Generalizing Closeness centrality to weighted networks using Newman method
#1385
Conversation
Signed-off-by: FedericoBruzzone <[email protected]>
|
|
FedericoBruzzone
changed the title
Closeness centrality to weighted networks using newman methodCloseness centrality to weighted networks using Newman method
rustworkx-core/src/centrality.rs
Outdated
| assert_eq!( | ||
| [ | ||
| Some(0.0), | ||
| Some(0.14285714285714285), |
There was a problem hiding this comment.
You probably want to compare the results using something like
rustworkx/rustworkx-core/src/centrality.rs
Line 538 in df38b0a
We should aim to have the most precision we can, but in practice minor changes to the algorithm can fudge non-significant digits of the float. Define an epsilon and check if |a - b| <= eps.
You might want a macro that loops over the arrays as well with zip as well.
Thanks for the review. I do not actually need the python implementation, but I can understand that this could improve usability for Python users. So, I will work on exposing the method to Python as well. Regarding the tests, I appreciate the feedback. I'll make sure to add checks that account for floating-point precision issues across all centrality metrics. |
Signed-off-by: FedericoBruzzone <[email protected]>
| #[test] | ||
| fn test_weighted_closeness_many_to_one_not_connected_2_digraph() { | ||
| let g = petgraph::graph::DiGraph::<u32, f64>::from_edges([ | ||
| (1, 0, 0.1), | ||
| (2, 0, 0.1), | ||
| (3, 0, 0.1), | ||
| (4, 0, 0.1), | ||
| (5, 0, 0.1), | ||
| (6, 0, 0.1), | ||
| (7, 0, 0.1), | ||
| (1, 7, 1.0), | ||
| ]); | ||
| let c = newman_weighted_closeness_centrality(&g, false, |x| *x.weight()); | ||
| let result = [ | ||
| Some(0.1), | ||
| Some(0.0), | ||
| Some(0.0), | ||
| Some(0.0), | ||
| Some(0.0), | ||
| Some(0.0), | ||
| Some(0.0), | ||
| Some(1.0), | ||
| ]; | ||
|
|
||
| assert_eq!(result, *c); | ||
| } |
There was a problem hiding this comment.
IMO this test should keep assert_eq! predicate but let me know :D
FedericoBruzzone
force-pushed
the
closeness-wgraph
branch
from
1fb850f to
b364913
Compare
|
@IvanIsCoding Make sure that the following commit is in line with the rest, but most importantly, make sure it is correct. Make available weighted_closeness for python users Why is CI not starting? 🤔 |
Signed-off-by: FedericoBruzzone <[email protected]>
Signed-off-by: FedericoBruzzone <[email protected]>
FedericoBruzzone
force-pushed
the
closeness-wgraph
branch
from
81813f3 to
7de1312
Compare
|
I need to manually approve CI. Also, I asked about Python just to confirm. Someone else can work on the Python bindings. |
There was no problem to implement it. It was a pleasure, I look forward to your review :D |
Signed-off-by: FedericoBruzzone <[email protected]>
FedericoBruzzone
force-pushed
the
closeness-wgraph
branch
from
0bc11f9 to
95c0166
Compare
Close #1384