fix: Fix direction of edges when restoring - #805

This appears to work also with igraph/igraphdata#5.

How can we test correct restoration in this repo?

On the current main branch:

options(conflicts.policy = list(warn = FALSE))
library(igraph)

g <- make_ring(3, directed = TRUE)
g
#> IGRAPH a6d007c D--- 3 3 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l)
#> + edges from a6d007c:
#> [1] 1->2 2->3 3->1

gs <- unserialize(serialize(g, NULL))
gs
#> IGRAPH a6d007c D--- 3 3 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l)
#> + edges from a6d007c:
#> [1] 2->1 3->2 1->3

^{Created on 2023-05-21 with reprex v2.0.2}

I also created small C example to play around with this in igraph. It should be from and to. Originally I was confused by the order in igraph_add_edges function.

#include <igraph.h>

void print_vector_int(igraph_vector_int_t *v) {
    igraph_integer_t i, n = igraph_vector_int_size(v);
    for (i = 0; i < n; i++) {
        printf(" %" IGRAPH_PRId, VECTOR(*v)[i]);
    }
    printf("\n");
}


int main(void) {

    igraph_t g;
    igraph_vector_int_t v1, v2;

    /* simple use */
    igraph_vector_int_init(&v1, 8);
    VECTOR(v1)[0] = 0;
    VECTOR(v1)[1] = 1;
    VECTOR(v1)[2] = 1;
    VECTOR(v1)[3] = 2;
    VECTOR(v1)[4] = 3;
    VECTOR(v1)[5] = 2;
    VECTOR(v1)[6] = 2;
    VECTOR(v1)[7] = 2;
    
    igraph_bool_t directed = 1;

    igraph_create(&g, &v1, 0, directed);

    igraph_vector_int_t v;
    igraph_integer_t i, s=igraph_vector_int_size(&g.from);
    igraph_vector_int_init(&v, s*2);

    for (i = 0; i < s; ++i) {
        igraph_integer_t from_v=VECTOR(g.from)[i];
        igraph_integer_t to_v=VECTOR(g.to)[i];
        
        igraph_vector_int_set(&v, i*2, from_v);
        igraph_vector_int_set(&v, i*2+1, to_v);
    }

    print_vector_int(&g.from);
    print_vector_int(&g.to);

    print_vector_int(&v);

    igraph_t g2;
    igraph_empty(&g2, g.n, directed);
    igraph_add_edges(&g2, &v, NULL);

    print_vector_int(&g2.from);
    print_vector_int(&g2.to);

    igraph_destroy(&g);
    igraph_destroy(&g2);
    igraph_vector_int_destroy(&v);

    return 0;
}

Working on a test here.

@Antonov548: The test failures are interesting. Can you replicate on Ubuntu? What happens here?

https://github.com/igraph/rigraph/actions/runs/5038508208/jobs/9036118195?pr=805#step:6:205

I can't replicate on Ubuntu.

I'll remove the snapshot test for now and merge.

I'll remove the snapshot test for now and merge.

It's very strange. But I can replicate it time to time. Sometime it prints edges some time not. I there any function which is responsible for this printing?

Yes, weird. We need a reprex first. Does it fail on your machine when you print in a loop? How many iterations do you need?

Yes, weird. We need a reprex first. Does it fail on your machine when you print in a loop? How many iterations do you need?

For me seems it failing when I don't watch how tests are running. I mean I just collapse VS code window 😕
I don't know if it really could be connected.

What version of R are you running? I'm not seeing failures with R 4.3.0 on Ubuntu.

What version of R are you running? I'm not seeing failures with R 4.3.0 on Ubuntu.

I also have 4.3.0

I see it coming from print.R file. And I see that printing of edges has some conditions. I will try to go over this code.