Hi Gary,
I'll start with your first mail:
> 1. we have two class hierachys ccp/format and ccpnmr/format/
> converters,
> what are the separation of concerns here? Does the first covers all
> the
> interaction with the file formats and the second handles all the
> interaction
> with ccpn data-model ?
The stuff in ccp/format does not use any API code. It parses the
files and prepares the information. The code in ccpnmr/format uses
this prepped info to create objects in the data model (because the
workflow is very similar if you're reading in, say, a sequence).
> 2. I was slightly confused if not decoyed by all the stuff in
> ccpnmr.format.converters.DataFormat.py specifically it builds a
> whole load
> of resonanceToAtom instances in setPeakAssignmentList (which it
> puts in
> self.peakAssignmentList) and then never seems to use them and in
> ccpnmr.format.converters.NmrViewFormat dips straight into the data-
> model. Is
> this because the NmrView conversion code predates this data
> structure and
> hasn't been updated yet?
It does use resonanceToAtom:
if self.resonanceToAtoms.has_key(resonance):
Basically what resonanceToAtom does is find the correct naming system
specific atom names for a particular resonance (along with all kinds
of other info).
> 3. what is the best way to go here? Is it to iterate over the data-
> model and
> use the peakContribs or to use the resonanceToAtoms in
> self.peakAssignmentList
This is in your next mail, I gather.
Wim
|