Re: [galaxy-dev] Multi-level datatype conversions
Hi David, On Jun 5, 2015, at 2:10 PM, David Kelly <davidkelly@uchicago.edu> wrote:
Hi Keith,
I am able to manually convert from a->b and then separately b->c, but not directly from a->c. I'm still thinking about how to approach this as well, but I'll share my thoughts so far.
Does Galaxy automatically do the conversions for you in the workflow editor? For example, if you have a tool A that produces format 'a' and another tool B that consumes format 'b' you should be able to drag the "noodle" from A to B and Galaxy will do an implicit format conversion. If this is working for you then someone from the Galaxy team will have to jump in. If Galaxy is not doing implicit format conversions then it likely means your datatype(s) are not fully defined. I didn't get the implicit format conversion working properly until I implemented the format sniffers.
In our use case, we have a collection of climate tools we'd like to use. The tools are written by different groups, so most use their own file formats. Ideally we would like users to be able to upload their data in any of these file formats and use all tools without having to explicitly convert their data.
Let's say there are 10 climate datatypes. My understanding is that in order to achieve this kind of interoperability, I'd have to write 90 converters.
If Galaxy could chain converters together, I could write 1 converter per datatype and chain them together like this: a -> b -> c -> d -> e -> f -> g -> h -> i -> j -> a
Another scenario I've been thinking about is picking one data type to be the 'standard'. Let's say it's 'a' in this example. You could write 18 converters: a->b, a->c, a->d,... b->a, c->a, d->a,...
This is what we do. We have a standard "pivot format" then converters from the other formats to/from the pivot format. Cheers, Keith
Then suppose you wanted to plug a 'd' file into a 'c' tool. You could first manually convert d->a, then use the automatic conversion for a->c. Not fully automatic, but not too bad.
On Thu, Jun 4, 2015 at 5:20 PM, Keith Suderman <suderman@anc.org> wrote: Hi David,
We just got it working for the single conversions, that is, getting Galaxy to automatically convert from a -> b when needed. Are you able to do the conversion, i.e a -> b and b -> c individually?
We are just coming to a similar use case and I'm definitely interested in how you make out, even if I can't be much help.
Cheers, Keith
On Jun 3, 2015, at 4:15 PM, David Kelly <davidkelly@uchicago.edu> wrote:
Hello,
Suppose I have three custom datatypes defined: a, b, and c
I have a datatype converter that converts from a -> b I have a datatype converter that converts from b -> c
Can I then send 'a' files as input into tools that accept type 'c'?
This doesn't seem to work in my test case, but I thought it might be worth asking to see if anyone has tried this before. Thanks!
Regards, David ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: https://lists.galaxyproject.org/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
------------------------------ Research Associate Department of Computer Science Vassar College Poughkeepsie, NY
------------------------------ Research Associate Department of Computer Science Vassar College Poughkeepsie, NY
participants (1)
-
Keith Suderman