Happy new year!
The reading of the PDC files is not implement yet as we had to work
out which sections and information were required. Now that most
things are sorted out for the relaxation times with the error scaling,
time to rate conversion, etc., relax will most likely just read this
section. I.e. walk to the results 'SECTION" and then read that
converting to relaxation rates and removing the scaling. How is the
peak name defined though? What about side chain TRP NH peaks? This
should not take long at all, though Michael and I are quite busy
finishing the relax GUI at the moment. Additional information would
be parsed from the PDC header and 'sample information' section. I'll
answer more of your questions below:
On 24 December 2010 07:58, Dr. Klaus-Peter Neidig
<peter.neidig@xxxxxxxxxxxxxxxxx> wrote:
Dear Edward & Michael,
I wish you some nice holidays and a good start into the new year.
The last time I had sent you examples of the PDC ouput. I would assume that
you can easily
read it via Python and that your Python programs are flexible. I will make
sure that the format
stays in its essential components, i.e. values are tab separated and the
sections start with
the keywords used so far.
May I ask if you in what detail you read other keywords ?
Example:
The section with the integrals looks like:
SECTION: used integrals
Mixing time [s]: 0.01 0.03 .........
Would you suffer if I changed Mixing time into Mixing times ? Or do you
just read a token up to : ?
We wouldn't need to read this section in relax as we know how to
convert the relaxation values and errors to the correct form. But to
find the relevant data, a fixed tag name would be useful. relax would
be able to handle changes, as long as the PDC file has version
information in it and the change occurs between different versions.
Then we can search for 'Mixing time [s]' in version 'x' and 'Mixing
times' in version 'y'.
The reason why I'm asking that someone looked through export and report and
found such minor
differences. Other examples are that in between the tab separated title
lines of the tables, e.g.
the title line Peak name F1[ppm] F2[ppm] ....
sometimes additional spaces are contained,
sometimes 2, sometimes 3.
In relax, the amount of spacing does not matter. It should not matter
- the only case where it does is the retarded PDB format. Parsers
should be able to handle this if written properly, just split the line
up by whitespace.
Do you evaluate these title lines at all or just the relevant table that
follows ?
This title line would be skipped. Can it be that the user can change
which columns are shown? For example can they rearrange these
columns, or say not show the F1 and F2 columns? If these columns are
variable, then we would have to read that line.
Early next year we should write down which parts of the output must no be
changed by myself
to avoid conflicts.
It would be nice to have these files versioned so that automatic
parsers can handle any future changes. That would make things much
easier.
Cheers,
Edward
Re: Happy holidays