Update master branch to v0.15.0 #229

raoulcollenteur · 2020-07-31T10:25:18Z

Short Description

Update master branch to v0.15.0

This method makes it possible to set all parameter properties in one go.

- Allow rfunc to be none, when no response function is used in the stressmodel - rfunc class is now instantiated in the stress model init, not in stressmodelbase

- Only parameter that are within the boundaries are now excepted when sampling from the multivariate normal distribution using truncation. - Solves issue #169 - Solves a new bug caused by Pandas reindexing

Methods otherwise take forever and run into memory problems

Only when the recharge model is Linear. This commit fixes the problem described in issue #120 Also, this commit fixes an error in StressModel2 where the evaporation times the factor was plotted in ml.plots.stresses. Now this factor is not applied when only plotting the stresses.

- Fix PEP8 - Add istress to docstring - Fix new Bug that was created when istress was not None, but nsplit !> 1 - improve efficiency

Make default rfunc in StressModelBase None #186

Closes issue #192

Don't run on travis because the Notebook depends on statsmodels.

Closing Issue #196

Introduce ArmaModel to Pastas

Bounds were unnecessarily small and improved initial values.

Differences compared to NoiseModel: 1. First value is residual 2. First weight is 1 / sig_residuals (i.e., delt = infty) 3. Sum of all weights is always 1

and update the initial parameters, these were the old ones.

Basicly only internal changes. Changes according to Issue #206, related to Issue #152.

Add a parameter to the Flexmodel that allows to update the potential evaporation. Default vary=False for this parameters, so it does not affect anything.

Much smaller makes no sense. This may cause problems in computation of the first few values in the four-parameter function, as it is integrated numerically.

New NoiseModel. See docs of NoiiseModel for differences with NoiseModelOld.

Ran the model to make sure it works with new noisemodel.

Closing Issue #152

… format

remove some old commented code

update notebook to include examples of other pastas read methods in future

Dinoloket read water level method

- Fix issues #205 and #203 - rename some keywords - new default value for bin_width = 0.5 - new number of lags is all up to 365 days, and remove those that could not be computed later using the min_obs argument (default min_obs=100) now.

.github/ISSUE_TEMPLATE/bug_report.md

raoulcollenteur and others added 30 commits April 1, 2020 08:05

Update Dev branch to 0.15.0b and update deprecations

173a1a4

Make set_parameter method public #190

d829a38

This method makes it possible to set all parameter properties in one go.

Support move_bounds in set_parameter

091e25b

#190

Make default rfunc in StressModelBase None #186

aaed07e

- Allow rfunc to be none, when no response function is used in the stressmodel - rfunc class is now instantiated in the stress model init, not in stressmodelbase

Fix StepModel

59dcd1b

Make sure ml.plots.results works without stressmodels

3025161

make sure notebook 12 and 14 will dsplay in docs

9a9bd55

Fix Issue #169 by using truncated multivariate normal distribution

7ca2864

- Only parameter that are within the boundaries are now excepted when sampling from the multivariate normal distribution using truncation. - Solves issue #169 - Solves a new bug caused by Pandas reindexing

fix ci_simulation and set n=1000 as default

a8e1204

Methods otherwise take forever and run into memory problems

Fix issues with commit 8a6da4c

cf40909

- Fix PEP8 - Add istress to docstring - Fix new Bug that was created when istress was not None, but nsplit !> 1 - improve efficiency

Merge pull request #191 from pastas/sm_rfunc

e8c682e

Make default rfunc in StressModelBase None #186

Add objective function value to fit report and ml.fit

2277974

Closes issue #192

Add numba and lmfit to show_versions method

65ababc

Introduce ArmaModel in Pastas

c64431d

Add test and don't run notebook on Travis

575a9d2

Don't run on travis because the Notebook depends on statsmodels.

Make get_stress methods consistent

0fb3a88

Closing Issue #196

Update LICENSE

013a7b7

change theta to beta

46ff758

Changed name from Bruggeman to Polder in table

cb32882

Generate ARMA(1,1) noise without statsmodels and test notebook

db79cc1

Merge pull request #198 from pastas/armamodel

14d4de3

Introduce ArmaModel to Pastas

Update some bounds and default values

86bb826

Bounds were unnecessarily small and improved initial values.

First attempt at NoiseModel3

950c451

Differences compared to NoiseModel: 1. First value is residual 2. First weight is 1 / sig_residuals (i.e., delt = infty) 3. Sum of all weights is always 1

Set first value of ARMA noise to the residual

5a20ab3

and update the initial parameters, these were the old ones.

Merge branch 'dev' of https://github.com/pastas/pasta into dev

4fc7418

Merge example notebook 4 and 11 #187

d752b31

separate noise and weights

0d8ea2e

Basicly only internal changes. Changes according to Issue #206, related to Issue #152.

Update Flexmodel and fix Issue #200

adb68e6

Add a parameter to the Flexmodel that allows to update the potential evaporation. Default vary=False for this parameters, so it does not affect anything.

Update examples in docs

4d59526

mbakker7 and others added 19 commits July 22, 2020 22:11

lower bound of b parameter set to 1e-6

52eb894

Much smaller makes no sense. This may cause problems in computation of the first few values in the four-parameter function, as it is integrated numerically.

New NoiseModel

310146b

New NoiseModel. See docs of NoiiseModel for differences with NoiseModelOld.

Update 1_basic_model.ipynb

23e3c98

Ran the model to make sure it works with new noisemodel.

Merge branch 'dev' of https://github.com/pastas/pastas into dev

8339d17

Remove noisemodel2 and update api docs for the noise models

8e3ac79

Closing Issue #152

example notebook and data for reading river level gauges in Dinoloket…

4d9f396

… format

top levels imports for reading river level gauges in Dinoloket format

045af7a

removed unused import

154b749

forward slashes for filepaths; fix for travis

73593fc

forward slashes in filepaths example notebook

358ee00

autopep8, docformatter

80ca82f

remove some old commented code

rename notebook

1ff9c32

update notebook to include examples of other pastas read methods in future

Merge pull request #227 from tomvansteijn/dev

b14023d

Dinoloket read water level method

min_obs in test because test time series are very short.

5c6ec47

Update all Notebooks and run PEP8 on all files

dcd9423

Up[date release notes

2dc7dd2

Update some parameter names in non-linear recharge model plus docs

d58a5e7

Update Master to v0.15

3154865

raoulcollenteur requested review from dbrakenhoff and rubencalje and removed request for dbrakenhoff July 31, 2020 10:25

Couple of Codacy fixes and NB3

26b0c31

dbrakenhoff reviewed Jul 31, 2020

View reviewed changes

.github/ISSUE_TEMPLATE/bug_report.md Outdated Show resolved Hide resolved

dbrakenhoff approved these changes Jul 31, 2020

View reviewed changes

Update bug_report.md

bd93902

dbrakenhoff approved these changes Jul 31, 2020

View reviewed changes

raoulcollenteur removed the request for review from rubencalje July 31, 2020 12:42

raoulcollenteur merged commit ff1a1b5 into master Jul 31, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update master branch to v0.15.0 #229

Update master branch to v0.15.0 #229

raoulcollenteur commented Jul 31, 2020

Update master branch to v0.15.0 #229

Update master branch to v0.15.0 #229

Conversation

raoulcollenteur commented Jul 31, 2020

Short Description