7fec373 | Matt Newville | 26 November 2018, 15:25:27 UTC | Merge pull request #519 from reneeotten/PR512 updates to complement PR512 | 26 November 2018, 15:25:27 UTC |
deff523 | reneeotten | 22 November 2018, 04:06:14 UTC | MAINT: small fixes after merge of PR #512 | 22 November 2018, 15:48:35 UTC |
f9e92d5 | reneeotten | 21 November 2018, 15:54:41 UTC | TST: add simple test for using method='emcee' | 22 November 2018, 15:48:35 UTC |
92fa6d7 | reneeotten | 21 November 2018, 02:25:45 UTC | DOC: update release notes | 22 November 2018, 15:48:35 UTC |
0a3396f | reneeotten | 21 November 2018, 02:18:45 UTC | DOC: update text and output of emcee example - keep colors in the documents consistent | 22 November 2018, 15:48:35 UTC |
69dfd02 | Matt Newville | 21 November 2018, 21:22:18 UTC | Merge pull request #512 from FaustinCarter/add_emcee Add emcee | 21 November 2018, 21:22:18 UTC |
33f6c13 | Faustin Carter | 19 November 2018, 00:00:46 UTC | Add newline at end of file for PEP8 | 19 November 2018, 00:00:46 UTC |
2debc53 | Faustin Carter | 18 November 2018, 23:50:09 UTC | Remove orphan comment that was missed a few commits back. | 18 November 2018, 23:58:40 UTC |
a344d12 | Faustin Carter | 18 November 2018, 23:49:38 UTC | Add emcee method to table in docs | 18 November 2018, 23:58:00 UTC |
9a5734e | Faustin Carter | 16 November 2018, 01:25:26 UTC | Small color changes to make figures easier to parse. This will be followed with appropriate updates to the actual images that go in the docs | 18 November 2018, 23:58:00 UTC |
66d1b88 | Faustin Carter | 16 November 2018, 01:21:25 UTC | Remove orphan comment left over from create-by-copy | 18 November 2018, 23:57:36 UTC |
2e1e9e2 | Faustin Carter | 16 November 2018, 01:17:57 UTC | Rework emcee statistics calculation to use the standard method for the known cases and only explicitly calculate the special case. | 16 November 2018, 01:17:57 UTC |
a8d29f3 | Matt Newville | 12 November 2018, 03:30:00 UTC | Merge pull request #515 from reneeotten/minor-code-maintenance Minor code maintenance | 12 November 2018, 03:30:00 UTC |
ef1bbfa | reneeotten | 11 November 2018, 16:37:31 UTC | DOC: doc/docstring updates | 12 November 2018, 03:02:28 UTC |
34ea61f | Faustin Carter | 12 November 2018, 02:01:37 UTC | fix spacing for PEP8 | 12 November 2018, 02:01:37 UTC |
f43cb1f | Faustin Carter | 11 November 2018, 21:57:57 UTC | add vscode prefs to gitignore | 11 November 2018, 21:57:57 UTC |
4864268 | Faustin Carter | 11 November 2018, 21:56:38 UTC | remove duplicate assign statement | 11 November 2018, 21:56:38 UTC |
12e87da | Faustin Carter | 11 November 2018, 21:23:36 UTC | Properly reassign likelihood | 11 November 2018, 21:23:36 UTC |
d81ab26 | Faustin Carter | 10 November 2018, 22:48:45 UTC | move custom emcee statstics handling back to emcee from calc stats. Also deal with some whitespace isses. | 10 November 2018, 22:48:45 UTC |
e334bfa | reneeotten | 10 November 2018, 22:46:23 UTC | MAINT: remove Python shebang line | 10 November 2018, 22:48:14 UTC |
60cdcde | reneeotten | 10 November 2018, 22:24:19 UTC | STY: unused variable 'info' (unused-variable) | 10 November 2018, 22:48:14 UTC |
f243681 | reneeotten | 10 November 2018, 22:19:40 UTC | STY: reimport 'warnings' (reimported) | 10 November 2018, 22:48:14 UTC |
d08268e | reneeotten | 10 November 2018, 22:17:03 UTC | STY: import alias does not rename original package (useless-import-alias) | 10 November 2018, 22:48:14 UTC |
3c30dd7 | reneeotten | 10 November 2018, 22:09:05 UTC | MAINT: propagate kws to json.dumps | 10 November 2018, 22:48:14 UTC |
49f4d81 | reneeotten | 10 November 2018, 17:28:45 UTC | MAINT: remove uncertainties import from __init__.py - reorder some imports as suggested by isort (cannot sort them alphabetically as they are not independent) | 10 November 2018, 22:48:14 UTC |
54bb538 | reneeotten | 10 November 2018, 17:10:23 UTC | STY: wrong continued indentation | 10 November 2018, 22:48:14 UTC |
d7671be | reneeotten | 10 November 2018, 17:05:50 UTC | STY: missing whitespace around modulo operator | 10 November 2018, 22:48:14 UTC |
b7c3418 | reneeotten | 10 November 2018, 17:01:53 UTC | STY: too many blank lines | 10 November 2018, 22:48:14 UTC |
71232cd | reneeotten | 10 November 2018, 13:21:50 UTC | MAINT: remove import of __future__.print_function - this is present in Python >= 2.7 | 10 November 2018, 22:48:11 UTC |
a4da432 | reneeotten | 10 November 2018, 02:47:01 UTC | DOC: update docstring of __init__.py | 10 November 2018, 22:36:53 UTC |
4b57b82 | reneeotten | 09 November 2018, 22:30:37 UTC | MAINT: make use of numpy.isclose function - identical function present in numpy since version 1.7.0 | 10 November 2018, 22:36:49 UTC |
cc3a224 | Faustin Carter | 10 November 2018, 00:14:44 UTC | bad spelling of variable | 10 November 2018, 00:14:44 UTC |
a19cb39 | Faustin Carter | 09 November 2018, 23:47:57 UTC | compare equals is different than assign equals... sigh | 09 November 2018, 23:47:57 UTC |
2051e72 | Faustin Carter | 05 November 2018, 15:06:10 UTC | fix stupid thing I did that was dumb | 05 November 2018, 15:06:10 UTC |
0c08c77 | Matt Newville | 04 November 2018, 21:37:44 UTC | Merge pull request #514 from lmfit/voigt_fwhm improved FWHM calculation for VoigtModel, especially for gamma!=sigma | 04 November 2018, 21:37:44 UTC |
e380007 | Faustin Carter | 04 November 2018, 15:33:55 UTC | Update emcee routine and _calculate_statistics routine to properly handle calculating statsitics output for emcee runs | 04 November 2018, 15:33:55 UTC |
38bc05c | Faustin Carter | 04 November 2018, 15:33:09 UTC | Update emcee example to use minimizer with residual function. Still need to update docs! | 04 November 2018, 15:33:09 UTC |
152524b | Matthew Newville | 03 November 2018, 14:27:15 UTC | update test of Voigt FWHM to match new value | 03 November 2018, 14:27:15 UTC |
3a989c6 | Matthew Newville | 03 November 2018, 14:19:21 UTC | improved FWHM calculation for VoigtModel, especially for gamma!=sigma | 03 November 2018, 14:19:21 UTC |
47fce05 | Matt Newville | 02 November 2018, 01:29:10 UTC | Merge pull request #513 from lmfit/license_clarify clarify License to be more standard BSD-3, some doc updates | 02 November 2018, 01:29:10 UTC |
ffcde29 | Faustin Carter | 01 November 2018, 18:00:49 UTC | Move updated example to new location in prep for restoring old example | 01 November 2018, 18:00:49 UTC |
e0682e6 | Faustin Carter | 01 November 2018, 17:59:05 UTC | Cast generator to list for python 3 compat | 01 November 2018, 17:59:05 UTC |
dabb061 | Matthew Newville | 01 November 2018, 02:35:48 UTC | more acknowledgments of prior work | 01 November 2018, 02:35:48 UTC |
fe4e645 | Matthew Newville | 01 November 2018, 01:17:07 UTC | make copyright be a true sub-section of installation doc | 01 November 2018, 01:17:07 UTC |
85ef355 | Matthew Newville | 01 November 2018, 01:13:27 UTC | fix many typos | 01 November 2018, 01:13:27 UTC |
982fe81 | Matthew Newville | 01 November 2018, 01:13:06 UTC | restore scipy license to LICENSE file | 01 November 2018, 01:13:06 UTC |
e1fabc7 | Faustin Carter | 31 October 2018, 21:12:09 UTC | Update emcee example to use lmfit.Model interface | 31 October 2018, 21:12:09 UTC |
4df4923 | Matthew Newville | 31 October 2018, 02:55:53 UTC | add FAQ entry on NaNs | 31 October 2018, 02:55:53 UTC |
3d95b08 | Matthew Newville | 31 October 2018, 02:55:37 UTC | some minorish doc updates | 31 October 2018, 02:55:37 UTC |
17f54b9 | Matthew Newville | 31 October 2018, 02:55:24 UTC | udpate THANKS.txt | 31 October 2018, 02:55:24 UTC |
db1b340 | Matthew Newville | 31 October 2018, 02:54:46 UTC | update LICENSE to clarify BSD-3 | 31 October 2018, 02:54:46 UTC |
f9b1b12 | Faustin Carter | 30 October 2018, 16:58:16 UTC | fix typo in docstrings | 30 October 2018, 16:58:16 UTC |
eb56b1f | Faustin Carter | 29 October 2018, 17:54:44 UTC | Remove unrelated unused variable | 29 October 2018, 17:55:02 UTC |
b2be396 | Faustin Carter | 28 October 2018, 19:21:36 UTC | Add 'emcee' as method kwarg to both Minimizer.minimize and minimize. | 29 October 2018, 17:55:01 UTC |
1b70f92 | Matt Newville | 24 October 2018, 21:29:48 UTC | Merge pull request #511 from reneeotten/build-tests-documentation updates for building, tests, and documentation | 24 October 2018, 21:29:48 UTC |
b3a1291 | reneeotten | 23 October 2018, 04:39:58 UTC | ENH: add all arguments to the minimize wrapper function | 24 October 2018, 15:47:58 UTC |
a8ddc0b | reneeotten | 23 October 2018, 04:25:13 UTC | DOC: update whatsnew.rst | 24 October 2018, 15:47:58 UTC |
787ac9a | reneeotten | 23 October 2018, 03:40:49 UTC | DOC: update installation.rst - update supported Python versions - remove references to nose, tests should be run with pytest | 24 October 2018, 15:47:58 UTC |
b9d0296 | reneeotten | 22 October 2018, 21:41:01 UTC | BLD/STY: add isort section to setup.cfg and run on source | 24 October 2018, 15:47:58 UTC |
cd3fc38 | reneeotten | 22 October 2018, 14:38:22 UTC | TST/STY: fixes in tests prompted by flake8 / pylint | 24 October 2018, 15:47:58 UTC |
0c9fe1c | reneeotten | 22 October 2018, 14:29:54 UTC | TST: remove redundant __main__ statements | 24 October 2018, 15:47:57 UTC |
7a40947 | reneeotten | 22 October 2018, 14:25:22 UTC | TST: merge unused _test_ci.py into test_confidence.py | 24 October 2018, 15:47:57 UTC |
f9b2647 | reneeotten | 22 October 2018, 14:19:55 UTC | TST: remove unused _test_make_paras_and_func.py | 24 October 2018, 15:47:57 UTC |
37cbc1b | reneeotten | 23 October 2018, 17:20:19 UTC | BLD/TST: use https for conda website | 24 October 2018, 15:47:57 UTC |
75ede2a | reneeotten | 24 October 2018, 15:32:28 UTC | BLD: update setup.py - correct typo --> tests_require - explicitly add Python versions - add license to classifiers as recommended - add keywords | 24 October 2018, 15:47:57 UTC |
e278951 | Matthew Newville | 23 October 2018, 11:48:40 UTC | use inspect.getfullargspec for Python3 | 23 October 2018, 11:48:40 UTC |
744cbd9 | Matt Newville | 22 October 2018, 23:56:51 UTC | Merge pull request #506 from reneeotten/numdifftools add numdifftools to estimate parameter uncertainties/correlations | 22 October 2018, 23:56:51 UTC |
6f62789 | Matt Newville | 22 October 2018, 20:42:29 UTC | Merge pull request #510 from lmfit/params_usersyms Add `usersyms` to Parameters() initialization | 22 October 2018, 20:42:29 UTC |
fb34f7e | reneeotten | 08 October 2018, 20:39:13 UTC | DOC: update documentation for numdifftools | 21 October 2018, 00:27:15 UTC |
453f584 | reneeotten | 07 October 2018, 21:08:10 UTC | FIX: catch exception if numdifftools encounters NaN | 21 October 2018, 00:27:15 UTC |
11b8220 | reneeotten | 07 October 2018, 20:57:36 UTC | FIX: numdifftools needs len(residual) > len(pars) Some tests (test_ampgo.py, test_basinhopping.py) were failing. To me it seems the reason is there we are trying to find the minimum of a function instead of fitting a function to data points. I *think* that numdifftools requires that the output of a function is an array with more elements than fitting parameters. Adding this requirement will not calculate a covariance_matrix when fitting a function, but probably not many people use lmfit for this anyway... | 21 October 2018, 00:27:15 UTC |
b05a90f | reneeotten | 17 September 2018, 22:49:55 UTC | TST: add tests for numdifftools cov_x and uncertainties/correlations | 21 October 2018, 00:27:15 UTC |
1455b05 | reneeotten | 17 September 2018, 21:53:00 UTC | ENH: calculate cov_x if not done by the minimizers - attempts to calculate the covariance matrix for minimizers other than (1) leastsq, least_square and emcee that already do this and (2) the brute force method - if successful, calculate parameter uncertainties and correlations | 21 October 2018, 00:27:15 UTC |
02f76e7 | reneeotten | 08 October 2018, 00:05:33 UTC | ENH: add calc_covar to Model class | 21 October 2018, 00:27:15 UTC |
64b4c24 | reneeotten | 13 June 2018, 01:54:07 UTC | ENH: add method to calculate the covariance matrix * calc_covar (default is True) added as an argument to the Minimizer class that determines whether or not to calculate the covariance matrix. * requires the numdifftools package to be installed | 21 October 2018, 00:27:15 UTC |
3a79237 | reneeotten | 07 October 2018, 15:42:07 UTC | FIX: nfree cannot be < 0 | 21 October 2018, 00:27:15 UTC |
e1f198b | Matthew Newville | 20 October 2018, 17:12:16 UTC | pep8 / pylint cleanups | 20 October 2018, 17:12:16 UTC |
600257c | Matthew Newville | 14 October 2018, 23:52:15 UTC | more uniform use of encode4js | 14 October 2018, 23:52:15 UTC |
b364a7e | Matthew Newville | 14 October 2018, 23:51:52 UTC | add storing and retrieving of callables using import path of callable if dill is unavailable | 14 October 2018, 23:51:52 UTC |
803da4c | Matthew Newville | 14 October 2018, 23:48:32 UTC | add test for Parameter usersyms | 14 October 2018, 23:49:53 UTC |
553488a | Matthew Newville | 14 October 2018, 23:47:55 UTC | add usersyms keyword argument to Parameters() to populate symbol table | 14 October 2018, 23:47:55 UTC |
cac2d38 | Matt Newville | 14 October 2018, 14:32:40 UTC | Merge pull request #508 from lmfit/small_fixes Small fixes | 14 October 2018, 14:32:40 UTC |
bc366ee | Matt Newville | 11 October 2018, 13:01:58 UTC | Merge pull request #509 from lmfit/revert-503-plot-complex Revert "Handle plotting of complex-valued models in ModelResult class" | 11 October 2018, 13:01:58 UTC |
476bffc | Matthew Newville | 07 October 2018, 14:55:26 UTC | numpy tests still require nose to be installed | 10 October 2018, 23:03:26 UTC |
1921a37 | Matthew Newville | 07 October 2018, 13:21:32 UTC | ModelResult.eval() and .eval_components() should not permanently overwrite .userkws | 10 October 2018, 23:03:26 UTC |
9efc6ed | Faustin Carter | 09 October 2018, 01:41:34 UTC | Handle plotting of complex-valued models in ModelResult class (#503) * Update plot routines in ModelResult class to gracefully handle models that return vectors of complex numbers. Adds options to convert the complex vector to a real one via several standard methods. * Update dosctrings and also add a line to the FAQ that is relevant to complex fitting. * Fix error propagation to not be wrong for phase error estimation when magnitude is zero * Change phase error to pi for case where magnitude is zero * Make a vector valued comparison not crash by using all * Update docstrings to be more compliant and refactor input-checking assertions to all be in central location * More docstring tweaking * Remove superfluous assert statement since this is handled with a raise later * consistent labeling of default option in docstrings * Replace asserts and default logic raises with explicit parameter test/raise * normalize angle errors by the square of the magnitude so that units are correct. Missed this earlier... * trim trailing whitespace and remove newlines between docstring and code | 09 October 2018, 01:41:34 UTC |
6a945a6 | Jonathan Fraine | 07 October 2018, 21:25:09 UTC | Added progress bar to emcee call (#496) * Add `sampler_kwargs` as kwarg to Minimizer.emcee By adding `sampler_kwargs` as a kwarg to `Minimizer.emcee`, we can now send kwargs directly to emcee that are not consider by default in the current Minimizer instance. Example: `emcee` was updated to use `tqdm` as a progress bar over the MCMC sampling; but this was not added into the `lmfit-py`. To avoid requiring `lmfit-py` to update itself with every new kwarg from `emcee`, passing a `sampler_kwargs` into the `Minimizer` instance allows users to access all `emcee` kwargs directly. * Deleted sampler_kwargs and added progress=True The opposite of what I hoped for, I had to manually add the flag `progress=True` (`line 879`) and `progress=progress` (`line 1200`), because the progress bar is allocated in the `self.sampler` method `self.sampler.run` (where `self.sampler = EnsembleSampler(...)`). It works great; but the precedent is that we need to add a new flag for each `.run` kwarg. * minimized the number of PR corrections * add tqdm to the dependencies * updated requirements for emcee * removed dependencies for sub features * include a emcee major version check Because the latest PYPI version of `emcee` does not include the `progress` kwarg -- to be released with v3.0.dev0 -- we added a version check to confirm that the `emcee` major version installed is >3. If it is, then `lmfit` with use progress with `progress=progress` on `line 1201`. The flag (`progress=True` `line 880`) can still be deactivated manually, even if `emcee.__version__ >= 3`. * Include emcee v3 check in self._lastpos line1200 * Configured EMCEE_VERSION Check and output Changes the variable name and behaviour of the EMCEE version check for version>=3, now it runs one set of 2 lines, as oppose to the default set of 2 lines | 07 October 2018, 21:25:09 UTC |
900095d | Matthew Newville | 07 October 2018, 13:17:54 UTC | convert tests to not rely on nose, using pytest instead where needed | 07 October 2018, 13:17:54 UTC |
aabf1f7 | Matthew Newville | 07 October 2018, 13:16:43 UTC | auto-convert independent var to array only if it is a list or tuple -- allow complex objects as independent vars | 07 October 2018, 13:16:43 UTC |
d4d8958 | Matt Newville | 05 October 2018, 21:51:27 UTC | Merge pull request #501 from reneeotten/least_squares calculate uncertainties in least_squares method | 05 October 2018, 21:51:27 UTC |
a7ca476 | reneeotten | 02 October 2018, 18:51:17 UTC | TST: fix DeprecationWarning random_integers --> randint | 02 October 2018, 18:51:17 UTC |
cc306ba | reneeotten | 24 September 2018, 18:05:01 UTC | ENH: set result.errorbars to True in _calculate_uncertainties_correlations() | 24 September 2018, 18:08:26 UTC |
c98bbc6 | reneeotten | 24 September 2018, 14:29:26 UTC | TST: correctness of parameter uncertainties/correlations with least_squares | 24 September 2018, 18:04:35 UTC |
71bcc8b | reneeotten | 24 September 2018, 13:54:19 UTC | ENH: estimate uncertainties when using least_squares - update docstring to reflect that we only support the 'trf' method - use the Jacobian returned by scipy.least_squares to estimate the covariance matrix and calculate uncertainties/correlations | 24 September 2018, 18:04:24 UTC |
7fbb75b | Matt Newville | 24 September 2018, 16:58:00 UTC | Merge pull request #498 from benmaier/master Allow for scipy dev version; fix #497 | 24 September 2018, 16:58:00 UTC |
db8ad40 | Benjamin F. Maier | 14 September 2018, 11:09:20 UTC | allow for scipy dev version; fix #497 | 19 September 2018, 14:24:52 UTC |
a288292 | Matt Newville | 12 September 2018, 01:58:53 UTC | Merge pull request #495 from lmfit/skewed_voigt add skewed voigt model | 12 September 2018, 01:58:53 UTC |
7504a0b | Matt Newville | 11 September 2018, 18:19:38 UTC | Merge pull request #494 from lmfit/fix_load_modelresult Fix load modelresult | 11 September 2018, 18:19:38 UTC |
df9c658 | Matthew Newville | 09 September 2018, 22:10:56 UTC | add skewed voigt model | 09 September 2018, 22:10:56 UTC |
5e93536 | Matthew Newville | 09 September 2018, 19:18:13 UTC | add test of data and init_params in loaded modelresult | 09 September 2018, 19:18:13 UTC |