spectro-kernel 0.1.0__tar.gz

spectro_kernel-0.1.0/.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
+
+# Local-only working docs (deployment runbooks, internal design notes).
+# These are kept on disk for the maintainer but never committed: not on
+# GitHub, not in the PyPI sdist, not on the docs site.
+stuff/

spectro_kernel-0.1.0/CHANGELOG.md

@@ -0,0 +1,363 @@
+# Changelog
+
+All notable changes to `spectro-kernel` are documented here. Format follows
+[Keep a Changelog](https://keepachangelog.com/); the project follows Semantic Versioning.
+Until `1.0.0` the public API may change between minor versions.
+
+## [Unreleased]
+
+### Changed — keep deployment runbooks + internal design notes out of the public artefacts
+
+- Moved `SPECTRO_KERNEL_IMPLEMENTATION.md`, `SPECTRO_PLATFORM_PROPOSAL.md`,
+  and the four deployment cookbook entries (`deploy-overview.md`,
+  `deploy-do.md`, `deploy-netlify.md`, `publish-pypi.md`) out of the
+  tracked tree into a local-only `stuff/` directory (gitignored).
+- Effect: these documents stay on the maintainer's disk for reference but
+  no longer ship in the PyPI sdist, no longer appear on the docs site,
+  and are not visible to anyone who clones the repository.
+- The `mkdocs.yml` nav and a handful of cross-references in
+  README/docs/ROADMAP were updated to drop the now-broken links.
+
+### Added — twelfth pass (algorithm source inspection)
+
+- **Wrapper-source audit** — every catalogue entry can now be inspected
+  from the outside:
+  - Python: `spectro_kernel.registry.get_algorithm_source(name)` returns
+    `{name, class_name, module, path, repo_relative_path, github_url,
+    backend, source}` — the wrapper's class, its file on disk, a GitHub
+    permalink, and the file contents as a string.
+  - CLI: `spectro source <name>` prints the source with a header
+    pointing at the local path and the GitHub URL; add `--url-only` to
+    print just the link, or `--json` for the structured payload.
+  - MCP: a new `get_algorithm_source` transverse tool returns the same
+    JSON, so an agent can fetch and reason about the wrapper code it is
+    about to call.
+- Use case: when a wrapper delegates to an external library
+  (``backend = "easyspec" / "astropy" / "scipy"``), the user can quickly
+  confirm what the glue does and where to read the upstream code.
+- 4 new unit tests (registry + CLI), suite size 115 → 119.
+
+### Added — eleventh pass (Tier B: DAG pipelines, échelle, sky subtraction, Keplerian, telluric fit, MCP uploads)
+
+- **DAG pipelines.** `PipelineStep.depends_on: list[str]` opt-in; when any
+  step declares dependencies the pipeline is executed in topological order
+  (Kahn's algorithm) instead of the declared sequence. Cycles and unknown
+  labels error at build time. Existing presets are untouched — no
+  `depends_on`, no behaviour change. `PipelineBuilder.add(..., depends_on=[…])`
+  and the YAML `depends_on:` field both supported.
+- **Échelle support — two new algorithms.**
+  - `read_echelle_fits` (`io`, backend `astropy`): handles multi-extension
+    `BinTableHDU` files and 2D images with per-order WCS; populates
+    `ctx.spectra` (sorted by central wavelength).
+  - `merge_echelle_orders` (`stacking`, backend `numpy`): resamples orders
+    onto a common log-λ grid, averages overlaps with inverse-variance
+    weighting (falls back to plain mean when uncertainties are absent),
+    leaves inter-order gaps as `NaN`. Three weighting modes:
+    `uncertainty` / `mean` / `first`.
+- **`subtract_sky_2d`** (`preprocessing`, backend `numpy`): long-slit sky
+  subtraction — per-column polynomial fit through off-trace rows, default
+  order 1. Stores median sky level in metrics.
+- **`fit_keplerian_orbit`** (`radial_velocity`, backend `scipy`):
+  single-companion Keplerian fit `(P, K, e, ω, T₀, γ)` with bounded
+  least-squares. Auto-detects an initial period via `astropy.LombScargle`
+  when the user does not provide one. Outputs full element set + RMS
+  residual + χ² in `ctx.extras['keplerian_orbit']`.
+- **`fit_telluric_scaling`** (`correction`, backend `scipy`): given a
+  template transmission spectrum (from `synth_telluric` or supplied
+  directly) and an observed spectrum, fits the Beer-Lambert airmass
+  exponent that best removes the telluric bands. Restricts the residual
+  to pixels inside the bands (default threshold T < 0.9). Chain into
+  `remove_telluric_division` to apply.
+- **MCP file upload tools.** Two new transverse MCP tools:
+  - `request_upload_url(session_id, filename, content_type)` returns a
+    pre-signed PUT URL plus a matching GET URL. Configured via env vars
+    `SPECTRO_MCP_S3_BUCKET` (required), `SPECTRO_MCP_S3_ENDPOINT` (DO
+    Spaces / MinIO), `SPECTRO_MCP_S3_REGION`, `SPECTRO_MCP_S3_PREFIX`,
+    `SPECTRO_MCP_S3_URL_TTL`. Gracefully reports a clear error if the
+    server is not configured or `boto3` is missing.
+  - `load_spectrum_from_url(session_id, url, format_hint)` downloads the
+    file (256 MiB cap) and loads it as the session's working spectrum.
+    The downloaded file is removed after loading.
+- **Tests.** 7 new unit tests covering: DAG topological order, DAG cycle
+  detection, DAG unknown-label rejection, échelle overlap averaging,
+  telluric airmass recovery on a synthetic injection, 2D sky subtraction
+  zero-ing the background, Keplerian RV recovery on a synthetic orbit.
+  Suite size 108 → 115 tests, all green; ruff clean; `mkdocs build
+  --strict` clean.
+
+### Added — tenth pass (Dependabot fix, Sun fixture, Netlify guide, 2 last notebooks)
+
+- **Dependabot alerts cleared.** Bumped Vite 5 → 6.3.6 in
+  `website/package.json` — kills the moderate `esbuild` advisory
+  (GHSA-67mh-4wv8-2f99) without breaking the build. `npm audit` clean,
+  `pip-audit` clean.
+- **Second LFS fixture — CALSPEC Sun** (`sun_reference_stis_002.fits`,
+  ~37 KB) committed to `tests/reference/data/sun/`. New `sun_fits_path`
+  fixture in `tests/reference/conftest.py` and a `test_sun.py` module that
+  verifies loading, SNR, Na D detection and a H-alpha fit on real data.
+- **Step-by-step deployment guides — polished**:
+  - `docs/cookbook/deploy-do.md` — now reflects the production hardening
+    that landed in Tier 1 (env vars `SPECTRO_MCP_API_KEY`,
+    `SPECTRO_MCP_RATE_PER_MINUTE`, `SPECTRO_MCP_REDIS_URL`, `SENTRY_DSN`)
+    instead of describing them as future work.
+  - `docs/cookbook/deploy-netlify.md` (new) — full Netlify walkthrough for
+    the React + Vite website (UI path + CLI path), custom domain, continuous
+    deploy, costs, plus how to wire the playground to your DO MCP across
+    origins.
+- **Tier 3 notebooks — completed**:
+  - *Native vs EasySpec showdown* — same operations through both backends
+    on the same bytes, tied to `compare_*` helpers; spells out when each
+    side wins.
+  - *Claude + spectro-kernel end-to-end* — Claude Desktop config, the first
+    conversation, the tools that get called under the hood, what works and
+    what doesn't (browser FITS upload, long jobs, image rendering).
+
+Counts: 70 algorithms across 22 categories; 6 presets; 7 notebooks;
+121 tests passing + 4 new on-Sun reference tests (1 skipped without Redis);
+ruff clean; mkdocs --strict clean; 2 LFS fixtures pushed.
+
+### Added — ninth pass (3 comparators, 2 notebooks, real LFS fixture)
+
+- **3 shelf comparators**:
+  - `compare_snr_methods` (quality) — runs `snr_der` + `snr_edge` +
+    `snr_linear_fit` on the same spectrum; the spread is recorded as a
+    metric so big disagreement flags a noisy or contaminated input.
+  - `compare_smoothings` (smoothing) — runs `smooth_savgol` + `smooth_gaussian`
+    side by side, stashes each output in `ctx.extras['smoothings']` plus
+    pairwise RMS.
+  - `compare_line_fits` (line_fitting) — fits one line with Gaussian /
+    Lorentzian / Voigt; reports R² of each and the best profile by R².
+- **2 new notebooks** in `docs/notebooks/`:
+  - *Aurora line monitor* — five short exposures, detect aurorae lines, fit
+    the [O I] 5577 green-line amplitude, render the `time_series_overview`
+    preset.
+  - *Exoplanet transit RV* — synthesise a Rossiter-McLaughlin signal,
+    recover with `cross_correlate_rv`, fit a sine, compare with the truth.
+- **First LFS-tracked fixture** — CALSPEC Vega FITS (~280 KB) lands under
+  `tests/reference/data/vega/alpha_lyr_stis_011.fits` via the `.gitattributes`
+  patterns. `tests/reference/conftest.py` now prefers the local file and falls
+  back to the Astropy on-the-fly download when it is missing or replaced by an
+  LFS pointer.
+- `BENCHMARKS.md` regenerated with Vega rows alongside the synthetic ones.
+
+Counts: 70 algorithms across 22 categories; 6 presets; 4 comparators on the
+guided shelf; 121 tests passing (1 skipped without Redis); ruff clean;
+mkdocs --strict clean.
+
+### Added — eighth pass (flux calibration, full-reduction preset, LFS, benchmarks, playground)
+
+EasySpec shelf is now end-to-end (raw FITS → flux-calibrated 1D), and the
+roadmap's Tier 3 polish items land:
+
+- `flux_calibrate_easyspec` (flux_calibration) — wraps
+  ``std_star_normalization`` + ``target_flux_calibration``. Final step of an
+  easyspec reduction; turns ADU/s into erg/s/cm²/Å using a bundled standard-star
+  archive (calspec, oke1990, …).
+- `full_reduction_easyspec` preset — template YAML chaining the 8 EasySpec
+  steps that need fixed inputs (combines, applies, CR, extract). User patches
+  `__REPLACE__` placeholders with their runtime paths before calling
+  `PipelineBuilder.from_config()`. Wavelength + extinction + flux calibration
+  stay separate (per-observation metadata).
+- New notebook **Full reduction walkthrough** — every step of the chain in one
+  page.
+- **Git LFS scaffold** — `.gitattributes` pinning `tests/reference/data/**/*.fits`
+  (and friends) to LFS, plus `tests/reference/data/README.md` documenting the
+  contributor / consumer workflow. Reference tests fall back to Astropy's
+  on-the-fly download when LFS fixtures are absent.
+- **`scripts/bench.py`** + first `BENCHMARKS.md` — per-algorithm median timing
+  on a synthetic spectrum (and optional CALSPEC Vega). Useful for catching
+  performance regressions across versions.
+- **Web playground** — `website/playground/index.html`, a no-build HTML page
+  that talks JSON-RPC to a `spectro-mcp` server (check `/health`, load a
+  spectrum by path, chain algorithms, see JSON output). Documented in
+  `docs/cookbook/web-playground.md`.
+
+Counts: 67 algorithms across 22 categories; 6 presets; 121 tests passing
+(1 skipped without Redis); ruff clean; mkdocs --strict clean.
+
+### Added — seventh pass (EasySpec extraction chain — raw FITS → flux-corrected 1D)
+
+The shelf now spans the full CCD-to-1D-spectrum chain end-to-end:
+
+- `extract_spectrum_easyspec` (extraction) — wraps
+  ``import_data`` + ``tracing`` + ``extracting`` into one call. Gaussian-weighted
+  aperture extraction with Monte-Carlo error estimation; the result is a 1D
+  Spectrum1D on a pixel-index axis, ready for wavelength calibration.
+- `wavelength_calibrate_easyspec` (wavelength_calibration) — fits a polynomial
+  from user-supplied lamp-peak identifications and applies it to
+  ``ctx.spectrum.wavelength``. easyspec's diagnostic path is called when
+  possible; the numpy.polyfit fallback guarantees the wrapper works in isolation.
+- `extinction_correct_easyspec` (correction) — applies easyspec's bundled
+  atmospheric extinction curves (lapalma, kpno, paranal, …) to ``ctx.spectrum``.
+  Wavelengths are wrapped as astropy Quantities to match easyspec's expectations.
+
+New ``algorithms/wavelength_calibration/`` and ``algorithms/extraction/``
+sub-packages with their ``__init__.py``.
+
+Counts: 66 algorithms across 21 categories; 11 EasySpec wrappers in total
+(masters · applies · extraction · wavelength · extinction); 121 tests passing
+(1 skipped without Redis); ruff clean; mkdocs --strict clean.
+
+### Added — sixth pass (EasySpec apply wrappers complete the reduction shelf)
+
+The four image-level apply operations join the EasySpec shelf, completing the
+roadmap "EasySpec wrappers" item end-to-end. Same patterns as the combine
+wrappers (paths or in-memory ImageFrames, temp-dir staging, headless plotting,
+silenced stdout); each one is a thin shim over the helper.
+
+- `subtract_bias_easyspec` (preprocessing) — wraps ``cleaning.debias``.
+- `subtract_dark_easyspec` (preprocessing) — wraps ``cleaning.sub_dark``.
+- `flat_normalize_easyspec` (preprocessing) — wraps ``cleaning.flatten``;
+  auto-normalises the input flat by its median unless told otherwise.
+- `cosmic_ray_remove_easyspec` (cosmic_ray) — wraps
+  ``cleaning.CR_and_gain_corrections``; gain/read-noise from FITS header or
+  explicit parameters.
+
+Plus `reduction/_easyspec_apply.py` factoring stage-target / resolve-master /
+result-unpacking, so each wrapper file stays ~50 lines.
+
+Counts: 63 algorithms across 21 categories; 118 tests passing (1 skipped
+without Redis); ruff clean.
+
+### Added — fifth pass (Tier 2 underway: telluric, presets, comparator, notebooks)
+
+- `synth_telluric` (corrections): synthesise a transmission template from the
+  dominant H₂O / O₂ bands, with Beer-Lambert airmass scaling. Pairs with
+  `remove_telluric_division`.
+- `compare_normalisations` (continuum, meta-algorithm): run every
+  `normalize_*` variant on the same spectrum, store each result in
+  `ctx.extras['normalisations']`, and compute pairwise RMS — turns the "shelf"
+  choice into a guided one.
+- 3 new pipeline presets in `presets/catalog/analysis/`:
+  - `rv_quick` — normalise + measure H-alpha radial velocity.
+  - `quality_report` — all three SNR estimators (DER, edge, linear-fit).
+  - `time_series_overview` — stack + dynamic spectrum + 3D surface + animation.
+- 2 notebook-style tutorials in `docs/notebooks/`:
+  - *Be-star variability* — synthesise multi-night H-alpha, run the
+    `time_series_overview` preset, inspect variability metrics.
+  - *Your first SB2* — synthesise a 6-epoch SB2, recover both components with
+    `disentangle_sb2`.
+
+Counts: 59 algorithms across 21 categories; 5 presets; 114 tests passing
+(1 skipped without Redis); ruff clean; mkdocs `--strict` clean.
+
+### Added — fourth pass (Tier 1 complete: reference tests, production MCP, EasySpec shelf)
+
+Closes the three Tier-1 items in `ROADMAP.md`:
+
+- **Reference scientific tests** — new `tests/reference/` directory with a
+  CALSPEC Vega download/cache fixture and acceptance tests that verify
+  `read_fits`, `snr_der`, continuum-normalised fitting and H-alpha detection
+  work on the canonical A0V standard. Auto-skip without network.
+- **Production-grade MCP** — the API-key middleware (`SPECTRO_MCP_API_KEY`),
+  sliding-window rate limit (`SPECTRO_MCP_RATE_PER_MINUTE`) and Redis-backed
+  session storage (`SPECTRO_MCP_REDIS_URL`) were already implemented; now
+  covered by tests (`test_mcp_production.py`) including a Redis-backend
+  lifecycle test that auto-skips when no Redis is reachable.
+- **EasySpec shelf, expanded** — extracted a shared helper
+  (`reduction/_easyspec_helpers.py`) that handles temp-directory staging,
+  headless matplotlib, and silenced stdout. `bias_combine_easyspec`
+  refactored onto it; added `dark_combine_easyspec` and
+  `flat_combine_easyspec`. End-to-end tests on synthetic FITS files.
+
+Counts: 57 algorithms across 21 categories; 113 tests passing (1 skipped
+without Redis); ruff clean.
+
+### Added — third pass (55 algorithms, 21 categories)
+
+- **5 more algorithms** rounding out the table:
+  - `plot_3d_surface_plotly`: 3D surface of stacked spectra (Plotly).
+  - `plot_animation_plotly`: animated Plotly figure that plays through `ctx.spectra`.
+  - `disentangle_sb2`: iterative wavelength-domain disentangling of an SB2 binary,
+    on a log-wavelength grid (Simon & Sturm 1994 spirit).
+  - `aperture_photometry`: differential aperture photometry on a 2D image, with
+    sky-annulus background subtraction (photutils).
+  - `bias_combine_easyspec`: first concrete EasySpec shelf wrapper — same operation
+    as the native `bias_combine`, routed through `easyspec.cleaning.master`.
+- `algorithms/advanced/` category populated (`disentangle_sb2`,
+  `aperture_photometry`).
+- `pyproject.toml`: `reduction` extra now also brings `photutils` for image-mode
+  algorithms.
+
+### Added — second pass (50 algorithms, 20 categories, cloud-ready MCP)
+
+- **14 new algorithms** raising the catalogue from 36 to 50 across 20 categories:
+  - I/O & export: `export_hdf5` (h5py).
+  - Resampling: `resample_flux_conserving` (specutils wrapper).
+  - Time-series viz: `plot_dynamic_spectrum` (Plotly heatmap of stacked spectra).
+  - Catalogues: `vizier_query`, `gaia_query` (astroquery).
+  - Radial velocity: `cross_correlate_rv` (scipy.signal cross-correlation in log-λ space),
+    `rv_precision_bouchy` (analytical photon-noise limit).
+  - Corrections: `remove_telluric_division`.
+  - CCD reduction: `bias_combine`, `dark_subtract`, `flat_normalize`,
+    `clip_cosmic_rays` (L.A.Cosmic / astroscrappy with a median-filter fallback),
+    `extract_spectrum_sum`, `wavelength_calibrate_polynomial`.
+- `adapters/easyspec.py`: scaffold for parallel "shelf" wrappers around the easyspec
+  CCD-reduction toolkit. Documented integration plan; native algorithms cover the
+  same operations today.
+- **MCP server, production hardening**:
+  - API-key auth middleware (`SPECTRO_MCP_API_KEY` env var).
+  - Sliding-window rate limiting (`SPECTRO_MCP_RATE_PER_MINUTE`).
+  - Redis-backed session storage (`SPECTRO_MCP_REDIS_URL`); in-memory backend remains
+    the default.
+  - `/health` and `/metrics` (Prometheus text format) endpoints via FastMCP's
+    `custom_route`.
+  - Structured JSON logging on every tool call (audit trail).
+  - Optional Sentry init (`SENTRY_DSN` env var).
+- **Reference scientific tests** (`tests/reference/`) — closed-form expected answers
+  with documented tolerances for SNR, line fitting, EW, Lomb-Scargle, continuum
+  normalisation, line identification, cross-correlation RV and flux preservation.
+- `pyproject.toml`: new `reduction` extra (easyspec, astroscrappy, ccdproc, matplotlib),
+  `monitoring` extra (sentry-sdk), `redis` and `uvicorn[standard]` in the `mcp` extra,
+  `h5py` in `storage`, `specutils` lifted into core deps.
+
+### Added — first pass
+
+- Provenance metadata on every algorithm — a `backend` field (the library it leans on:
+  astropy / specutils / astroquery / scipy / numpy / plotly) and `references` (literature
+  citations) — surfaced in `describe_algorithm`, `spectro describe`, the MCP tool
+  descriptions and the documentation.
+- Full MkDocs Material documentation site: "Why spectro-kernel?" (vs astropy/specutils),
+  concepts with diagrams, per-access-path guides, tutorials, and an algorithm-catalogue
+  page generated from the live registry. Read the Docs config and a docs CI workflow.
+
+## [0.1.0] - 2026-05-22
+
+Initial foundation release.
+
+### Added
+- Core types: `Spectrum1D`, `ImageFrame`, `WorkContext`, `ProcessingStep`,
+  `LineFitResult`, `Periodogram`, `AlgorithmCategory`.
+- Plugin registry: `@register_algorithm`, `list_algorithms`, `get_algorithm`,
+  `describe_algorithm`, with automatic algorithm discovery.
+- `BaseAlgorithm` / `AlgorithmOutput` with a reproducibility-aware `execute()` wrapper
+  (parameter merging, input/output hashing, audit trail).
+- `Pipeline` / `PipelineBuilder` / `PipelineResult` for composable, reproducible runs.
+- FITS and ASCII I/O (`read_fits`, `write_fits`, `read_ascii_spectrum`,
+  `write_ascii_spectrum`) with WCS reconstruction and local-path/URL support.
+- Algorithm catalogue — 36 algorithms across 15 categories (33 with the core install;
+  `simbad_query` needs `[catalogs]`, the Plotly plots need `[viz]`):
+  - I/O: `read_fits`, `read_ascii_spectrum`, `read_votable_spectrum`.
+  - Continuum: `normalize_polynomial`, `normalize_percentile`, `normalize_max`,
+    `normalize_edges`, `subtract_continuum`.
+  - Quality: `snr_der`, `snr_edge`, `snr_linear_fit`.
+  - Smoothing: `smooth_savgol`, `smooth_gaussian`.
+  - Transforms: `resample_linear`, `extract_region`, `clip_sigma`, `mask_range`.
+  - Lines: `detect_lines`, `fit_gaussian_line`, `fit_lorentzian_line`,
+    `fit_voigt_line`, `equivalent_width`.
+  - Corrections: `barycentric_correction`, `doppler_shift`, `air_to_vacuum`,
+    `vacuum_to_air`.
+  - Radial velocity: `measure_radial_velocity`.
+  - Stacking: `stack_spectra`. Time series: `lomb_scargle`, `phase_fold`.
+  - Catalogues: `simbad_query` (with `[catalogs]`).
+  - Visualisation: `plot_spectrum_plotly`, `plot_overlay_plotly` (with `[viz]`).
+  - Export: `export_csv`, `export_fits`, `export_votable`.
+- Shared Gaussian/Lorentzian/Voigt profile-fitting core (`lines/_profiles.py`).
+- VOTable I/O (`read_votable_spectrum`, `write_votable_spectrum`).
+- Spectral line catalogues: Balmer, telluric, nebular, aurorae.
+- YAML presets with loader (`snr_check`, `balmer_quick`).
+- `spectro` CLI: `list`, `describe`, `run`, `pipeline`, `presets`, `info`.
+- `spectro_mcp` MCP server auto-exposing the catalogue as tools (`[mcp]` extra).
+
+[Unreleased]: https://github.com/matthieulel/spectro-kernel/compare/v0.1.0...HEAD
+[0.1.0]: https://github.com/matthieulel/spectro-kernel/releases/tag/v0.1.0

spectro_kernel-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 spectro-kernel contributors
+
+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.