Calling methods using parent class

Question

I'm very new to object-oriented programing, and I'm having trouble wrapping my head around how to manage the relationship between classes that have related data.

I'm trying to represent some genomic information about a series of organisms, which I am eventually going to compare. So I have a Species class:

class Species(object):
    def __init__(self, species_name, genes={}, contigs={}):
        self.species_name = species_name
        self.genes = genes
        self.contigs = contigs

Then I'd like to have a Gene class,

class Gene(object):
    def __init__(self, locus_tag, annotation, dna_seq, aa_seq):
        self.locus_tag = locus_tag
        self.annotation = annotation
        self.dna_seq = dna_seq
        self.aa_seq = aa_seq

The idea is that the genes method of Species will be a dictionary of Gene objects, where the key is the locus_tag of the Gene . I think I know how to implement this part.

But one of the things I'd like to be able to do is call a function on a Gene object that requires knowing what Species it's in (it would output a file structure that includes the species_name and locus_tag , but these are coming from two different classes. The easiest thing I can think of would be to add a species_name directly to the Gene class, but since each Gene will be within a Species , this seems redundant.

I'm not even sure where to begin - I did quite a bit of searching, but the other questions I've found are either not relevant, or I don't understand enough to grasp the relevance.

Answer 1

Create the Gene objects, then create the Species object by passing the dictionary for genes and then update the Gene objects by setting the Species instance to it, like so:

for gene in genes:
    gene.species = species

After that each Gene instance ( gene ) can refer to its Species object by doing self.species or, for example, self.species.species_name to get the name.

Answer 2

I would generally not do what you've described. It creates a circular reference, which should be avoided where possible. Objects should generally have a hierarchical relationship. You should be try to find a way to keep a gene and a species reference together where you might need to serialise it.

To be able to serialise a gene you could write a function that takes a species and a gene as arguments. eg.

# if using python 2 (I think you might be)
from __future__ import print_function

class Species(object):
    def __init__(self, species_name, genes={}, contigs={}):
        self.species_name = species_name
        self.genes = genes
        self.contigs = contigs

    def serialise(self, fileobj):
        for gene in self.genes.values():
            self.serialise_gene(gene, fileobj)

    def serialise_gene(self, gene, fileobj):
        print(self.name, gene.locus_tag, sep=", ", file=fileobj)
        # prints "name, locus_tag" to the file

s = Species("homo erectus")
with open("some_file.txt", "w") as f:
    s.serialise(f)