Skip to content

v0.4.0

Choose a tag to compare

@benjaminpope benjaminpope released this 21 Oct 22:38
· 1146 commits to main since this release
8c51dc5

What's Changed

  • Development carbon page, ledger and README badge by @benjaminpope in #253
  • Add Claude PR review workflow by @benjaminpope in #254
  • Design note: one argument order (model before data) for 0.4 by @benjaminpope in #255
  • Add the private _deprecate shim for the 0.4 argument order by @benjaminpope in #256
  • Orbits fitted to visibilities across epochs: design and stage 1 (Epochs, OrbitalBinary) by @benjaminpope in #257
  • Add virgil.metrics: image-recovery scores for reconstructions by @benjaminpope in #258
  • Add virgil.ensemble: PYRA/MYTHRA-style reconstruction ensembles by @benjaminpope in #259
  • AGENTS.md: relevant tests locally, zensical docs, direct venv python by @benjaminpope in #260
  • Model before data in the grid, limit and detection tools (argument order PR 2) by @benjaminpope in #261
  • Model before data in the likelihoods; one name per argument (argument order PR 3) by @benjaminpope in #263
  • Visibility orbits stage 2: rank_orbits, start from a positions fit, per-chain starts by @benjaminpope in #262
  • Carry argument order PR 3 (#263) to main by @benjaminpope in #266
  • CI: stop the py3.11 lowest job running out of memory by @benjaminpope in #264
  • OIData.select by wavelength windows and observables; audit of the spectro-interferometry plan by @benjaminpope in #265
  • Design note: plate-scale nuisances can absorb inter-instrument motion by @benjaminpope in #267
  • ensemble: judge the running mean on the members' own grids by @benjaminpope in #272
  • Argument order PR 4: last call sites and pytest filters (stacked on #266) by @benjaminpope in #269
  • CI: drop the legacy mkdocs workflow; lint script uses the venv python by @benjaminpope in #271
  • metrics: rms_convolved without pixel scale; clarify Lawson normalisation by @benjaminpope in #274
  • Orbit tutorial: executed outputs for visibility orbits stage 2 by @benjaminpope in #273
  • CI: two xdist workers on the lowest-deps job by @benjaminpope in #275
  • mwe_sparse_imaging: match the summary to the latest run by @benjaminpope in #276
  • Tests: clear JAX caches after each module (lowest-deps CI OOM) by @benjaminpope in #277
  • Faster CI: lighter test_epochs starts, factorized TruncatedCone render by @benjaminpope in #278
  • Design note: automatic binary orbits from multi-epoch interferometry by @benjaminpope in #279
  • PR B0: orbit benchmark harness (benchmarks/orbits) by @benjaminpope in #280
  • Orbits: noise-marginalised gap and scale-aware ranking (#268) by @benjaminpope in #281
  • Avoid zero-size gain arrays that crash jax.pmap (parallel chains) by @benjaminpope in #282
  • Fix main CI: NaN-robust ensemble selection, macOS-safe memory sampler by @benjaminpope in #283
  • Orbit tutorial: outputs re-executed on main (0 divergences) by @benjaminpope in #284
  • Orbit tutorial: outputs from rerun at 33e1390 by @benjaminpope in #285
  • Add OzSTAR sync gate and stale-notebook-output check by @benjaminpope in #286
  • Design: orbit-fitting review (published practice vs virgil's plan) by @benjaminpope in #287
  • Drop unused limits.chi2ppf by @benjaminpope in #293
  • epochs.marginal_loglike: adaptive ln s integral at a steep s_max bound (F18) by @benjaminpope in #295
  • limits.chi2ppf: work for any df without tensorflow_probability (F17) by @benjaminpope in #291
  • Fit budgets and guards; phase-coherent period grid (orbit review PR B) by @benjaminpope in #288
  • epoch_positions: peak catalogue, refine the top grid peaks before committing (PR C) by @benjaminpope in #289
  • Orbit search design (§14 revision) by @benjaminpope in #292
  • read_oifits: place a T3 leg with no visibility row at the T3 coordinates by @benjaminpope in #300
  • TruncatedCone: azimuthal modulation of the walls (leading-edge brightening) by @benjaminpope in #303
  • ensemble: keep weights below the L-curve corner; widen weight ranges; fields 1–4× by @benjaminpope in #304
  • SVI: virgil.svi.variational, flow guides started from a fit by @benjaminpope in #302
  • Orbit search D1: shared-nuisance scorer (score_orbits) by @benjaminpope in #298
  • virgil.pipeline PR 1: core, BinaryPipeline, docs and tutorial sections by @benjaminpope in #305
  • pipeline: capture stage warnings as checks, quieten quicklook, clean tutorial sections by @benjaminpope in #308
  • One import convention: jax.numpy as np, numpy as onp (#306 s1) by @benjaminpope in #312
  • Use model/data in public signatures (#306 s2) by @benjaminpope in #313
  • Oxford -ize spelling in names, with deprecated -ise keywords (#306 s3) by @benjaminpope in #314
  • Warnings, precision and small inconsistencies (#306 s4, s5, s8) by @benjaminpope in #315
  • Dead code and trivial dedups (part of #310) by @benjaminpope in #311
  • Drifted copies unified, with regression tests (part of #310) by @benjaminpope in #316
  • Render dev carbon badge as an image by @benjaminpope in #318
  • StarPipeline: stellar diameter and limb darkening by @benjaminpope in #307
  • Period aliases of undersampled orbits (fit_orbit_aliases) by @benjaminpope in #319
  • aliases: robust to singular Hessians, per-band checkpoints, flat table scales by @benjaminpope in #321
  • numpyro_model: sample each prior in its flat coordinate, as fit does by @benjaminpope in #301
  • Import zodiax as zdx by @benjaminpope in #322
  • Pre-0.4.0 review: StarPipeline model choice, pipeline and alias robustness, 0.3 regressions, de-spaghettification by @DePrinsT in #323
  • Fix the slow-test failures on main; faster, steadier CI by @benjaminpope in #325
  • Alias fit: sign-free seeds for near-equal binaries, Newton polish by @benjaminpope in #324
  • L-BFGS restarts: fix the macOS-only error_scale="fit" test failure by @benjaminpope in #327
  • Docs: comparison with other packages by @benjaminpope in #326
  • Evidence helpers take several field Images (path=None or a tuple) by @benjaminpope in #330
  • aliases: UD primary, per-epoch V² amplitude, flux flip, init warm start by @benjaminpope in #331
  • Provenance: references file, ADS fetch script, docstring citations by @benjaminpope in #332
  • contributors: indent nested bullets by 4 spaces so they render nested by @benjaminpope in #333
  • docs: remove 'Using these docs' section from index by @benjaminpope in #334
  • P1: PolarDisk engine (M=0) by @benjaminpope in #335
  • PolarDisk P2: height, linear harmonics, TruncatedCone on the engine by @benjaminpope in #336
  • CI: review-response label for unanswered Claude review points by @benjaminpope in #341
  • Design note: composite scenes (identifiability toy) by @benjaminpope in #342
  • Broadcast time-dependent models over samples; batch orbit candidates by default by @benjaminpope in #339
  • spectra.Scaled, chromatic companion-search tests and tutorial by @benjaminpope in #346
  • GN1: dense damped-Newton kernel (_gauss_newton.py) and the Gauss–Newton plan by @benjaminpope in #349
  • design note: FlaredDisk* stays; execution log by @benjaminpope in #348
  • PolarDisk P3: positive Harmonics, VonMises, virgil.harmonics by @benjaminpope in #338
  • P6: PolarField for Image; FlaredDisk vs PolarDisk comparison by @benjaminpope in #340
  • PolarDisk P4: RadialGP, PolarKernel, spectral_index, freeze_geometry by @benjaminpope in #343
  • PolarDisk P5: tutorial, a planet in a gap between inclined rings by @benjaminpope in #347
  • CI: fix review-response label workflow (invalid trigger) by @benjaminpope in #353
  • PolarDisk P4+P5: RadialGP, PolarKernel, freeze_geometry, tutorial (stacked merge landed on polar-p3) by @benjaminpope in #355
  • GN2: curvature rows, V19 mass with fitted error scales, size and PSD guards by @benjaminpope in #352
  • GN4: alias optimizer and per-epoch refinement on the damped-Newton kernel by @benjaminpope in #350
  • Opt-in bandwidth smearing (EFF_BAND, boxcar channels) by @benjaminpope in #345
  • CI: tiered matrix, install from uv.lock, per-test budget, nightly report by @benjaminpope in #354
  • Fix spurious failing review-response label check by @benjaminpope in #358
  • PolarField: share RadialGP/PolarKernel kernels and tie geometry to the disk by @benjaminpope in #357
  • CI: raise the per-test budget to 30 s by @benjaminpope in #363
  • Orbit 1a: AliasResult load, scene() and residuals() by @benjaminpope in #359
  • Upgrade uv.lock by @github-actions[bot] in #360
  • Size the image field from the data; resolve from unflagged samples only by @benjaminpope in #361
  • Fix contributors nesting; drop duplicate docs list from README and index by @benjaminpope in #364
  • Design: DiskPipeline (provisional), with open questions by @benjaminpope in #365
  • Docs: PolarDisk reference page, streamlined tutorial, honest GP length scale by @benjaminpope in #356
  • Orbit 1b: night flags with reasons and UD-aware seeds (to main) by @benjaminpope in #366
  • Orbit 1c: automatic engine defaults and a data-driven period window by @benjaminpope in #368
  • PolarDisk tutorial: five-cell prose, folded code, reference folded into API by @benjaminpope in #367
  • imaging: per-observable chi2, short-baseline residual, residual image (ImagePipeline A1) by @benjaminpope in #370
  • ensemble: GP members (evidence_grid, gp_members) and a combine that takes them by @benjaminpope in #371
  • Orbit 1d: tempered evidence for correlated spectral channels by @benjaminpope in #372
  • Process rules to slow code growth: AGENTS.md, CONTRIBUTING, size comment by @benjaminpope in #377
  • Stability tiers for 0.4.0: provisional features by @benjaminpope in #373
  • Correctness bundle from the 0.4.0 release review by @benjaminpope in #375
  • Docstring diet part 1: move derivations to docs, cut redundancy by @benjaminpope in #379
  • Settle core public names before 0.4.0 by @benjaminpope in #376
  • Fix x64 night-flag fixture (R1); run slow tests on orbit PRs (CI-02) by @benjaminpope in #374
  • Compile the LM kernel once; one dense-matrix limit for fitting and svi by @benjaminpope in #381
  • Orbit 1d follow-up: direct effective-channel estimator and calibration test by @benjaminpope in #380
  • Streamline CI: main = gate + slow, x64 numerical only, readable workflows by @benjaminpope in #384
  • Sort once in the group builders; trace the prior widths of extra observables by @benjaminpope in #383
  • design/: orbit plan, status tables, archive, 0.4.0 release review by @benjaminpope in #386
  • Docs for 0.4.0: quickstart, Which orbit tool?, extras notes, metadata by @benjaminpope in #385
  • Consistent names for the MultipoleDisk family (renames PolarDisk) by @benjaminpope in #378
  • Docstring diet part 2a: detection, ensemble, grid_fit, inference, spectra, svi by @benjaminpope in #387
  • Perf: evidence SVD on device (#328), chunked dirty_image, jitted metrics.align by @benjaminpope in #382
  • Release gate: publish needs a green full-tier run; red-main issue; orbits at the floors; checklist by @benjaminpope in #388
  • Orbit 0.4.0 performance by @benjaminpope in #389
  • Upgrade uv.lock by @github-actions[bot] in #392
  • Orbit 0.4.0 API fixes by @benjaminpope in #390
  • Remove virgil.spectra.Tabulated (use Nodes) by @benjaminpope in #394
  • Guard plot_chainconsumer_diagnostics against slow calls (could not reproduce 7,971 s) by @benjaminpope in #393
  • Fix main: slow-tier tempering calibration timeout after #389 by @benjaminpope in #400
  • harmonix tutorial: grid-scan the spot before the L-BFGS fit by @benjaminpope in #404
  • Add DiskPipeline (provisional): rings and GP disk models by @benjaminpope in #369
  • Slim data, observables and plotting modules (0.4.0 size review) by @benjaminpope in #396
  • Simplify virgil.models and fields: shared helpers, notes page, tidier tests by @benjaminpope in #397
  • Simplify imaging, gains, fitting, ensemble and friends before 0.4.0 by @benjaminpope in #398
  • Orbit module cleanup for 0.4.0 by @benjaminpope in #403
  • Pipelines: start NUTS inside the prior when the best fit is at a bound by @benjaminpope in #402
  • Test the frozen refit's cost deterministically (nightly macOS flake) by @benjaminpope in #405
  • tests: keep PR CI under the per-test budget (DiskPipeline, ensemble) by @benjaminpope in #407
  • Docs examples: fix failing pages and stale outputs from the GPU gate by @benjaminpope in #401
  • tests: move the no-companion pipeline test to the slow tier by @benjaminpope in #410
  • Upgrade uv.lock by @github-actions[bot] in #409
  • DiskPipeline rings: a ring split in two does not count as a ring by @benjaminpope in #406
  • pipelines: start NUTS and the ring fits where the gradient is finite (docs/pipeline.md) by @benjaminpope in #415
  • docs: Pipelines tutorial notebook by @benjaminpope in #414
  • docs: condense prose across the docs; fold import and plotting boilerplate by @benjaminpope in #413
  • tests: move the compile-bound tests over the 30 s CI budget to the slow tier by @benjaminpope in #412
  • DiskPipeline: defaults from imaging nine post-AGB discs by @benjaminpope in #416
  • Results-neutral perf: cache DiskPipeline _builder, vmap Epochs.snapshots by @benjaminpope in #418
  • Remove dead code and merge duplicates (0.4.0 batch B) by @benjaminpope in #419
  • limb_darkening: fix understated chi2/N (N = independent data) by @benjaminpope in #417
  • starting_image: pad_to_mas pads beyond λ/B_min; warn when largest_mas is capped by @benjaminpope in #421
  • Design: instrument noise models in pipelines (0.4.1) by @benjaminpope in #420
  • Design: generic instrument noise contract for pipelines (0.4.1) by @benjaminpope in #423
  • Keep the best point of a refinement that does not converge by @benjaminpope in #427
  • Seed alias bands from short arcs: night velocities narrow the cycle counts by @benjaminpope in #425
  • Upgrade uv.lock by @github-actions[bot] in #424
  • Automatic nufftax NUFFT for large images with many uv points by @benjaminpope in #426
  • Fix short-arc seeding: main's ValueError and displaced position seeds by @benjaminpope in #430
  • CI: shard the py3.13 and dependency-floor jobs by @benjaminpope in #434
  • Add virgil.radio: complex visibilities and calibration nuisances by @benjaminpope in #432
  • Fit summary: forward-only chi-squared, L-BFGS gradient norm from its state by @benjaminpope in #433
  • Docs: fix multipole_disk_rings DiskPipeline block (ell_halffringes below the floor) by @benjaminpope in #431
  • Pipelines: parsimony and robustness, phase A (disk, star, binary, _common) by @benjaminpope in #428
  • OrbitPipeline: blind multi-epoch binary-orbit alias search (orbit pipeline 1e) by @benjaminpope in #422
  • Temper the short-arc sign gate and the cut's errors by f (Gl 229 gate) by @benjaminpope in #435
  • radio: document and test that term data stay jit arguments by @benjaminpope in #438
  • R1: score data with gains or closure offsets on the marginal surface by @benjaminpope in #437
  • Test the gains quadrature at 16 and 32 nodes by @benjaminpope in #442
  • LM never accepts a step that raises the loss (fixes a noise-free runaway) by @benjaminpope in #443
  • radio: sigma_scale and dirty/residual image diagnostics by @benjaminpope in #441
  • Add virgil.crossval: generic k-fold cross-validation helper by @benjaminpope in #440
  • Image: optional Gaussian smoothing, regularisers on the smoothed pixels by @benjaminpope in #444
  • Pipeline parsimony and robustness, phase B by @benjaminpope in #446
  • Design: orbit fits with calibrated data and virgil-vlti noise models by @benjaminpope in #436
  • Docs: orbit tutorial leads with OrbitPipeline (orbit plan 1f) by @benjaminpope in #447
  • Upgrade uv.lock by @github-actions[bot] in #449
  • Epoch surface with gains: memory-budgeted batches and coarse-then-exact scoring by @benjaminpope in #453
  • Closure offsets for three-telescope data (triangle and modes groups) by @benjaminpope in #452
  • Matrix Fourier transform for rendered disks on DISCO uv lattices by @benjaminpope in #450
  • FlaredDisk: drop arctan2/cos from the render by @benjaminpope in #454
  • Fix origin gradients on JAX 0.12 and the CLEAN stall test; explain the cold-cache alias timeout by @benjaminpope in #455
  • Night flags inform instead of excluding (outlier mixture, per-mode chi2, sidelobe flag) by @benjaminpope in #451
  • BinaryPipeline: start the fit inside the position priors (fixes the docs/pipelines.md TypeError) by @benjaminpope in #456
  • ImagingPipeline: pipeline skeleton (PR B) by @benjaminpope in #457
  • OrbitPipeline: search on the plain likelihood, polish with the outlier mixture (Gl 229 gate regression from #451) by @benjaminpope in #459
  • CI: move two over-budget tests to the slow set by @benjaminpope in #465
  • Orbit pipeline: simplify pass by @benjaminpope in #461
  • Pipelines: simplify binary.py and star.py by @benjaminpope in #464
  • ImagingPipeline C: ensemble stage by @benjaminpope in #460
  • Docs: requote multipole_disk_rings block 7 (A100 run at 5ca0bc9) by @benjaminpope in #458
  • Simplify the shared pipeline core by @benjaminpope in #463
  • Pipeline disk/image: simplify pass by @benjaminpope in #462
  • Upgrade uv.lock by @github-actions[bot] in #467
  • 0.4.0 review fixes: NaN CV scores, non-blocking night flags, validation, docs by @benjaminpope in #468
  • 0.4.0 pre-tag docs fix: pipelines page recovers the true companion and reads a real OIFITS by @benjaminpope in #469
  • Release 0.4.0: version, uv.lock and restructured changelog by @benjaminpope in #399

New Contributors

  • @github-actions[bot] made their first contribution in #360

Full Changelog: v0.3.0...v0.4.0