ModelChainResult.__repr__ by cwhanse · Pull Request #1236 · pvlib/pvlib-python

cwhanse · 2021-05-18T03:42:22Z

ClosesImprove ModelChainResult.__repr__ #1110
I am familiar with the contributing guidelines
~~[ ] Tests added~~
Updates entries to docs/sphinx/source/api.rst for API changes.
[ ] Adds description and name entries in the appropriate "what's new" file in docs/sphinx/source/whatsnew for all changes. Includes link to the GitHub Issue with :issue:`num` or this Pull Request with :pull:`num`. Includes contributor name and/or GitHub username (link with :ghuser:`user`).
New code is fully documented. Includes numpydoc compliant docstrings, examples, and comments where necessary.
Pull request is nearly complete and ready for detailed review.
Maintainer: Appropriate GitHub Labels and Milestone are assigned to the Pull Request and linked Issue.

…es_repr

cwhanse · 2023-06-22T16:08:12Z

I'm not unhappy with the print of ModelChainResult here. The code is clumsy, because ModelChainResult doesn't hew to any standards for the type of each attribute (e.g., with 2 Arrays, the value can be None, a single float, or a tuple of something).

Here's what the report looks like for a simple ModelChain (system with one Array). I'm "head"ing any Series or DataFrames.

Any if you read this far, look at the name of the last Series for ModelChainResult.ac.

=== ModelChainResult === weather ghi dhi ... temp_air wind_speed
2023-06-10 10:00:00-07:00 775.999173 108.933862 ... 25 0
2023-06-10 11:00:00-07:00 898.985531 116.633882 ... 25 0
2023-06-10 12:00:00-07:00 969.592761 120.879109 ... 25 0
2023-06-10 13:00:00-07:00 982.786788 121.660833 ... 25 0
2023-06-10 14:00:00-07:00 937.637996 118.971374 ... 25 0
[5 rows x 5 columns] solar_position apparent_zenith ... equation_of_time
2023-06-10 10:00:00-07:00 37.382226 ... 0.539325
2023-06-10 11:00:00-07:00 25.474558 ... 0.530968
2023-06-10 12:00:00-07:00 15.254955 ... 0.522605
2023-06-10 13:00:00-07:00 12.488151 ... 0.514235
2023-06-10 14:00:00-07:00 20.492312 ... 0.505857
[5 rows x 6 columns] airmass airmass_relative airmass_absolute
2023-06-10 10:00:00-07:00 1.257469 1.257469
2023-06-10 11:00:00-07:00 1.107109 1.107109
2023-06-10 12:00:00-07:00 1.036112 1.036112
2023-06-10 13:00:00-07:00 1.023853 1.023853
2023-06-10 14:00:00-07:00 1.067070 1.067070 Number of Arrays: 1 ------------------- Array 0 ------------------- tracking None aoi 2023-06-10 10:00:00-07:00 37.382226
2023-06-10 11:00:00-07:00 25.474558
2023-06-10 12:00:00-07:00 15.254955
2023-06-10 13:00:00-07:00 12.488151
2023-06-10 14:00:00-07:00 20.492312
Freq: H, Name: aoi, dtype: float64 aoi_modifier 1.0 total_irrad poa_global ... poa_ground_diffuse
2023-06-10 10:00:00-07:00 775.994182 ... 0.0
2023-06-10 11:00:00-07:00 898.983256 ... 0.0
2023-06-10 12:00:00-07:00 969.591954 ... 0.0
2023-06-10 13:00:00-07:00 982.786249 ... 0.0
2023-06-10 14:00:00-07:00 937.636532 ... 0.0
[5 rows x 5 columns] spectral_modifier 1 effective_irradiance 2023-06-10 10:00:00-07:00 775.994182
2023-06-10 11:00:00-07:00 898.983256
2023-06-10 12:00:00-07:00 969.591954
2023-06-10 13:00:00-07:00 982.786249
2023-06-10 14:00:00-07:00 937.636532
Freq: H, dtype: float64 cell_temperature 2023-06-10 10:00:00-07:00 51.474617
2023-06-10 11:00:00-07:00 55.670639
2023-06-10 12:00:00-07:00 58.079598
2023-06-10 13:00:00-07:00 58.529749
2023-06-10 14:00:00-07:00 56.989374
Freq: H, dtype: float64 dc 2023-06-10 10:00:00-07:00 173.454396
2023-06-10 11:00:00-07:00 197.173423
2023-06-10 12:00:00-07:00 210.324276
2023-06-10 13:00:00-07:00 212.743986
2023-06-10 14:00:00-07:00 204.414727
Freq: H, Name: p_mp, dtype: float64 dc_ohmic_losses None End of Arrays ------------------- losses 1 ac 2023-06-10 10:00:00-07:00 166.923393
2023-06-10 11:00:00-07:00 189.648371
2023-06-10 12:00:00-07:00 202.216782
2023-06-10 13:00:00-07:00 204.526893
2023-06-10 14:00:00-07:00 196.571725
Freq: H, Name: p_mp, dtype: float64

cwhanse · 2023-06-22T16:44:13Z

@kandersolar your call if this is for v0.10 or later.

kandersolar · 2023-06-22T18:38:39Z

I think using head for each attribute takes up a ton of space in the output just to display what is going to be mostly boring nighttime values for most simulations. How about just showing the type and shape of the data, but not any actual values? For example, something like this:

ModelChainResult:
weather: DataFrame (145 rows x 5 columns)
solar_position: DataFrame (145 rows x 6 columns)
airmass: DataFrame (145 rows x 2 columns)
tracking: None
aoi: Tuple (Series (length 145), Series (length 145))
aoi_modifier: Tuple (Series (length 145), Series (length 145))
total_irrad: Tuple (DataFrame (145 rows x 5 columns), DataFrame (145 rows x 5 columns))
spectral_modifier: Tuple (1, 1)
effective_irradiance: Tuple (Series (length 145), Series (length 145))
cell_temperature: Tuple (Series (length 145), Series (length 145))
dc: Tuple (Series (length 145), Series (length 145))
dc_ohmic_losses: None
losses: 1
ac: Series (length 145)

Code for the above:

def_mcr_repr(obj):
ifisinstance(obj, tuple):
return"Tuple ("+", ".join([_mcr_repr(o) foroinobj]) +")"ifisinstance(obj, pd.DataFrame):
return"DataFrame ({} rows x {} columns)".format(*obj.shape)
ifisinstance(obj, pd.Series):
return"Series (length {})".format(len(obj))
# scalar, None, other?returnrepr(obj)
def__repr__(self):
lines= ['ModelChainResult:']
attrs= [
'weather', 'solar_position', 'airmass',
'tracking', 'aoi', 'aoi_modifier', 'total_irrad', 'spectral_modifier',
'effective_irradiance', 'cell_temperature', 'dc', 'dc_ohmic_losses', 'losses', 'ac'
]
forattrinattrs:
lines.append(f" {attr}: "+_mcr_repr(getattr(self, attr)))
return"\n".join(lines)

cwhanse · 2023-06-23T14:55:45Z

I had three outcomes in mind:

report number of Arrays
report attributes and their shape. I was doing this by writing a report for each Array, but beyond 2 or 3 Arrays, that will be text barf on the screen.
provide a summary of key data, e.g., the DatetimeIndex. I could do that once rather than using head() for each DataFrame or Series.

So I'll drop the head except for times

cwhanse · 2023-06-23T15:20:09Z

Code I'm using to display the report for different PVSystems

import pytz
import pandas as pd
import pvlib
loc = pvlib.location.Location(35, -116)
dr = pd.date_range(start='2023-06-10 10:00:00', end='2023-06-11 10:00:00',
freq='H', tz=pytz.timezone('Etc/GMT+7'))
weather = loc.get_clearsky(dr)
weather['temp_air'] = 25
module_parameters = {'pdc0': 250, 'gamma_pdc': -0.004}
inverter_parameters = {'pdc0': 250}
temperature_parameters = pvlib.temperature.TEMPERATURE_MODEL_PARAMETERS['sapm']['open_rack_glass_glass']
pvs = pvlib.pvsystem.PVSystem(module_parameters=module_parameters,
inverter_parameters=inverter_parameters,
temperature_model_parameters=temperature_parameters)
mc = pvlib.modelchain.ModelChain(pvs, loc, aoi_model='no_loss',
spectral_model='no_loss',
temperature_model='sapm')
mc.run_model(weather)
mcres = mc.results
print(mcres)
mount = pvlib.pvsystem.FixedMount(surface_tilt=30, surface_azimuth=180,
racking_model='open_rack')
array = pvlib.pvsystem.Array(mount=mount, albedo=0.3,
module_parameters=module_parameters,
temperature_model_parameters=temperature_parameters)
pvs2 = pvlib.pvsystem.PVSystem(arrays=(array, array),
inverter_parameters=inverter_parameters)
mc2 = pvlib.modelchain.ModelChain(pvs2, loc, aoi_model='no_loss',
spectral_model='no_loss',
temperature_model='sapm')
mc2.run_model(weather)
mcres2 = mc2.results
print(mcres2)
mc3 = pvlib.modelchain.ModelChain(pvs2, loc, aoi_model='no_loss',
spectral_model='no_loss',
temperature_model='sapm')
mc3.run_model((weather, weather))
mcres3 = mc3.results
print(mcres3)

kandersolar

Probably worth a test? We used to have a ModelChain.__repr__ test, but it looks like I incorrectly removed it in #1181, oops. We should probably bring that back.

kandersolar · 2023-06-23T16:01:13Z

+ desc2 = (f'Number of Arrays: {num_arrays} \n')
+ attr = 'times'
+ desc3 = ('Times (first 3)\n' +
+ f'{_head(_getmcattr(self, attr))}' +


Suggested change
f'{_head(_getmcattr(self, attr))}'+
f'{self.times[:3]}'+
Seems like this should be fine as long as it's safe to assume self.times is a DatetimeIndex

Won't work if times is a DatetimeStamp. Maybe that can't happen, idk.

cwhanse · 2023-06-23T18:19:59Z

Probably worth a test? We used to have a ModelChain.__repr__ test, but it looks like I incorrectly removed it in #1181, oops. We should probably bring that back.

I added the Modelchain.__repr__ test back.

For the ModelChainResult.__repr__ test, I'm checking that all the non-private attributes appear. That provides some future-proofing, without a cumbersome string==string check.

kandersolar · 2023-06-23T18:44:53Z

+@pytest.mark.parametrize('strategy, strategy_str', [
+ ('south_at_latitude_tilt', 'south_at_latitude_tilt'),
+ (None, 'None')]) # GitHub issue 352


Thanks for putting this test back. I think this decorator is no longer needed and can be deleted.

kandersolar · 2023-06-23T18:49:18Z

- 'aoi_modifier', 'total_irrad', 'spectral_modifier',
- 'effective_irradiance', 'cell_temperature', 'diode_params',
- 'dc', 'dc_ohmic_losses', 'losses', 'ac']
+ mc_attrs = dir(self)


Using dir sorts the list of attributes alphabetically instead of conceptually. Not necessarily a downside.

Yeah, once I saw the alphabetized list my thought was "that's easier to use"

kandersolar · 2023-06-23T20:13:28Z

Thanks @cwhanse!

cwhanse added 6 commits May 17, 2021 21:40

initial __repr__ for ModelChainResult
edf3b53

fix timezone handling
0551447

merge upstream
2b52088

working on it
068c72b

Merge branch 'main' ofhttps://github.com/pvlib/pvlib-pythoninto mcr…
df268d0
…es_repr

working now
db68853

cwhanse added 3 commits June 22, 2023 10:10

jettison that space
3b005a2

whatsnew
0963ca9

inline comments
db1a33d

cwhanse marked this pull request as ready for review June 22, 2023 16:42

cwhanse added the enhancement label Jun 22, 2023

shorten it up
7f66f43

kandersolar reviewed Jun 23, 2023
View reviewed changes

test, add ModelChain.__repr__ test also
ce3b77b

kandersolar added this to the 0.10.0 milestone Jun 23, 2023

kandersolar approved these changes Jun 23, 2023
View reviewed changes

remove unneeded decorator
dde8221

kandersolar merged commit 209376e into pvlib:mainJun 23, 2023

Uh oh!

Conversation

cwhanse commented May 18, 2021• edited by kandersolar Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cwhanse commented Jun 22, 2023

Uh oh!

cwhanse commented Jun 22, 2023

Uh oh!

kandersolar commented Jun 22, 2023

Uh oh!

cwhanse commented Jun 23, 2023• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cwhanse commented Jun 23, 2023

Uh oh!

kandersolar left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

kandersolarJun 23, 2023

Choose a reason for hiding this comment

Uh oh!

cwhanseJun 23, 2023

Choose a reason for hiding this comment

Uh oh!

Uh oh!

cwhanse commented Jun 23, 2023

Uh oh!

kandersolarJun 23, 2023

Choose a reason for hiding this comment

Uh oh!

kandersolarJun 23, 2023

Choose a reason for hiding this comment

Uh oh!

cwhanseJun 23, 2023

Choose a reason for hiding this comment

Uh oh!

kandersolar commented Jun 23, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

cwhanse commented May 18, 2021•
edited by kandersolar
Loading

cwhanse commented Jun 23, 2023•
edited
Loading