Update master for release #73

mdeceglie · 2018-03-23T18:16:05Z

This incorporates changes that address issues: #67, #61, #59, #57, #56 as well as:

improvements/typo fixes to docstrings
Fixes error in check for two years of data in degradation_year_on_year
Improves the calculations underlying irradiance_rescale

All changes have ben through review when incorporated into the development branch. The purpose of this PR is to check that this version of rdtools installs and runs correctly.

For me, it passes everything in the release procedure, with the exception of the minimum requirement check on windows, but failure was due to problems building old versions of some requirements.

Update the normalized_energy info in degradation_year_on_year

…es from production-weather merged files (csv or pickle)

…tion rates from production-weather merged files (csv or pickle)" This reverts commit 09fbb33.

…tion rates from production-weather merged files (csv or pickle)" This reverts commit 47c996e.

This reverts commit 0b2f357.

This reverts commit 92faee7.

Merge issue_57 branch with development branch

Enables users to specify the size of the returned confidence interval

Fix erroneous check for two years of data

Irradiance rescale stability

Some changes in the notebook were overwritten by update from development branch. Also explicitly set method = ‘iterative” in rdtools.irradiance_rescale()

…rcent/year

… nosie Previous version used approximations such as 52 weeks in a year. This version uses 1 year = 365 days regardless of the frequency. It also adds noise so that the confidence interval calculations can be tested. Finally I simplified the structure.

Issue 59 confidence interval

* closes #56 * running `2to3 rdtools/` shows the only incompatibilities are the relative imports already addressed in #67 so merge #68 to fix imports * running `2to3 tests/` shows python 2 print commands, so rerun 2to3 with `-f print -w -n` to add parentheses to print statements to make them calls instead of commands, altho these should really be replaced by logging, so TODO: make another issue and PR to add logging to tests Signed-off-by: Mark Mikofski <bwana.marko@yahoo.com>

Signed-off-by: Mark Mikofski <bwana.marko@yahoo.com>

support Python 3

abshinn

Hi @mdeceglie. Thanks for putting this release together.

I ran python setup.py install in clean environments for both python 2.7 and python 3.6 using the latest development branch. Both installed without error.

When running tests/run_tests, in both python versions, all tests pass but I received the familiar FutureWarnings about pandas-statsmodels. Issue is already noted in #27. Not a show stopper for this release.

cdeline · 2018-03-30T15:37:58Z

I get installation errors in new environments of Python 2.7 and Python 3.6 with my Win10 machine. When installed into my existing environment, there are no errors.

Python 2.7:
from numpy.distutils.misc_util import get_info
ModuleNotFoundError: No module named 'numpy'

Python 3.6:
from numpy.distutils.misc_util import get_info
ImportError: No module named numpy.distutils.misc_util

This appears to be a known bug with statsmodels for Windows 10 machines:

statsmodels/statsmodels#3207

https://stackoverflow.com/questions/45492810/statsmodels-installation-no-module-named-numpy-distutils-msvccompiler-in-num

cdeline · 2018-03-30T15:56:42Z

Update: if I use 'conda install statsmodels' followed by 'pip install .' to install the development branch, then everything works in py2.7 and py3.6. Is there any way to install statsmodels through conda rather than pip in our setup.py file?

mdeceglie · 2018-03-30T17:25:44Z

@cdeline I was able to recreate this in a clean environment with only pip and python. However, I suggest we can proceed without directly addressing this. If a user has installed the anaconda python distribution (that's the most likely way they would get the conda package manager), they will already have statsmodels installed. If they are setting up environments, hopefully they have enough experience to navigate the requirements gymnastics.

We could add a note in the readme encouraging people to install requirements by hand if they run into trouble. Do you think that would help?

cdeline · 2018-03-30T18:37:30Z

OK, I'm good with this approach. If you could include an extra line in the readme about manually installing components using conda if they get errors during install, I'm fine with this.
I suspect that we'd still have installation issues if someone has a fresh non-Anaconda python distribution and just tries 'pip install rdtools', We don't specifically require Anaconda in the readme anywhere.

cdeline

Update readme to include a line about manually installing components using conda if necessary. Otherwise, everything installs and runs correctly.

mdeceglie · 2018-03-30T19:24:12Z

Agreed @cdeline, how's the updated readme look?

mdeceglie · 2018-03-30T22:11:38Z

Thanks for the input everyone!

Michael Deceglie and others added 30 commits December 15, 2017 13:23

Update the normalized_energy info in degradation_year_on_year

4b79d8d

Fix typo in degradation_year_on_year doc string

5a9a661

Another typo fix

ece87a2

Merge pull request #58 from NREL/degradation_doc_string

6954bad

Update the normalized_energy info in degradation_year_on_year

Make frequency inference more robust in clearsky_temperature.py

9a7ff83

typo fix

35311d7

Fix erroneous check for two years of data

3d11204

Fix issue 61

dbdba01

Added degradation_example_merged notebook to issue_57

92faee7

Added degradation_example notebook to issue_57

0b2f357

Updated NREL and public facing notebooks to calculate degradation rat…

47c996e

…es from production-weather merged files (csv or pickle)

Updated NREL and public facing notebooks to calculate degradation rat…

09fbb33

…es from production-weather merged files (csv or pickle)

Revert "Updated NREL and public facing notebooks to calculate degrada…

be9bf89

…tion rates from production-weather merged files (csv or pickle)" This reverts commit 09fbb33.

Revert "Updated NREL and public facing notebooks to calculate degrada…

05cdb10

…tion rates from production-weather merged files (csv or pickle)" This reverts commit 47c996e.

Revert "Added degradation_example notebook to issue_57"

19a5cff

This reverts commit 0b2f357.

Revert "Added degradation_example_merged notebook to issue_57"

92217b3

This reverts commit 92faee7.

Merge pull request #62 from NREL/issue_57

5005072

Merge issue_57 branch with development branch

Add confidence_level to degradation calcs

384209e

Enables users to specify the size of the returned confidence interval

Change confidence level to 95% in example notebook

cf44e7b

Change example confidence interval back to 68.2%

d729366

update irradiance_rescale with more stable iterative calculation

5cf0f32

Update example notebook to use iterative irradiance_rescale

a4799c6

Merge pull request #60 from NREL/yoy_2_year_check_fix

01dfbcc

Fix erroneous check for two years of data

Fix spelling errors

e892508

Merge pull request #64 from NREL/irradiance_rescale_stability

19d880e

Irradiance rescale stability

Update from development branch

97bf53f

Reinstate changes to notebook (confidence interval set to 68.2%)

9e7adda

Some changes in the notebook were overwritten by update from development branch. Also explicitly set method = ‘iterative” in rdtools.irradiance_rescale()

Merge branch 'development' into issue_59_confidence_interval

2ba88ff

add confidence interval tests

99d0e4c

Check confidence intervals relative to nominal degradation rate in pe…

0ef1315

…rcent/year

Michael Deceglie and others added 13 commits February 13, 2018 21:12

Add print statement for running test_confidence_intervals()

abfbb88

lock np seed to make deterministic

872ff44

Merge pull request #63 from NREL/issue_59_confidence_interval

9085d75

Issue 59 confidence interval

use full absolute path

a5dbc1f

Merge remote-tracking branch 'upstream/development' into futurize

2c8debf

Signed-off-by: Mark Mikofski <bwana.marko@yahoo.com>

use unicode

43c3dca

decode byte strings to unicode if Python-2

f154b30

use Python-3 compatible print()

56cedd3

use integer divide

9f0898a

Merge pull request #72 from mikofski/futurize

c42406d

support Python 3

Add python 3 to classifiers

c80d4e9

mdeceglie requested review from abshinn and cdeline March 23, 2018 18:16

abshinn approved these changes Mar 26, 2018

View reviewed changes

cdeline requested changes Mar 30, 2018

View reviewed changes

Update readme re:installation of requirements

d304c52

cdeline approved these changes Mar 30, 2018

View reviewed changes

mdeceglie merged commit af466d1 into master Mar 30, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update master for release #73

Update master for release #73

mdeceglie commented Mar 23, 2018

abshinn left a comment

cdeline commented Mar 30, 2018

cdeline commented Mar 30, 2018

mdeceglie commented Mar 30, 2018

cdeline commented Mar 30, 2018

cdeline left a comment

mdeceglie commented Mar 30, 2018

mdeceglie commented Mar 30, 2018

Update master for release #73

Update master for release #73

Conversation

mdeceglie commented Mar 23, 2018

abshinn left a comment

Choose a reason for hiding this comment

cdeline commented Mar 30, 2018

cdeline commented Mar 30, 2018

mdeceglie commented Mar 30, 2018

cdeline commented Mar 30, 2018

cdeline left a comment

Choose a reason for hiding this comment

mdeceglie commented Mar 30, 2018

mdeceglie commented Mar 30, 2018