ModelChainResult.__repr__ #1236
Conversation
|
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 |
|
@kandersolar your call if this is for v0.10 or later. |
|
I think using Code for the above: def _mcr_repr(obj):
if isinstance(obj, tuple):
return "Tuple (" + ", ".join([_mcr_repr(o) for o in obj]) + ")"
if isinstance(obj, pd.DataFrame):
return "DataFrame ({} rows x {} columns)".format(*obj.shape)
if isinstance(obj, pd.Series):
return "Series (length {})".format(len(obj))
# scalar, None, other?
return repr(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'
]
for attr in attrs:
lines.append(f" {attr}: " + _mcr_repr(getattr(self, attr)))
return "\n".join(lines) |
|
I had three outcomes in mind:
So I'll drop the |
|
Code I'm using to display the report for different PVSystems |
There was a problem hiding this comment.
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.
| desc2 = (f'Number of Arrays: {num_arrays} \n') | ||
| attr = 'times' | ||
| desc3 = ('Times (first 3)\n' + | ||
| f'{_head(_getmcattr(self, attr))}' + |
There was a problem hiding this comment.
| 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
There was a problem hiding this comment.
Won't work if times is a DatetimeStamp. Maybe that can't happen, idk.
I added the For the |
pvlib/tests/test_modelchain.py
Outdated
| @pytest.mark.parametrize('strategy, strategy_str', [ | ||
| ('south_at_latitude_tilt', 'south_at_latitude_tilt'), | ||
| (None, 'None')]) # GitHub issue 352 |
There was a problem hiding this comment.
Thanks for putting this test back. I think this decorator is no longer needed and can be deleted.
| 'aoi_modifier', 'total_irrad', 'spectral_modifier', | ||
| 'effective_irradiance', 'cell_temperature', 'diode_params', | ||
| 'dc', 'dc_ohmic_losses', 'losses', 'ac'] | ||
| mc_attrs = dir(self) |
There was a problem hiding this comment.
Using dir sorts the list of attributes alphabetically instead of conceptually. Not necessarily a downside.
There was a problem hiding this comment.
Yeah, once I saw the alphabetized list my thought was "that's easier to use"
|
Thanks @cwhanse! |
[ ] Tests addeddocs/sphinx/source/api.rstfor API changes.[ ] Adds description and name entries in the appropriate "what's new" file indocs/sphinx/source/whatsnewfor 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`).