Add binomial tree graph generator

[nahumsa]

Related to #150

Add binomial tree graph for a given order.

[coveralls]

Pull Request Test Coverage Report for Build 863756254

• 135 of 135 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.07%) to 96.849%

---

Totals
Change from base Build 858086935:	0.07%
Covered Lines:	7591
Relevant Lines:	7838

---

💛 - Coveralls

[mtreinish]

Is there a reason to make this optional if it's always required? Couldn't we change the order argument to be: order: usize, above and just remove this?

[nahumsa]

Sure. I was thinking to use weights to initialize as well, but was having some issues with rust's Option.
However, not making order optional would work.

[nahumsa]

I ran some benchmarks comparing to networkx:

import time
import retworkx as rx
import networkx as nx 


libraries = ["retworkx", "networkx"] # change for each branch
sizes = [5, 10, 20]

start = 0
stop = 0

for lib in libraries:
    
    for N in sizes:
        if lib == "retworkx":
            start = time.time()
            graph = rx.generators.binomial_tree_graph(N)
            stop = time.time()
        
        if lib == "networkx":
            start = time.time()
            graph = nx.generators.binomial_tree(N)
            stop = time.time()

        print(f"N = {N} {lib}: {str(stop - start)}")

Library	N=5	N=10	N=20
retworkx	1.5e-05s	0.000129s	0.076s
networkx	0.0198s	0.0075s	5.8s

[mtreinish]

Can we add some assertions here on the edge list. It would be good to assert that the output graph is actually a binomial tree instead of just a graph with 2**n nodes and 2**n - 1 edges

[nahumsa]

Added assertions to check if the edges match what is expected for a binomial tree graph

[mtreinish]

Looks good, just a couple of quick inline comments. I think this is almost ready.

[mtreinish]

Is there a reason to not do a multigraph flag here too?

[nahumsa]

Not really, I just did this because all other directed generators have this pattern. Should we add multigraph on all of them?

[mtreinish]

Ah, I didn't realize it was missing as an option from all the digraph generators. I think then it's fine to exclude it here for now, but we should push a follow up PR and add a multigraph option to all the digraph generators.

[nahumsa]

Sure, I'll raise an issue and submit a PR for that

[mtreinish]

LGTM, I found a couple more places to use is_none() but other than that I think this is ready.