[PyWrapper-devel] FW: PyWrapper and Species 2000
Status: Alpha
Brought to you by:
jatorre
From: <m.d...@BG...> - 2006-11-10 09:59:23
|
-----Original Message----- From: D=F6ring, Markus=20 Sent: Freitag, 10. November 2006 10:08 To: 'Peter Brewer'; Torre de la, Javier Subject: RE: PyWrapper and Species 2000 Peter, We are happy to hear that interest from species2000. The current version = of the pywrapper that works with TAPIR and BioMOBY unfortunately doesn't = implement SPICE at all. But I don't hink this will be hard to add (I = will come back to it later). The previous version, pywrapper 1.x, that = run with biocase supported SPICE. So I know pretty well how the = specification works. With TAPIR there are more powerful configurations possible, mainly = through the use of "output models". Using this for SPICE will make it = much easier to implement it. The only thing we would need is an up to = date XML Schema. Then we will have to create a separate output model for = each request-type - we could entirely reuse the SPICE schema. Then there would need to be a small and newly developed "protocol" layer = for SPICE in pywrapper. That also shouldn't be too bad. Overall I would = expect something like 2 weeks for the implementation and another week = for testing and configuring a database like the berlin model. The only real problem with having a generic and configurable wrapper is = that taxonomic data is build on tree hierarchies. This is often = represented as self-references in databases but not always. SO inorder = to "map" your local database to the SPICE schema it usually takes quite = some transformation which you often have to do with views, cause generic = wrappers like pywrapper cant handle this (you cannot simply add some = custom SQL). So all I wanted to say here is that there still remains a = bit of the complexity. But you can be sure the wrappers do understand = SPICE correctly and respond properly - just the data they carry might = not be making sense if they were badly configured. Anyway. I would be very happy if we can make pywrapper SPICE compliant = again. Unfortunately there is not much manpower left from Javier and = myself this year, so we could only do little things. But we would be = very happy to assist someone trying to get into the python code and do = it himself. Its not that hard. Would that be a possibility? With thanks, Markus > -----Original Message----- > From: Peter Brewer [mailto:p.w...@re...] > Sent: Freitag, 10. November 2006 09:32 > To: D=F6ring, Markus; Torre de la, Javier > Subject: PyWrapper and Species 2000 >=20 >=20 > Dear Markus and Javier, >=20 > I'm currently at a Species 2000 meeting in Amsterdam and one of the > issues that has been raised is that of assisting Global=20 > Species Database=20 > managers to provide wrappers for their databases. As the new Species=20 > 2000 Systems Manager that responsibility apparently falls on=20 > my shoulders! >=20 > It seems clear to me that the best approach is for me to > assist with the=20 > development of PyWrapper by adding support for the SPICE protocol. >=20 > I've downloaded your latest development snapshot and I'm very > impressed=20 > - especially with the excellent level of documentation! It's such a=20 > relief to come across a piece of scientific software with a=20 > installation=20 > manual more than 3 lines long. :-) >=20 > What is the current level of support for SPICE in the PyWrapper? Has > anything been done yet? I'd be quite surprised if there has=20 > because the=20 > level of documentation of SPICE is very poor - something I'm=20 > also going=20 > to work on over the next couple of months. >=20 > I'll get in touch again next week once I get back to Reading > and hope to=20 > make progress with this quite quickly. >=20 > Best wishes >=20 > Peter Brewer >=20 > =20 >=20 >=20 >=20 |