Migration to juliacall#1721
Draft
PaulJonasJost wants to merge 15 commits into
Draft
Conversation
* try to clarify `hessian_update` * typo * review; clarify docstring
Bumps [julia-actions/setup-julia](https://github.com/julia-actions/setup-julia) from 2 to 3. - [Release notes](https://github.com/julia-actions/setup-julia/releases) - [Commits](julia-actions/setup-julia@v2...v3) --- updated-dependencies: - dependency-name: julia-actions/setup-julia dependency-version: '3' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [codecov/codecov-action](https://github.com/codecov/codecov-action) from 5 to 6. - [Release notes](https://github.com/codecov/codecov-action/releases) - [Changelog](https://github.com/codecov/codecov-action/blob/main/CHANGELOG.md) - [Commits](codecov/codecov-action@v5...v6) --- updated-dependencies: - dependency-name: codecov/codecov-action dependency-version: '6' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* add low level 2d profile plot function and 2d profile function * Polish 2D profile visualization - Fix colorbar layout using constrained_layout (no more overlap) - Always extend axes to parameter bounds; remove show_bounds parameter - Auto-generate scale-aware axis labels from problem.x_scales - Accept optional x_labels override to match 1D profiles API - Pairplot-style label cleanup: parameter names on edges only, ratio label kept on off-edge diagonal subplots - Reset MaxNLocator on diagonal subplots to avoid dense tick marks - Add smoke tests for visualize_2d_profile and profile_lowlevel_2d - Add 2D profile cell to getting_started.ipynb * Add shared colorbar range and color customization to 2D profile plots - Compute global vmin/vmax across all off-diagonal panels so the colorbar accurately represents all subplots - Add profile_color parameter for diagonal 1D profile line color - Add vmin/vmax parameters to profile_lowlevel_2d for external control * Paul&Daniel review, polish plot Addresses reviewer feedback on #1652 and polishes the 2D profile grid so every subplot is legible on its own. - Self-describing subplots: every panel now carries its own axis label and tick labels instead of relying on grid position. Diagonal panels show "Log-posterior ratio" / "Objective value"; off-diagonals show both parameter labels. - Off-diagonal scatter points are sorted by color so dense clusters reveal actual ratio structure rather than reading as uniformly dark. Marker edges removed for the same reason. - New `_add_bound_lines_1d` / `_add_bound_lines_2d` helpers draw dashed lower/upper bound lines on every panel as a cross-plot anchor. - New `_add_panel_legend` helper adds compact legends on the top-left diagonal and the first 2D off-diagonal, explaining the glyphs. - New `label_fontsize` parameter; labels are rendered bold with a larger default, tick labels follow two points smaller, colorbar label matches. Tightened gridspec so y-labels hug their own axis rather than drifting between columns. - Tests cover the new styling (fontsize, spines hidden, legend presence on the expected panels, dashed bound line count). * small update of margins * small fix of layout --------- Co-authored-by: Doresic <domagoj.doresic@gmail.com> Co-authored-by: Doresic <85789271+Doresic@users.noreply.github.com>
* Fix remaining AMICI 1.0 calls and extend tests This updates the remaining AMICI calls in the model-fit visualization code (get_timepoints, set_sensitivity_order, set_timepoints, get_state_names) and aligns related docstrings and error messages with the same naming. This also expands visualization test coverage by: directly covering _get_simulation_rdatas(..., simulation_timepoints=None) to exercise default AMICI timepoint handling extending time_trajectory_model(...) coverage to include hierarchical problems and named-state selection * Daniel review changes * fix setting x_guesses in test
* introduce _style.py and rename par_indices - Add pypesto/visualize/_style.py with shared helpers get_ax and process_deprecated_kwarg. The latter is the canonical pattern for the visualization signature harmonization that follows. - Drop the module-level cmap global in sampling.py; inline at the single use site. - Rename par_indices → parameter_indices in sampling_parameter_traces, sampling_1d_marginals, and the internal get_data_to_plot. Public functions accept par_indices as a deprecated alias (DeprecationWarning). * harmonize type hints across public visualize functions - ax: canonicalize to `matplotlib.axes.Axes | None = None` everywhere (was `plt.Axes | None`, `matplotlib.axes.Axes` without `| None`, or untyped `ax=None` in profiles.py) - size: fix `tuple[float]` → `tuple[float, float]` in optimizer_convergence.py and waterfall_lowlevel; bare `tuple` → `tuple[float, float]` in optimizer_history_lowlevel - title/suptitle: `str = None` → `str | None = None` in sampling functions - Add `import matplotlib.axes` to waterfall.py, optimizer_convergence.py, optimizer_history.py, profiles.py * harmonize single-panel ax handling - switch remaining single-panel plotters and lowlevel helpers to the shared `get_ax` helper - remove ad hoc pyplot-based axes creation in favor of the shared path - ensure plots still fully configure passed-in axes, including labels and other axis-level metadata - make `projection_scatter_umap_original` explicitly accept and return an axes object - preserve backward compatibility for `optimization_run_properties_one_plot` by keeping positional argument order intact - keep tests broad and figure-focused, with targeted checks only where behavior actually changed * harmonize multi-panel axes grid-return contract - `get_axes_array(axes, nrows, ncols, size)` — create or normalize a 2-D axes grid; raises on shape mismatch - `plot_diagonal_marginal(ax, values, diag_kind)` — KDE/hist marginal extracted from `optimization_scatter` so `sampling_scatter` can share it Functions updated: `optimization_scatter`, `sampling_scatter`, `sampling_1d_marginals`, `sampling_prediction_trajectories`, `visualize_estimated_observable_mapping`, `plot_linear_observable_mappings_from_pypesto_result`, `plot_splines_from_pypesto_result`, `plot_splines_from_inner_result`, `ensemble_scatter_lowlevel`, `projection_scatter_umap_original`, `optimization_run_properties_one_plot`. Also removes all `sns.set(style="ticks")` calls from `sampling.py` — these were the root cause of global matplotlib style mutations that made plots look inconsistent depending on call order. * rename alpha to confidence_levels `sampling_parameter_cis` accepted `alpha: Sequence[int]` (e.g. `[95]`) as a credibility level in integer percentages. This was mis-named (not matplotlib transparency) and used inconsistent units compared to every other CI parameter in the codebase (all others use 0–1 fractions). * add confidence_level convenience par to profiles `profiles` accepted only `ratio_min` (a likelihood-ratio threshold) to filter profile points. Users wanting to cut off at a standard confidence level had to call `chi2_quantile_to_ratio` manually first. New additive kwarg `confidence_level: float | None = None` — pass e.g. `confidence_level=0.95` and it is converted to the equivalent `ratio_min` via `chi2_quantile_to_ratio`. The two parameters are mutually exclusive; passing both raises `ValueError`. `ratio_min` is unchanged and keeps its existing meaning and default. * add return types, remove deprecation test * Small update of type * small fix for docs test * Daniel&Paul&Vincent review adjustments - move shared visualize helpers from _style.py into misc.py - add hide_unused_axes and reuse it across plotting functions that trim unused subplot panels - switch internal figure creation to constrained layout and stop forcing tight_layout afterwards - keep deprecated-kwarg handling with _UNSET-based detection of omitted vs explicit None - update observable mapping axes preparation to use the shared helpers - move helper tests from test_style.py to test_misc.py - move close_fig to test/conftest.py and ensure figure cleanup also happens on test failure * small additional change * Document deprecated sampling plot arguments
* Fix PyMC posterior extraction for newer ArviZ * Fix PyMC outputs with newer ArviZ * add comment to change when bumped * Bump arviz>=0.21 and drop old code * Fix: drop incorrect .dataset access * require arviz 1.1.0 -- that's the correct one * Support PyMC posterior extraction across ArviZ versions - Use Python-version-specific ArviZ requirements so Python 3.11 installs ArviZ 0.x while Python 3.12+ uses ArviZ 1.1+ - Keep PyMC posterior extraction on the common posterior interface shared by old and new ArviZ result objects * some final cleanups * add TODOs
* Improve profiling step-size defaults and step-size robustness * Fix extrapolation trust-region not using resolved * Fix small bug We were changing the step size of profiling parameter to 0 by accident. * Improve profile step-size resolution - explicit absolute/relative profile step-size options - deprecated old absolute step-size option names - per-parameter step-size family resolution from `ub - lb` - finite-bound validation before profiling - single upfront step-size resolution before profile walking - resolved step sizes passed through profiling tasks/proposals - lightweight many-steps precheck warning/error mode - focused profile tests for options, resolution, precheck * Change relative defaults
* Changed Codeowners: Vincent added to sampling, Domagoj replacing Stephan in visualization * Changed Maintainer from Vincent to Moritz
|
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #1721 +/- ##
==========================================
+ Coverage 83.17% 83.58% +0.40%
==========================================
Files 164 164
Lines 14345 14766 +421
==========================================
+ Hits 11932 12342 +410
- Misses 2413 2424 +11 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
As tests for 3.14 are failing, trying to migrate once to juliacall and see whether it fixes it.