#   that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#   and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#   you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+# Temporary file for partial code execution
+tempCodeRunnerFile.py
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml
+
+.vscode/
+# mkdocs strict build output (docs.yml)
+_site/
+trade_study_review/

pyturb-0.2.0/CHANGELOG.md

@@ -0,0 +1,116 @@
+# Changelog
+
+All notable changes to pyturb are documented here. The format follows
+[Keep a Changelog](https://keepachangelog.com/), and the project aims to adhere
+to [Semantic Versioning](https://semver.org/).
+
+## [0.2.0] 
+
+The "atmosphere" release: pyturb goes from a phase-screen library to a complete,
+benchmarked, GPU-native AO atmosphere.
+
+### Added
+
+- **`Atmosphere`** — layered atmosphere summed to pupil OPD, with per-layer
+  wind, airmass/zenith scaling, off-axis `directions=`, field-of-view
+  oversizing, and integrated `r0` / `seeing` / `theta0` / `tau0` /
+  `greenwood_frequency`. Built from named profiles via `from_profile`.
+- **Two frozen-flow engines.** `engine="spectral"` (default): exact sub-pixel
+  shift-theorem translation, all layers in one FFT, but periodic. Boiling
+  (`tau_boil`) via a spectral AR(1). `engine="extrude"`: Assémat–Wilson row
+  extrusion in a wind-aligned ring buffer with rotated sub-pixel sampling —
+  unbounded, non-periodic, any wind direction.
+- **`InfinitePhaseScreen`** rewritten with a ring buffer and sub-pixel
+  `advance()` (Catmull-Rom / linear), memory bounded over arbitrarily long runs.
+- **Named profiles**: `paranal-median`, `mauna-kea`, `keck`, `las-campanas`,
+  `cerro-pachon`, `armazones`, `hv57`, `single-layer`, `two-layer`;
+  `discretize_cn2(method=...)` with moment-conserving `"equivalent"`
+  (conserves `theta0` and `tau0`), `"centroid"`, and `"optimal_grouping"` —
+  the last chooses bin edges by dynamic programming to minimise the
+  Cn²-weighted within-group spread of `h^{5/3}`, for MCAO/tomography layer
+  compression (Saxenhuber et al. 2017).
+- **`Atmosphere.evolve(dt)`** — single-step, in-seconds frozen-flow stepper
+  (mirrors HCIPy's `evolve_until`); repeated calls reproduce `frames(dt)`.
+- **`interp="lanczos"`** (6-tap Lanczos-3) sub-pixel readout on the extruder and
+  `InfinitePhaseScreen` — a flatter sub-Nyquist kernel that cuts the extruder's
+  finest-scale travel-phase flicker (~10% → ~3.5%) and structure-function
+  deficit versus the default cubic, with no change to the extrusion statistics.
+- **`pyturb.analysis`**: Zernike basis/decomposition, Noll (1976) mode
+  variances, temporal PSD + power-law fit, angular decorrelation.
+- **I/O**: `pyturb.save` / `pyturb.load` for `.npz` and FITS (optional astropy)
+  with metadata; `Atmosphere.metadata`.
+- **Chromatic OPD**: `dispersion="edlen"` (dry air) and `dispersion="ciddor"`
+  with a `wet_fraction` water-vapour term for the mid-IR/interferometric
+  "wet–dry" problem; `pyturb.air_refractivity` and
+  `pyturb.water_vapour_refractivity`.
+- **LGS cone effect**: `Atmosphere(lgs_altitude=...)` on **both** engines — the
+  extruder samples its ring buffer on a magnified grid, the spectral engine
+  zoom-resamples each layer's screen about the pupil centre by the same factor.
+  On the spectral engine the cone now **composes with `tau_boil` boiling**,
+  closing the previous cone/boiling mutual exclusivity.
+- **Non-Kolmogorov spectra**: `PhaseScreen(power_law=..., inner_scale=...)`.
+- **Threaded CPU FFT**: `pyturb.set_fft_workers()`.
+- **GPU test path**: GPU tests marked `@pytest.mark.gpu`, run with
+  `pytest --run-gpu` (a `device` fixture parameterises statistics tests over
+  CPU/GPU); `.github/workflows/gpu.yml` runs them on a self-hosted GPU runner.
+- **`pyturb.benchmark()`** convenience; `benchmarks/bench_suite.py`
+  (per-use-case throughput sweep across CPU/GPU) and
+  `benchmarks/bench_compare.py` head-to-head vs aotools/soapy/HCIPy;
+  `validation/validate.py` gallery.
+- Docs: `docs/comparison.md`, `docs/interop.md`, `docs/validation.md`; examples
+  gallery (`examples/01`–`05`).
+- `py.typed` marker; version single-sourced from package metadata.
+
+### Changed
+
+- `Atmosphere` output is **OPD in metres** (achromatic); pass `wavelength=` for
+  phase. `PhaseScreen` / `InfinitePhaseScreen` still return radians.
+- `discretize_cn2` default method is now `"equivalent"` (moment-conserving).
+
+### Performance
+
+- **Spectral engine: collapse the layer axis before the transform.** The
+  inverse FFT and subharmonic outer product are linear and shared across
+  layers, so `Atmosphere._integrate` now sums the shifted spectra to one
+  `(n, n)` array and inverse-FFTs *once* instead of once per layer (and sums
+  each subharmonic level's `3x3` coefficients before the shared basis product).
+  Identical output; measured on an RTX 5090, 9-layer paranal-median: **CPU
+  25 → 87 fps at 512² (3.4×)**, **GPU 865 → 1232 fps (1.4×)**.
+- **Batch all subharmonic levels into one matmul.** The low-frequency
+  subharmonic correction shares one `(3, n)` sinusoid basis across levels and
+  layers, so `Atmosphere._integrate`, `PhaseScreen.generate`,
+  `FourierFlowScreen.translate` and the boiling update now evaluate every level
+  in a couple of batched matmuls instead of a Python loop over levels (which was
+  launch-latency bound on the GPU — ~78% of a frame). Identical output.
+  Measured on an RTX 5090, 9-layer paranal-median frozen flow: **1,232 → 3,004
+  fps at 512² GPU**, **3,234 fps at 256²**; single-layer Monte-Carlo generation
+  **14,000 → 31,000 screens/s at 512²** (55,000 → 108,000 at 256²); CPU
+  spectral **~87 → ~130 fps at 512²** before the accel extra below.
+- **Fused GPU/CPU extruder readout kernel.** Every layer's ring buffer is a slab
+  of one contiguous `(L, cap, W)` array, and the per-frame rotated, sub-pixel,
+  per-layer-wind-shifted pupil gather runs in a single pass for the `"cubic"`
+  and `"lanczos"` interpolators: a hand-written CUDA kernel on the GPU and a
+  fused `prange` Numba kernel on the CPU (see the accel extra below), bit-exact
+  with the previous tap-broadcast gather. Measured on an RTX 5090, 9-layer
+  paranal-median: **121 → 4,484 fps at 256² GPU (37×)**, **118 → 1,730 fps at
+  512² (15×)**, **50 → 602 fps at 1024²**; the `"lanczos"` readout is now a
+  fused kernel too (~334 fps at 512² GPU, from ~120 fps).
+- **Optional Numba CPU acceleration (`pip install pyturb[accel]`).** The CPU
+  frozen-flow hot paths — the spectral engine's fused layer sum and the
+  extruder's fused bicubic/Lanczos readout — run through Numba when it is
+  importable, with a NumPy fallback otherwise (identical results to float
+  round-off). Measured on a 32-core CPU, 9-layer paranal-median: spectral
+  **~130 → 270 fps at 512²**, extruder **6 → 164 fps at 512² (27×)** and
+  **28 → 966 fps at 256² (34×)**.
+- **Geometry-derived extruder buffer sizing.** The shared ring buffer is now
+  sized to the largest along-wind/off-axis requirement actually present among
+  the layers (each layer's own wind direction and altitude), not a blanket
+  every-layer-at-45-degrees-and-max-altitude assumption. Measured (n=512): a
+  ground-layer-only atmosphere with `field_of_view=30"` uses ~68% less buffer
+  memory; an axis-aligned atmosphere uses ~41% less even at
+  `field_of_view=0`. Never worse than before.
+
+## [0.1.0]
+
+- Initial release: `PhaseScreen` (FFT + subharmonics) and `InfinitePhaseScreen`
+  (Assémat–Wilson extrusion), NumPy/CuPy backends, structure-function tests.

pyturb-0.2.0/CLAUDE.md

@@ -0,0 +1,73 @@
+# CLAUDE.md
+
+Guidance for agents working in this repo. Keep it current if the CI workflow changes.
+
+## Before considering any change done
+
+Run these from the repo root (activate an env with the project installed
+editable, e.g. `pip install -e ".[test,fits,docs]"`). All four mirror
+`.github/workflows/ci.yml` exactly — if they pass locally, CI passes.
+
+```bash
+ruff check .                                    # lint (must be clean, zero errors)
+python -m pytest -q --cov=pyturb --cov-report=term-missing   # full test suite + coverage
+mkdocs build --strict                           # docs (only if you touched README/docs/mkdocs.yml)
+python -c "import pyturb"                       # sanity import after any src/ change
+```
+
+CI additionally runs the test suite on Python 3.9, 3.11, 3.12, and 3.13. If
+you only have one interpreter available, at minimum grep your diff for
+anything that needs Python >=3.10 (`match` statements, `X | Y` type unions
+used at runtime, etc.) — the project floor is `>=3.9`, and code should not
+silently assume a newer numpy either (e.g. `np.trapezoid` requires NumPy
+>= 2.0 and `np.trapz` was removed in a later release; the `numpy>=1.22` floor
+needs a `np.trapezoid if hasattr(np, "trapezoid") else np.trapz` fallback,
+already used in `profiles.py` — note `hasattr`, not `getattr`'s default,
+since `getattr(np, "trapezoid", np.trapz)` still evaluates `np.trapz` eagerly
+and breaks on NumPy releases that no longer have it). If in doubt, spin up a throwaway
+`conda create -n py39check python=3.9` and run the suite there — this has
+caught real bugs before.
+
+## What "done" means here, beyond green tests
+
+- **Exercise the actual behavior, not just the code path.** A test that
+  calls a function and checks it doesn't throw is not a correctness test.
+  Assert on values, statistics, or invariants that would actually catch the
+  bug you just fixed or could plausibly introduce. See
+  `tests/test_extrude.py::test_finescale_readout_flicker_is_bounded` or
+  `tests/test_atmosphere.py::test_boiling_is_scale_dependent_not_uniform`
+  for the pattern: characterize the real physical/statistical behavior with
+  a bounded assertion, not just "it ran."
+- **Check edge cases the existing tests don't reach**: a single large jump
+  vs. many small steps (ring-buffer code in `extrude.py`/`infinite.py` has
+  been bitten by this — compaction logic that only gets exercised by tiny
+  steps hides bugs that surface on one big one), off-grid/boundary requests,
+  values outside a declared range.
+- **If you touch statistical/physical code**, verify against theory or a
+  known reference where one exists (structure function vs. von Kármán
+  theory, θ₀/τ₀ formulas, a cited profile table) rather than just checking
+  the code runs. Don't trust a single-realization measurement — several
+  seeds, or an ensemble average, distinguish a real effect from noise.
+- **Match error message quality to the rest of the codebase**: when
+  rejecting an invalid combination (see `Atmosphere.__init__`'s many
+  `ValueError`s), say *why*, not just *that*. A bare "X requires Y" forces
+  the next reader to spelunk the source to find out if it's a permanent
+  architectural fact or a gap that might get lifted.
+- **Update docstrings/README/RESULTS.md claims when behavior changes.**
+  Several of the bugs found were docs stating something the code didn't
+  actually do (or a claim that didn't survive scrutiny, e.g. a benchmark
+  ranking within its own noise). A code fix that leaves a stale claim in
+  place isn't finished.
+
+## Style notes specific to this repo
+
+- Comments and docstrings describe **current** behavior only — never
+  "no more X" / "previously Y, now Z" / references to a past bug or a
+  specific review. A future reader has no context for what "before" means;
+  state what the code does now. (`CHANGELOG.md` is the one place that's
+  supposed to narrate change over time.)
+- Type annotations use `from __future__ import annotations` +
+  `typing.Optional`/`Union` (not bare `X | Y`), to stay valid on the
+  `>=3.9` floor.
+- `ruff` line length is 90 (`pyproject.toml`); wrap before that, don't
+  disable the rule.

pyturb-0.2.0/CONTRIBUTING.md

@@ -0,0 +1,54 @@
+# Contributing to pyturb
+
+Thanks for your interest! pyturb aims to be the fastest, GPU-native, and
+statistically-careful way to get atmospheric OPD into an AO workflow. A few
+conventions keep it that way.
+
+## Development setup
+
+```bash
+git clone https://github.com/jacotay7/pyturb
+cd pyturb
+pip install -e ".[test]"          # add ",fits" for the FITS I/O tests
+pytest -q
+```
+
+For GPU work, install a CuPy build matching your CUDA toolkit
+(`pip install cupy-cuda12x`); the suite skips GPU tests when CuPy is absent.
+
+## The bar for a change
+
+pyturb's credibility rests on three habits — please keep them:
+
+1. **Every physics feature lands with an ensemble-statistics test against
+   theory.** New turbulence behaviour must be shown to match a closed form
+   (structure function, Noll variances, a PSD slope, …), not just "look right".
+   See `tests/` for the pattern and `validation/validate.py` for the gallery.
+2. **Every performance claim lands with a benchmark.** If you speed something
+   up, add or update a script under `benchmarks/`.
+3. **Every user-facing feature lands with docs.** A docstring at minimum; a
+   `docs/` page or example if it's a new capability.
+
+## Scope
+
+pyturb is *the atmosphere*, not a full AO system. Please keep out of scope:
+WFS/DM/controller simulation, tomographic reconstructors and slope
+covariance, and Fresnel/scintillation propagation. We output
+phase/OPD and hand those effects to the tools that own them — see
+`docs/comparison.md`.
+
+## Style
+
+- `ruff check` must pass (`pip install ruff`) — CI enforces it. Match the
+  surrounding style; the repo is hand-formatted, so `ruff format` is not imposed.
+- NumPy-style docstrings; type hints on public signatures.
+- Write backend-agnostic array code (works on NumPy and CuPy); avoid
+  host↔device syncs inside hot loops.
+- Prefer `float32` defaults for GPU throughput; keep a `float64` path for
+  accuracy-sensitive setup.
+
+## Pull requests
+
+Small, focused PRs with tests are easiest to review. Note in the description
+which of the three habits above your change satisfies. Update `CHANGELOG.md`
+under *unreleased*.

pyturb-0.2.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Jacob Taylor
+
+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.