f8668e8 | Matthew Newville | 22 June 2018, 17:21:19 UTC | remove cruft | 22 June 2018, 17:21:19 UTC |
c042afd | Matthew Newville | 22 June 2018, 01:34:45 UTC | UNRELATED: allow json encoding of unicode on Py2 | 22 June 2018, 01:34:45 UTC |
049533c | Matthew Newville | 21 June 2018, 14:37:21 UTC | add test for adding parameter with expr, bounds, but no value | 21 June 2018, 14:37:21 UTC |
b8a6e86 | Matthew Newville | 21 June 2018, 14:33:16 UTC | force val=-inf if val is None earlier in init_bounds | 21 June 2018, 14:33:16 UTC |
8e0a063 | Matt Newville | 15 June 2018, 01:45:37 UTC | Merge pull request #481 from reneeotten/covariance_matrix Transform covariance matrix and refactor code for parameter uncertainties/correlations | 15 June 2018, 01:45:37 UTC |
3a57a7c | reneeotten | 10 June 2018, 23:43:34 UTC | TST: add tests for the covariance matrix To make sure that (i) the new recipe to directly transform the covariance matrix and (ii) the refactoring of calculating the uncertainties and correlations yields the same results as before. | 12 June 2018, 03:19:42 UTC |
a30ff73 | reneeotten | 12 June 2018, 01:59:49 UTC | TST: if result.errorbars is False, then .stderr is None | 12 June 2018, 02:45:39 UTC |
1b3c6de | reneeotten | 08 June 2018, 03:14:37 UTC | MAINT: refactor estimation of uncertainties/correlations * move the code into a separate function so that it can be used later on when estimating the covariance matrix for other solvers * scale the covariance matrix if scale_covar is True * calculate uvars only once if there is an expression set for any parameter | 12 June 2018, 02:45:34 UTC |
c539764 | reneeotten | 08 June 2018, 01:40:36 UTC | MAINT: use the covariance matrix from scipy.leastsq In the leastsq solver, there is no need to (re-)calculate the covariance matrix from "fjac" and "ipvt". One can directly use "cov_x" from scipy.leastsq and transform it to the external parameter space. | 12 June 2018, 02:12:01 UTC |
f3648f9 | reneeotten | 08 June 2018, 00:42:51 UTC | ENH: transform the covariance matrix directly to external parameters cov_ext = np.dot(grad.T, grad) * cov_int | 12 June 2018, 02:12:01 UTC |
9829178 | Matt Newville | 18 May 2018, 14:26:54 UTC | Merge pull request #479 from reneeotten/examples add missing Jupyter notebook examples | 18 May 2018, 14:26:54 UTC |
6eb1b07 | reneeotten | 18 May 2018, 13:57:42 UTC | BLD: add missing Jupyter notebook examples | 18 May 2018, 13:57:42 UTC |
57f04ab | Matt Newville | 18 May 2018, 10:53:52 UTC | Merge pull request #478 from reneeotten/description_pr_issues Updates to contributing and Issue/PR templates | 18 May 2018, 10:53:52 UTC |
461abdf | reneeotten | 18 May 2018, 01:41:47 UTC | DOC: update pull_request_template.md | 18 May 2018, 02:03:16 UTC |
aa05bdf | reneeotten | 18 May 2018, 01:20:46 UTC | DOC: update issue_template.md | 18 May 2018, 02:03:06 UTC |
6b4fce7 | reneeotten | 18 May 2018, 00:58:49 UTC | DOC: update contributing.md | 18 May 2018, 02:02:49 UTC |
e1f725a | Matt Newville | 16 May 2018, 15:18:13 UTC | Merge pull request #477 from reneeotten/fix_tests Fixes to testing | 16 May 2018, 15:18:13 UTC |
aebf59c | reneeotten | 15 May 2018, 18:00:14 UTC | TST: remove intermittently failing 'assert(out.residual <= ret.fun)' from test_basinhopping.py | 15 May 2018, 18:00:14 UTC |
777ab01 | reneeotten | 15 May 2018, 17:30:23 UTC | BLD: add the examples to the distribution and ignore .pytest* files | 15 May 2018, 17:33:25 UTC |
a4eefbb | reneeotten | 15 May 2018, 17:15:26 UTC | BUG: fix imports to asteval | 15 May 2018, 17:19:05 UTC |
2786527 | reneeotten | 15 May 2018, 17:09:14 UTC | TST: do not generate any matplotlib output in automatic testing | 15 May 2018, 17:19:05 UTC |
0e3abbf | Matthew Newville | 14 May 2018, 01:33:26 UTC | update THANKS | 14 May 2018, 01:33:26 UTC |
9f52c7d | Matt Newville | 14 May 2018, 01:14:00 UTC | Merge pull request #476 from reneeotten/various_updates various small changes | 14 May 2018, 01:14:00 UTC |
5e5faed | reneeotten | 13 May 2018, 01:05:15 UTC | DOC: update docstring for least_squares | 13 May 2018, 01:15:45 UTC |
bca8f80 | reneeotten | 12 May 2018, 22:06:50 UTC | DOC: update solvers in docstrings/documentation. | 13 May 2018, 01:15:45 UTC |
974649d | reneeotten | 12 May 2018, 21:55:48 UTC | TST: exclude the new trust-* methods during testing. These methods need a gradient and/or Hessian; also the other methods that need this are excluded in the "test_scalar_minimizers". | 13 May 2018, 01:15:45 UTC |
ef5b979 | reneeotten | 12 May 2018, 21:24:52 UTC | ENH: support all scipy.optimize.minimize solvers. | 13 May 2018, 01:15:45 UTC |
e45e5b4 | reneeotten | 12 May 2018, 21:38:48 UTC | DOC: update output for doc_fitting_withreport.py | 13 May 2018, 01:15:45 UTC |
b48e6cc | reneeotten | 12 May 2018, 19:40:33 UTC | DOC: set random.seed for reproducible examples in the documentation. Update the fit_report and figures in the documentation accordingly. | 13 May 2018, 01:15:45 UTC |
3e78678 | reneeotten | 09 May 2018, 01:20:27 UTC | DOC: update release notes. | 13 May 2018, 01:15:41 UTC |
5ebe5c6 | reneeotten | 09 May 2018, 01:07:29 UTC | DOC: update INSTALL. | 13 May 2018, 01:06:30 UTC |
0eafb1d | reneeotten | 09 May 2018, 01:02:37 UTC | MAINT: remove unnecessary files from repository. | 13 May 2018, 01:06:30 UTC |
0ab6eac | Matt Newville | 09 May 2018, 01:47:27 UTC | Merge pull request #475 from reneeotten/calculate_statistics add function to calculate fitting statistics | 09 May 2018, 01:47:27 UTC |
dcea5a6 | reneeotten | 04 May 2018, 14:59:52 UTC | TST: compare function values, not chisqr. When fitting to a function the methods brute, basinhopping, and ampgo will return the function value. In the tests this should be compared to result.residual and not result.chisqr | 04 May 2018, 15:20:27 UTC |
89ef15d | reneeotten | 04 May 2018, 02:04:52 UTC | ENH: add _calculate_statistics method Consolidate the calculation of fitting statistics in a function to avoid code duplication in all solvers. Still need to make sure that result.residual is calculated/set correctly before calling this method! | 04 May 2018, 15:20:27 UTC |
b7d87b9 | Matt Newville | 01 May 2018, 14:21:48 UTC | Merge pull request #466 from reneeotten/ampgo Add the AMPGO algorithm | 01 May 2018, 14:21:48 UTC |
0bf71d0 | reneeotten | 29 April 2018, 14:36:52 UTC | TST: add test for ampgo method | 01 May 2018, 02:44:51 UTC |
1626e8c | reneeotten | 29 April 2018, 20:10:02 UTC | ENH: maxfuneval is None means no limit on the number of function evaluations. The optimization will stop after the specified number of iterations ("totaliter"). | 01 May 2018, 02:27:15 UTC |
2441547 | reneeotten | 29 April 2018, 14:36:44 UTC | MAINT: function value should be a scalar not an array. When fitting a function (i.e., in the test) AMPGO returns an array instead of a scalar for the function value. | 01 May 2018, 02:27:15 UTC |
58588f8 | reneeotten | 30 April 2018, 01:10:01 UTC | MAINT: handle AbortFitException. | 01 May 2018, 02:27:14 UTC |
64ebd1f | reneeotten | 29 April 2018, 14:02:10 UTC | MAINT: use apply_bounds_transformation. Do not pass "bounds" to the underlying minimizer anymore; use the "lmfit" way of making sure that parameters stay within bounds. | 01 May 2018, 02:27:14 UTC |
b032feb | reneeotten | 09 April 2018, 21:49:13 UTC | MAINT: remove 'fmin' parameter and update description of 'glbtol'. | 01 May 2018, 02:27:14 UTC |
07c5519 | reneeotten | 05 April 2018, 22:51:33 UTC | MAINT: remove original AMPGO function. | 01 May 2018, 02:27:14 UTC |
2b3b5a5 | reneeotten | 02 April 2018, 20:39:02 UTC | STY: minor PEP8 style changes and comment updates. | 01 May 2018, 02:27:14 UTC |
205fafb | reneeotten | 02 April 2018, 20:31:38 UTC | MAINT: use the calculated numerator in inverse_tunnel function. | 01 May 2018, 02:27:14 UTC |
50cc742 | reneeotten | 02 April 2018, 01:05:29 UTC | MAINT: import numpy as np. | 01 May 2018, 02:27:14 UTC |
92424f5 | reneeotten | 01 April 2018, 20:30:20 UTC | MAINT: simplify setting of local_tol. | 01 May 2018, 02:27:14 UTC |
7a357ce | reneeotten | 31 March 2018, 20:04:42 UTC | MAINT: simplify code to handle bounds. | 01 May 2018, 02:27:14 UTC |
04ff527 | reneeotten | 31 March 2018, 17:06:45 UTC | MAINT: use "n" instead of calculating "len(x0)" again". | 01 May 2018, 02:27:14 UTC |
f7c290c | reneeotten | 31 March 2018, 02:14:36 UTC | MAINT: work on the duplicated AMPGO code. - remove main() function - add module docstring - remove OPENOPT solvers - add/update function docstrings - update "disp" statements (now boolean) - add comments to the code | 01 May 2018, 02:27:14 UTC |
fd37d22 | reneeotten | 21 March 2018, 02:03:55 UTC | ENH: add AMPGO minimization algorithm and update docs - add AMPGO code from Andrea Gavana | 01 May 2018, 02:27:14 UTC |
1a009ed | Matt Newville | 01 May 2018, 01:25:18 UTC | Merge pull request #474 from lmfit/fix_isfinite_bug simplify and fix corner-case errors testinf closeness of large integers | 01 May 2018, 01:25:18 UTC |
ca66242 | Matthew Newville | 30 April 2018, 01:38:08 UTC | simplify and fix corner-case errors testinf closeness of large integers | 30 April 2018, 01:38:08 UTC |
a6cc4d6 | Matt Newville | 29 April 2018, 11:50:33 UTC | Merge pull request #473 from reneeotten/external_uncertainties_package External uncertainties package | 29 April 2018, 11:50:33 UTC |
beca5a6 | reneeotten | 19 April 2018, 23:34:50 UTC | DOC: update documentation. | 29 April 2018, 02:47:13 UTC |
c1d8153 | reneeotten | 19 April 2018, 20:11:16 UTC | TST: fix obsolete warnings related to uncertainties | 29 April 2018, 02:45:51 UTC |
160b6c4 | reneeotten | 19 April 2018, 19:23:22 UTC | MAINT: remove fork of uncertainties package. | 29 April 2018, 02:45:51 UTC |
7ca6558 | reneeotten | 19 April 2018, 18:43:39 UTC | MAINT: don't call std_dev() anymore [obsolete]. | 29 April 2018, 02:45:51 UTC |
7c11aa7 | reneeotten | 19 April 2018, 18:45:56 UTC | MAINT: uncertainties.Variable --> uncertainties.core.Variable | 29 April 2018, 02:45:51 UTC |
cc096e6 | reneeotten | 19 April 2018, 18:35:50 UTC | MAINT: use external 'uncertainties' package. | 29 April 2018, 02:45:51 UTC |
1f22c6f | Matt Newville | 28 April 2018, 17:35:09 UTC | Merge pull request #472 from lmfit/mpl_importtest tolerate nearly any exception when trying to import matplotlib | 28 April 2018, 17:35:09 UTC |
0319860 | Matt Newville | 28 April 2018, 17:34:00 UTC | Merge pull request #468 from reneeotten/basinhopping Add support for scipy.basinhopping | 28 April 2018, 17:34:00 UTC |
fe68f5d | reneeotten | 15 April 2018, 20:42:22 UTC | TST: add test for basinhopping. | 28 April 2018, 13:41:05 UTC |
14718a8 | reneeotten | 19 April 2018, 15:17:20 UTC | MAINT: do use apply_bounds_transformation. Do not pass "bounds" to underlying local minimizer... Instead make use of "apply_bounds_transformation" so that bounds are respected even if the local minimizer does not support them. That also avoids the need of using an "accept_test". | 28 April 2018, 13:41:05 UTC |
2d3d838 | reneeotten | 16 April 2018, 18:41:43 UTC | MAINT: fix basinhopping for scipy version < 0.19 This removes the "seed" argument and prints a warning message. | 28 April 2018, 13:41:05 UTC |
5e84fd3 | reneeotten | 16 April 2018, 13:22:02 UTC | MAINT: refactor penalty_brute into penalty function. | 28 April 2018, 13:41:05 UTC |
b0a0e01 | reneeotten | 15 April 2018, 14:46:49 UTC | ENH: add basinhopping algorithm. | 28 April 2018, 13:41:05 UTC |
e73ae1d | reneeotten | 15 April 2018, 15:06:46 UTC | DOC: update filename for brute example. | 28 April 2018, 13:41:05 UTC |
ae5fc87 | Matthew Newville | 28 April 2018, 12:27:24 UTC | tolerate nearly any exception when trying to import matplotlib | 28 April 2018, 12:27:24 UTC |
5c27c1c | Matt Newville | 28 April 2018, 12:20:07 UTC | Merge pull request #471 from reneeotten/requirements update handling of installation requirements | 28 April 2018, 12:20:07 UTC |
6f64f2d | reneeotten | 25 April 2018, 18:21:46 UTC | MAINT: use 'python_requires' to specify allowed Python versions. See: https://packaging.python.org/tutorials/distributing-packages/#python-requires | 25 April 2018, 18:24:23 UTC |
a2730a0 | reneeotten | 25 April 2018, 18:14:15 UTC | MAINT: read dependencies from file. Only "install_requires" in setup.py is automatically analyzed by pip during an install. Requirement files are only used when invoking "pip install -r". Previously, the dependencies were declared but no minimum versions were given. Now all requirements and minimum versions are consolidated in "requirements.txt". See: https://packaging.python.org/discussions/install-requires-vs-requirements/#install-requires-vs-requirements-files | 25 April 2018, 18:24:23 UTC |
8662627 | Matt Newville | 16 April 2018, 02:57:54 UTC | Merge pull request #467 from reneeotten/AbortFitException Raise an AbortFitException when fit is aborted by user callback | 16 April 2018, 02:57:54 UTC |
2c39d76 | reneeotten | 14 April 2018, 15:26:27 UTC | MAINT: handle AbortFitException in least_squares. | 14 April 2018, 15:35:48 UTC |
3480442 | reneeotten | 14 April 2018, 15:08:40 UTC | MAINT: handle AbortFitException in least_squares. | 14 April 2018, 15:35:48 UTC |
6a9f62b | reneeotten | 14 April 2018, 14:13:52 UTC | MAINT: handle AbortFitException in scalar_minimize. | 14 April 2018, 15:35:48 UTC |
63e44f0 | reneeotten | 13 April 2018, 15:44:27 UTC | ENH: raise AbortFitException when fit is aborted by the user. - set some attributes before raising the exception - handle the exception in leastsq | 14 April 2018, 15:35:48 UTC |
846cd6c | Matt Newville | 01 April 2018, 22:25:16 UTC | Merge pull request #465 from reneeotten/consistent_nfev consistent nfev | 01 April 2018, 22:25:16 UTC |
feb0be6 | reneeotten | 30 March 2018, 15:30:45 UTC | TST: fix failing test; now nfev -= 2 for leastsq. | 30 March 2018, 15:30:45 UTC |
e6bcdfb | reneeotten | 30 March 2018, 15:05:48 UTC | MAINT: do not add calls to __jacobian to nfev. | 30 March 2018, 15:05:48 UTC |
5abb892 | reneeotten | 30 March 2018, 15:04:18 UTC | MAINT: correct nfev for pre/post calls to __residual. | 30 March 2018, 15:04:18 UTC |
8c4a1f1 | reneeotten | 29 March 2018, 19:01:49 UTC | DOC: fix typo in docstring | 29 March 2018, 19:02:33 UTC |
2122b92 | reneeotten | 29 March 2018, 18:57:12 UTC | MAINT: no need to calculate the residual array, use 'fun' attribute from least_squares. | 29 March 2018, 19:02:24 UTC |
cfd34d5 | reneeotten | 29 March 2018, 18:08:17 UTC | BUG: fix upper bounds in least_squares. | 29 March 2018, 18:10:37 UTC |
b44786b | Matthew Newville | 21 March 2018, 15:00:25 UTC | doc should display version without versioneer tags | 21 March 2018, 15:00:25 UTC |
ff140a7 | Matthew Newville | 21 March 2018, 00:28:31 UTC | brevity | 21 March 2018, 00:28:31 UTC |
1e90b69 | Matthew Newville | 21 March 2018, 00:19:59 UTC | installation points to 2 conda channels, as conda-forge is perenially broken | 21 March 2018, 00:19:59 UTC |
b6f5789 | Matt Newville | 20 March 2018, 15:59:27 UTC | Merge pull request #462 from reneeotten/documentation_docstrings updates to documentation and docstring | 20 March 2018, 15:59:27 UTC |
f02bc9e | reneeotten | 20 March 2018, 03:09:48 UTC | DOC: updates for new fit_report output | 20 March 2018, 12:50:03 UTC |
3d0caea | reneeotten | 13 March 2018, 02:19:41 UTC | DOC: update installation requirements. | 19 March 2018, 16:10:09 UTC |
71e6427 | reneeotten | 13 March 2018, 02:05:49 UTC | DOC: correct typo in docstring. | 19 March 2018, 16:10:09 UTC |
cd02480 | Matt Newville | 19 March 2018, 03:16:44 UTC | Merge pull request #461 from lmfit/stats_with_gformat use gformat for fit statistics | 19 March 2018, 03:16:44 UTC |
cffdab4 | Matthew Newville | 19 March 2018, 02:32:26 UTC | trim leading whitespace for reported statistics | 19 March 2018, 02:32:26 UTC |
f6cbb72 | Matthew Newville | 18 March 2018, 20:34:00 UTC | use gformat for fit statistics | 18 March 2018, 20:34:00 UTC |
7b6bbaf | Matt Newville | 18 March 2018, 19:20:02 UTC | Merge pull request #460 from reneeotten/fit_report Improve output of fit_report | 18 March 2018, 19:20:02 UTC |
4f2b92f | reneeotten | 18 March 2018, 18:09:55 UTC | ENH: improve output of fit_report. | 18 March 2018, 18:09:55 UTC |
d42d72f | reneeotten | 18 March 2018, 18:09:00 UTC | STY: small changes to conform to PEP8. | 18 March 2018, 18:09:00 UTC |
aac5897 | Matt Newville | 18 March 2018, 17:45:53 UTC | Merge pull request #459 from lmfit/gformat_fix fix formatting error for values near 9.99e-5 | 18 March 2018, 17:45:53 UTC |
3c65f1a | Matthew Newville | 18 March 2018, 13:29:25 UTC | better handling of all range of formats 1e-15 to 1e15 | 18 March 2018, 13:29:25 UTC |
d4f7fe7 | Matthew Newville | 18 March 2018, 12:50:00 UTC | add missing import, use same lengt for val and stderr | 18 March 2018, 12:50:00 UTC |