Fixed reprs to match 0.15 behaviour

[richardjgowers]

Fixes #1221

Changes made in this Pull Request:

PR Checklist

• Tests?
• Docs?
• CHANGELOG updated?
• Issue raised/referenced?

[orbeckst]

See also #1223 (comment) before merging... maybe keep group reprs simpler than before but make sure that list(group) (and array(group)) show the representations of the individual elements.

[richardjgowers]

@orbeckst this restores old behaviour, so can we merge as is? I did think that #1223 might be a good starter issue for someone (who doesn't have mountains of docs still to write)

[jeiros]

Hi,

I'm happy to finish up this PR. If I understood this correctly, what's missing is defining a __str___ method in the GroupBase class that is smart and only prints the first and last three elements when it's longer than 10?

[richardjgowers]

Yes relative to current develop branch

…

On Wed, 1 Mar 2017, 5:30 p.m. Juan Eiros, ***@***.***> wrote: Hi, I'm happy to finish up this PR. If I understood this correctly, what's missing is defining a __str___ method in the GroupBase class that is smart and only prints the first and last three elements when it's longer than 10? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1222 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AI0jB7P0GeEnUoGNkj_5_-ZvZkDLF5bZks5rhatCgaJpZM4MOay8> .

[jeiros]

I've got something that is sort of working, implemented in the GroupBase class. It's like this:

def __str__(self):
    name = self.level.name
    if len(self) <= 10:
        return '<{}Group {}>'.format(name.capitalize(), repr(list(self.level)))
    else:
        return '<{}Group {}, ..., {}>'.format(name.capitalize(),
                                              repr(list(self.level)[:3])[:-1],
                                              repr(list(self.level)[-3:])[1:])

My guess is something like this is what you want:

repr(rg)
>>> '<ResidueGroup with 10 residues>'
str(rg)
>>> <ResidueGroup [<Residue MET, 1>,
 <Residue ARG, 2>,
 <Residue ILE, 3>,
...,
 <Residue ALA, 8>,
 <Residue PRO, 9>,
 <Residue GLY, 10>]>

So that the repr from the individual elements that form the group are printed out.
But what would be the equivalent of self.segments or self.residues in the GroupBase class? Or is that not accessible from there, since ResidueGroup, AtomGroup and SegmentGroup inherit from it?

[jeiros]

Actually, changing self.level to self does that, I think. Since list(self) gives the repr of its elements.

[richardjgowers]

Yeah that looks like a good start. Tests too!

…

On Wed, 1 Mar 2017, 6:17 p.m. Juan Eiros, ***@***.***> wrote: I've got something that is sort of working, implemented in the GroupBase class. It's like this: def __str__(self): name = self.level.name if len(self) <= 10: return '<{}Group {}>'.format(name.capitalize(), repr(list(self.level))) else: return '<{}Group {}, ..., {}>'.format(name.capitalize(), repr(list(self.level)[:3])[:-1], repr(list(self.level)[-3:])[1:]) My guess is something like this is what you want: repr(rg)>>> '<ResidueGroup with 10 residues>'str(rg)>>> <ResidueGroup [<Residue MET, 1>, <Residue ARG, 2>, <Residue ILE, 3>,..., <Residue ALA, 8>, <Residue PRO, 9>, <Residue GLY, 10>]> So that the repr from the individual elements that form the group are printed out. But what would be the equivalent of self.segments or self.residues in the GroupBase class? Or is that not accessible from there, since ResidueGroup, AtomGroup and SegmentGroup inherit from it? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1222 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AI0jB9IzTfTSHHfNOLWpa3_mjYwONJ8tks5rhbWWgaJpZM4MOay8> .

[orbeckst]

good – restoring 0.15.0 as far as I can tell; we'll then move forward with #1223 using this as a starting point.

[orbeckst]

We are probably going to go back to the terse representation, see #1223 and comments therein.

[orbeckst]

Will need similar tests again for #1223.

[orbeckst]

@jeiros thanks for taking this on. Please continue discussion at #1223 ; once you have a PR referencing #1223 , most of the detailed discussion will happen on the PR.

[jeiros]

I actually just have finished writing the tests, where should I do the PR so you can start looking at it?

[jeiros]

Ah sorry, I see. I'll do a new PR referencing #1223