Skip to content

v0.13.0

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 19 May 15:37
· 50 commits to main since this release
v0.13.0
2624b75

πŸš€ Features

  • (tecan) Improve loggings and warnings
  • (tecan) Mcmc delta pH as TruncatedNormal
  • (tecan) Multi mcmc with scheme return not only trace
  • (tecan) Drop emcee fit
  • Add logging_config
  • (tecan) Pymc_odr
  • (tecan) Pymc_odr use exact xc not x_true
  • (tecan) Add Reweighted LS and glob IRLS
  • (tecan) Add ye_mag for each label
  • (tecan) Add result_multi_mcmc (#1000)
  • New pymc_multi and outlier2 fit (#1025)
  • Enhance CLI error handling
  • (fitting) Add residuals analysis module and bayes robustness fixes
  • (benchmarks) Final fitting method comparison
  • (benchmarks) Comprehensive fitting method comparison
  • (bayes) Add fit_binding_pymc_multi_noise with learned noise model
  • (cli) Add --mcmc multi-noise option for noise-aware MCMC fitting
  • (prtecan) Export trace NetCDF and shared noise params for multi-noise MCMC
  • (prtecan) Auto-export residual CSVs alongside fit results
  • (residuals) Add diagnostic plots exported alongside fit results
  • (bayes) Add fit_binding_pymc_multi_noise_xrw with per-well pH random walk
  • (bayes) Add --nuts-sampler CLI option with GPU support and clear import errors
  • (bayes) Use per-well x_per_well for lb4 xrw .dat/.png export
  • (cli) Rationalize ppr tecan options and add --fit-method/--outlier
  • (bayes) Add --ctr-free-k hierarchical CTR replicate K
  • (cli) Add --mcmc-samples option and outlier/sampler comparison scripts
  • (scripts) Test outlier zscore at 3.0, 2.5, and 2.0 thresholds
  • (scripts) Extend compare to L1/L3; document Ξ± estimation analysis
  • (prtecan) Add --noise-alpha for proportional noise correction
  • Add mcmc_multi_noise_xrw_free_ctr mode and noise calibration helpers
  • (fitting) Add detect_bad_wells() in diagnostics module
  • (diagnostics) Add detect_bad_wells_from_dat() and ppr detect-bad-wells CLI
  • (tecan) Extend benchmark runner
  • (tecan) Add factorized Tecan registry
  • (tecan) Add real data benchmark
  • Add mcmc noise plots and detect and discard bad wells
  • (fitting) Add robust options to consolidated multi mcmc
  • (fitting) Add buffer bg_noise as sigma_read estimate
  • Add outlier masking, bad-well detection, and infer_gain support

πŸ› Bug Fixes

  • (tecan) Typing in pymc ODR fit
  • (tecan) Tutorial
  • Empty result when mcmc fails
  • (tecan) Returned FitResult ds
  • (tecan) Global fit test
  • Renovate configuration file
  • Logger for outliers in glob_reweighted
  • Yanked click 8.2.2
  • Cruft-update
  • (ci) Remove .rej files in cruft-update workflow
  • (prtecan) Switch global fit from outlier2 to Huber-loss WLS
  • (prtecan) Stream per-well PNG+data export as each fit completes
  • (prtecan) Move noise-extras export after per-well loop to restore streaming
  • (bayes) Disable progressbar for JAX vectorized chain_method
  • (plot) Sort wells within each CTR group in K{i}.png top panel
  • (bayes) Handle None stderr in weighted_stats and n_sd
  • (scripts) Capture ppr exit code via temp file, not SIGPIPE-prone pipeline
  • Typing error
  • Guard against NaN z-scores in outlier2 for zero-noise datasets
  • (bayes) Fix zip length mismatch in fit_binding_pymc_multi_noise_xrw
  • (bayes) Replace hierarchical CTR K with flat per-well priors in ctr_free_k mode
  • Lmfit covariance edge case in test
  • Make type|xdoc after diagnostic
  • (bayes,synthetic,tests) Resolve pre-commit lint issues from new CTR K init and tecan error model
  • Set optimizer=fast_compile in PYTENSOR_FLAGS to prevent nutpie kernel limit
  • Set JAX_PLATFORM_NAME=cpu to prevent GPU OOM with blackjax/numpyro
  • Accept underscores in --mcmc option (e.g. multi_noise_xrw)
  • Update plot_ppc_well variable names to match new likelihood naming conventions
  • (docs) Update notebooks and plotting.py for arviz 1.x API
  • Plot_qc_mean_vs_std after Arviz 1.x
  • Detect and discard logic

🚜 Refactor

  • (tecan) Redesign Titration fit; results as cached_property
  • (tecan) Add TitrationResults remove Plotter and print_fitting
  • (tecan) Removed verbose from cli
  • (tecan) Create dataset (any number)
  • (tecan) Dict (with label as key) for data, results, bg
  • (tecan) Simplify plot_k and dataframes export
  • (tecan) Pymc removing priors with cov
  • (tecan) Core data structure in data.py
  • (tecan) Pymc_many_scheme
  • (tecan) Fit_binding_pymc_multi
  • Logger config
  • Rename binding to fitting
  • (tecan) Move plot_fit into plotting
  • (fitting) Move out ODR
  • (fitting) Move out PyMC
  • (fitting) Rename fitting -> core
  • FitResult and scipy-stubs (#1035)
  • Improve prtecan error handling and validation
  • Improve fit_bayes test coverage
  • The logger; add tests for tecan
  • Dilution correction and removed redundant tests
  • Add benchmarks to compare fit and error models
  • Merge refactor_benchmark branch
  • Remove redundant fit; improve synthetic data
  • Ds.plot() make_dataset
  • Bayes tests and add transitive deps
  • (odr) Switch ODR backend to odrpack and add shared utils
  • (fitting) [breaking] Unify fit_binding_glob with method/reweight/remove_outliers params
  • (fitting) Remove redundant wls and iterative fit methods
  • Remove outlier2, recursive, reweighted because redundant
  • Fit-combination registry
  • (fitting) Consodidate odr and single mcmc
  • (fitting) Remove redundant functions
  • Plot_qc_mean_vs_std
  • Clean uo bayes
  • Decouple error modeling and orchestrator, fix pipeline tests

πŸ“š Documentation

  • Fix prtecan tutorial
  • Add back ReadTheDocs
  • Try fix RtD
  • Add pdf and epub to RtD
  • Try not using requirements.txt
  • Update readme to cruft template
  • Improve documentation for core fitting models
  • Devel
  • Fix make docs warnings and errors
  • Add expand prtecan docstrings
  • (todo) Record L2/L4 fit-method and MCMC comparison conclusions
  • (todo) Update fit-method tables with L1/L3 results (all 4 plates)
  • (todo) Add L1 MCMC results; buffer well contamination findings; correct xrw assessment
  • Further adjust tutorial after Arviz 1.x

🎨 Styling

  • Consistent test dependency requirements

πŸ§ͺ Testing

  • For fitting
  • Few prtecan edge cases" (#1109)
  • Remove duplicate prtecan test classes

🦾 Build

  • (deps) Bump pymc from <=5.18.2 to 6.0.0
  • (deps) Bump codecov/codecov-action from 5.0.7 to v5.5.1
  • (deps) Bump ruff from <=0.8.1 to <=0.11.10
  • (deps) Bump coverage[toml] from <=7.6.8 to <=7.8.1
  • (deps) Bump commitizen from <=4.0.0 to <=4.8.1
  • (deps) Bump numpy from <=2.1.3 to <=2.2.3
  • (pre-commit) Update hooks
  • (deps) Bump hatch from 1.13.0 to 1.14.1
  • (deps) Bump matplotlib from <=3.9.3 to 3.10.9
  • (deps) Update pydata-sphinx-theme requirement
  • Update ruff codes in pyproject
  • (deps) Bump mypy from <=1.13.0 to <=1.15.0
  • (deps) Update types-setuptools
  • (deps) Bump click from <=8.1.7 to 8.3.1
  • (deps) Bump nbsphinx from <=0.9.5 to <=0.9.7
  • (deps) Update sphinx-autodoc-typehints
  • (deps) Bump scipy from <=1.14.1 to 1.17.1
  • (deps) Bump pygments from <=2.18.0 to <=2.19.1
  • (deps) Bump pre-commit from <=4.0.1 to <=4.2.0
  • (deps) Bump pip from 24.3.1 to 25.1.1
  • Drop ruff-lsp
  • Remove isort from toml
  • Add uv to hatch
  • Lint with npm prettier
  • (deps) Bump sphinx from <=8.1.3 to <=8.2.3
  • (deps) Bump pytest from <=8.3.4 to <=8.3.5
  • (deps) Bump arviz from <=0.20.0 to 0.23.4
  • (deps) Bump pandas-stubs from <=2.2.3.241126 to <=2.2.3.250308
  • (deps) Bump lmfit from <=1.3.2 to 1.3.4
  • Add uv.lock
  • Fix python version for default hatch env
  • Adopt .venv and auto update uv.lock
  • (deps) Bump actions/setup-python from 4 to v6
  • Update renovate lint and release (#936)
  • (deps) Bump python Docker tag to v3.13 (#940)
  • (deps) Bump sympy from 1.13.3 to 1.14.0
  • (deps) Bump pandas from 2.2.3 to 3.0.3
  • (deps) Bump xlrd from 2.0.1 to 2.0.2
  • Drop py3.10
  • (deps) Bump actions/download-artifact from v4 to v7
  • (deps) Bump actions/checkout from v4 to v6
  • (deps) Bump actions/upload-pages-artifact from v3 to v4
  • (deps) Bump git-cliff from 2.10.0 to 2.10.1
  • (deps) Bump coverage from 7.10.6 to 7.10.7
  • Add cruft
  • Cruft update
  • Drop Python 3.11
  • (deps) Bump astral-sh/setup-uv from v6 to v7
  • (deps) Bump peter-evans/create-pull-request from v7 to v8
  • (deps) Bump actions/cache from v4 to v5
  • (deps) Bump python Docker tag to v3.14 (#1121)
  • Manual cruft upgrade
  • Manual cruft update
  • (deps) Bump tqdm from 4.67.1 to 4.67.3
  • Manual cruft update to autoapi docs

πŸ™ CI/CD

  • Add back pre-commit.ci to autofix uv lock with yanked pkgs
  • (fix) Pre-commit.ci autofix labl
  • Drop hatch; improve release
  • (fix) Cruft-update
  • Add uv lock --upgrade into renovate
  • Cruft manages workflows and docs/dev/lint/test deps
  • Test temporary downgrade in cookiecutter
  • Fix cruft-update
  • Fix cruft-update uv lock error
  • Manual update typos in pre-commit-config
  • Skip slow real-data benchmark in default CI
  • Cruft manual update
Assets 2
Loading