Re: [igraph] python-igraph 1.0?

[Tamas Nepusz]

Dear Alex,

So I've had some time to look into the situation a bit. It seems like
there was no official release of the C interface since 0.7.1, so in
practice, the latest-and-greatest stable version of the C interface
still contains the "old" Fruchterman-Reingold layout implementation.
That's a bummer because I cannot release a new version of the Python
interface with the new implementation, since the Python interface
always links to the C core of igraph that is on the user's machine (or
downloads and compiles the latest stable version of the C core if the
user did not install the C core yet).

On the other hand, the development version of the Python interface
(the one that you can get from Github) has already been updated to use
the new FR layout and not the old one. So, if you check out the source
code of both the C core and the Python interface from Github and
compile them, you should already have a version with the new FR
layout.

The other improvements in the C core that I could identify were:

[+] the Davidson-Harel layout (see Graph.layout_davidson_harel)
[-] the GEM layout
[+] the mindist argument of the neighborhood functions
[-] the igraph_all_simple_paths function
[-] the igraph_triangles function
[+] the random walk generator (see Graph.random_walk)

The ones that I marked with [+] are already in the development version
of the Python interface; the remaining ones are waiting to be ported.

All the best,
T.

T.


On Fri, Sep 11, 2015 at 3:52 PM, Tamás Nepusz <address@hidden> wrote:
> Dear Alex,
>
> I am a bit short on spare time recently, so I won't be able to release a 
> full, updated version of the Python interface for a while (I did not even 
> have the time to figure out what the major changes are between 0.7.1 and 
> 1.0), but I'll try to prepare a patched version for you in a few days that 
> includes at least the new FR layout. Stay tuned.
>
> T.
>
>> On 11 Sep 2015, at 07:43, Alex A <address@hidden> wrote:
>>
>> Dear Tamas,
>>
>> Any idea when the improvements mentioned in R/igraph 1.0 are coming to 
>> python-igraph?  In particular, I'm excited about the "much faster" 
>> implementation of the Fruchterman-Reingold layout.
>>
>> Thank you,
>> Alex
>> _______________________________________________
>> igraph-help mailing list
>> address@hidden
>> https://lists.nongnu.org/mailman/listinfo/igraph-help