Re: [cclib-devel] github fork

[Clyde F. <cly...@gm...>]

Hi Noel,

I've changed the readme to specifically state it's an unofficial fork and
for those who wish to contribute to follow your link.

I'm trying to construct, submit to the server, collect from the server,
extract data from the subsequent files and analyse them entirely from
within an IPython notebook.Consequently I'm using several different python
quantum chemistry wrappers. Part of the reason I've put cclib into a git is
because I'd like to have all the codes I'm modifying in one code management
system and partly it's because I assumed some of the modifications that
would be useful to me would not necessarily be useful to the majority of
the package users.

In terms of cclib i've modified the units not because I think they are
wrong but simply because I want the three parsers I use to produce
consistent results. For my Gaussian calculations I use cclib to extract
data from the main body of the log file, I use ASE's gaussian reader (
https://trac.fysik.dtu.dk/projects/ase/browser/trunk/ase/io/gaussian_reader.py)
to extract data from the machine readable content at the end of the
logfile, and I use molmod's fchk reader (
https://github.com/molmod/molmod/blob/master/molmod/io/fchk.py) to extract
data from the .fchk point files.

At the moment ASE's gaussian_reader and cclib use a different set of units
(and gaussian itself uses a third set of units).

Of course I would be thrilled if I could convince cclib and ase to
standardize their units. In terms of any changes I make, I'm happy to push
them back to the official version if it's thought suitable. (Of course if
you were on github that would be easier :D)

Cheers

Clyde


On 8 November 2013 09:47, Noel O'Boyle <bao...@gm...> wrote:

> (ccing to the list)
>
> Hi Clyde,
>
> Thanks for letting me know. You're free to do this of course - might
> be good to make it clear that this is an "unofficial" fork though.
> Just to save people contacting us asking about ASE.
>
> Naturally for us we would prefer if you contribute any fixes directly
> to cclib. In this case, if any of our units are incorrect we would
> like to fix our source.
>
> This might be the push we need to consider moving to Github. Git on
> Windows though is a joke. The interface is not a lot better than the
> old CVS one (you probably don't remember this) which also used to call
> out to the command line. If only Github ran mercurial...:-/
>
> - Noel
>
>
>
> On 7 November 2013 15:05, Clyde Fare <cly...@gm...> wrote:
> > Hi Noel
> >
> > I've got a git repository with cclib in it because I am faffing about
> trying
> > to put together a bunch of python of tools associated with quantum
> chemistry
> > to use with the ipython notebook. In my version I've changed the units
> that
> > cclib uses in it's conversions because I want the various parsers I use
> to
> > all use the same units.
> >
> > As I wanted an easy way of loading the package into Anaconda I've put the
> > result on github here:
> > https://github.com/Clyde-fare/cclib
> >
> > Github  have a very fork-me friendly mentality but it occurred to me that
> > you might not be so happy that I'd done this so just wanted to check in
> with
> > you - I could write a version that means the units used are chosen based
> on
> > some flag (and your initial units are the default) which would mean I
> could
> > keep my version synced with yours if that's something you would prefer.
> >
> > Cheers
> >
> > Clyde
> >
>