Pypesto

-------------------

- **Breaking Changes**
- **PETab select**: There are some deprecated features that will show up as warnings. In addition:

- The plotting methods ignore some arguments. You will need to reimplement these with the newer approach, which uses
plotting methods from the PEtab Select library instead. See the model selection notebook for examples.

- All objects containing multiple models (e.g., dictionaries or lists) are now replaced by `petab_select.Models`,
which supports dictionary and list methods.

To convert your old list of models:

python
petab_select.Models(list_of_Model)

- General
- Exclude nlopt==2.9.0 from setup (1519)
- Improve CI (1521, 1523, 1532, 1536, 1508, 1544, 1531)
- Update references/documentation (1506, 1491, 1516, 1543)
- **Docker Image** (1083, 1538)
- Hierarchical
- Fix no error if inner observable parameter in noise formula & viceversa (1504)
- Remove inner datas from relative calculator (1505)
- Fix not scaling inner pars when applying to rdatas (1534)
- Optimization
- ESSOptimizer: Fix priority for local search startpoints (1503)
- Fix NLoptOptimizer.__repr__ (1518)
- Improve exception-handling in SacessOptimizer (1517)
- Fix ESSOptimizer min of empty sequence (1510)
- Don't modify sys.path for amici model imports (1522)
- Set OptimizerResult.optimizer in Optimizer.minimize (1525)
- SacessOptimizer: More efficient saving of intermediate results (1529)
- Objective
- AmiciObjective/PEtab import: Fix plist (1493)
- PEtab: Fix warning from fill_in_parameters with fixed parameters (1509)
- Amici: Fix handling of PEtab fixed parameters (1514)
- Fix get_parameter_prior_dict docstring (1537)
- Select
- Support user-provided calibration results (1338)
- Problem-specific minimize method for SaCeSS (1339)
- Update for the latest PEtab Select version; see example notebook or the PEtab Select repo (1530)
- Storage
- Enable writing Optimize(r)Result directly in Writer (1528)
- Update parameter scale storage (1542
- Visualize
- Fix flatten of observable mapping with one observable (1515)

-------------------

- **Breaking Changes**
- Remove Aesara support (1453, 1455)
- General
- CI improvements (1436, 1437, 1438, 1439, 1440, 1443, 1473, 1484, 1486, 1490, 1485)
- Update references/documentation (1404, 1456, 1474, 1479, 1483, 1470, 1498)
- Profile
- Improve Profiling Code (1447)
- Visualize
- allow log and/or linear scale for visualization (1435)
- More informative error message for start indices. (1472)
- Optimization
- SacessOptimizer: Fix acceptance threshold for objective improvement (1457)
- SacessOptimizer: expose more hyperparameters + minor fixes (1459, 1476)
- SacessOptimizer, ESSOptimizer: Bound-normalize parameters for proximity check (1462)
- ESSOptimizer: Fix bug in recombination and go-beyond (1477, 1480)
- Objective
- FD-objective correctly working with fixed parameters (1446)
- Petab Importer reforge (1442, 1502)
- Use cloudpickle for serializing NegLogParameterPriors (1467)
- Update PEtab.jl integration to match version 3.X (1489)
- Sampling
- Bayes Factor Tutorial (1444)
- Ensemble
- Added HPD calculation to ensemble (1431)

-------------------

* General
* Notebook on history usage and comparison of multiple results. (1389)
* GHA/test improvements (1423, 1408, 1430)
* Numpy 2.0 compatibility (1420, 1433)
* PEtab
* Issue a warning if a fixed parameter has a prior defined (1413)
* Update to libpetab 0.4.0 (1422)
* Optimize
* Added a Result object with lazy hdf5 loading (1421)
* RoadRunner
* Roadrunner handling of petab issue 0019 (1419)
* Disentangle amici and roadrunner (1429)
* Amici
* Require `return_dict` in `ObjectiveBase.call_unprocessed` (fixes AMICI posterior RData) (1424)
* Hierarchical
* Visualize: visualization of estimated observable mapping (1409)
* Hierarchical: avoid recomputing inner parameters if simulation failed (1426)
* Visualization
* Fixing Aggregated Objective Visualisations (1411)

-------------------

* **New Feature**: Variational inference with PyMC (1306)
* PEtab
* Import of petab independent of amici (1355)
* Problem
* Added option to sample startpoints for a problem, from the problem directly. (1364)
* More detailed defaults for problem.get_full_vector (1393)
* Save pypesto and python version to the problem. (1382)
* Objective
* Fix calling priors in sampling with fixed parameters (1378)
* Fix JaxObjective (1400)
* Optimize
* ESS optimizers: suppress divide-by-zero warnings; report n_eval (1380)
* SacessOptimizer: collect worker stats (1381)
* Add load method to Hdf5AmiciHistory (1370)
* Hierarchical
* Relative: fix log of zero for default 0 sigma values (1377)
* Sample
* Fix pypesto.sample.geweke_test.spectrum for nfft<=3 (1388)
* Visualize
* Handle correlation plot with nans (1365)
* General
* Remove scipy requirement from pypesto[pymc] (1376)
* Require and test python >=3.10 according to NEP 29 (1379)
* Fix various warnings (1384)
* Small changes to GHA actions and tests (1386, 1387, 1402, 1385)
* Improve Documentation (1394, 1391, 1399, 1292, 1390)

..........

-------------------

* General
* Include pymc in the documentation. (1305)
* Ruff Codechecks (1307)
* Support RoadRunner as simulator for PEtab problems (1336, 1347, 1348, 1363)
* Hierarchical
* Semiquant: Fix spline knot initialization (1313, 1323)
* Semiquant: Add spline knots to the optimization result (1314)
* Semiquant: fix inner opt tolerance (1330)
* Relative: Fix return of relative calculator if sim fails (1315)
* Relative: Hierarchical optimization: fix unnecessary simulation (1327)
* Relative: Fix return of inner parameters on objective call (1333)
* Optimize
* Support ipopt with gradient approximation (1310)
* Deprecate CmaesOptimizer in favor of CmaOptimizer (1311)
* ESSOptimizer: Respect local_n2 in case of failed initial local search (1328)
* Remove CESSOptimizer (1320)
* SacessOptimizer: use 'spawn' start method for multiprocessing (1353)
* PEtab
* Fix unwanted amici model recompilation in PEtab importer (1319)
* Sample
* Adding Thermodynamic Integration (1326, 1361)
* Dynesty warnings added (1324)
* Dynesty: method to save raw results (1331)
* Ensembles
* Ensembles: don't expect OptimizerResult.id to be convertible to `int` (1351)
* Misc
* Updated Code to match dependency updates (1316, 1344, 1346, 1345)
* Ignore code formatting in git blame (1317)
* Updated deployment method (1341, 1371, 1373)
* add pyupgrade to codechecks (1352)
* Temporarily require scipy<1.13.0 for pypesto[pymc] (1360)