On Fri, Dec 5, 2008 at 8:19 PM, Sébastien Morin
<sebastien.morin.1@xxxxxxxxx> wrote:
Hi Edward,
I agree with you. I had not think about all the other files in relax
since I was so focused on the 'generic' file... but your idea of
implementing this at a higher level is quite logical and would give much
more flexibility to relax. Maybe we could create a branch for this, as
you proposed...
Concerning the generic format, what do you think we should do ? Should
we just introduce variables so the user tells relax in which column
things are..? The user could also tell that he uses the generic format,
as well as how many header lines there are, etc... In fact, the user
could, for now (until we get the automatic recognition stuff working),
tell everything concerning the formatting of his file... What do you think ?
I think it would be best to do what all the other code does for now
and require the user to specify all columns manually. That could be
the test for the generic format for now, that some of the columns have
been specified, so that we can still automatically determine the file
format. This automatic format determination is specific to the
generic_fns.spectrum code and not that special data object I
previously talked about. In the future, the test could be to pass the
file into this data object as a last resort if no other format is
found, and if a RelaxError occurs, catch it and throw a different
RelaxError stating that the automatic file format determination
failed.
Lots of work ahead !
Like I said, this is not a simple task ;)
Regards,
Edward
Re: relax formatted files for relax reading and writing.