Implement Variant.__repr__

[jeromekelleher]

The __str__ implementation of Variant looks nice but is quite unhelpful when you want to quickly inspect the actual values. We should implement a simple __repr__ that gives a string version of the raw values.

[charandeepsinghb]

I want to work on this issue. Please give me one chance. I have understood where str is used in Variant.

I am very keen to work in open source.
I want to learn and continue to work on tskit.
I'd really appreciate any kind of response to this comment.

Thank you

[benjeffery]

Hi @charandeepsinghb and thanks for your interest in tskit!

The aim is to add a __repr__ function to Variant that shows the genotypes. Something like:
return ','.join(self.alleles[gt] for gt in self.genotypes)

There is a development guide here: https://tskit.dev/tskit/docs/latest/development.html

[jeromekelleher]

Thanks for your interest in contributing to tskit @charandeepsinghb ❤️

I think we want something simpler than that @benjeffery, just something that shows the raw data held in the variant object literally, like say:

def __repr__(self):
     d = {
         "isolated_as_missing": self.isolated_as_missing,
         "site": self.site,
         "samples": self.samples,
         "alleles": self.alleles,
         "genotypes": self.genotypes,
    } 
    return f"Variant({repr(d)})"

I think this will do it? Hopefully this will give enough information for debugging without spewing out megabytes of text when the arrays are big, since:

>>> import numpy as np
>>> a = np.zeros(10000)
>>> a
array([0., 0., 0., ..., 0., 0., 0.])
>>> repr(a)
'array([0., 0., 0., ..., 0., 0., 0.])'

The Python docs for repr say:

Return a string containing a printable representation of an object. For many types, this function makes an attempt to return a string that would yield an object with the same value when passed to eval(), otherwise the representation is a string enclosed in angle brackets that contains the name of the type of the object together with additional information often including the name and address of the object. A class can control what this function returns for its instances by defining a repr() method.

So I think this is roughly OK?

[charandeepsinghb]

I'm really am sorry to say that I couldn't setup the project in the time available to me, maybe, also I can't set it up in my current system (non root environment to install and setup conda and stuff). I'm really sorry to disappoint you guys. It really meant a lot to me to hear anything from you. I'll for sure try to set it up in future and make some development if it will be possible. For now I can't work on this. although you explained the issue pretty well to me and I understand that how one would need to add a repr for quick inspection while debugging (because you guys explained so well). Thank you for your support.

[jeromekelleher]

No problem @charandeepsinghb, totally understood. Thanks for your interest anyway!

[chriscrsmith]

I'd like to work on this issue!

[chriscrsmith]

Opened PR #2695

I noticed if a site has experienced many mutations the repr() output is still quite large. Do you think leave it as is, or do something to abbreviate the number of mutations?