[ome-devel] Python import API

[Niko Ehrenfeuchter]

Hi Colin,
On 01.03.2016 15:28, Colin Blackburn (Staff) wrote:
> Hi Niko,
>
>> On 1 Mar 2016, at 14:01, Niko Ehrenfeuchter
>> <nikolaus.ehrenfeuchter at unibas.ch
>> <mailto:nikolaus.ehrenfeuchter at unibas.ch>> wrote:
>>
> [...]
>>> Is this the sort of thing you are after or are you looking at getting
>>> these IDs at some other point in your workflow?
>>
>> This is kind of the thing I am after, yes. I was just wondering if it
>> didn't make sense to have this "officially" supported / exposed via
>> the API, instead of everyone having to write their own code to carve
>> out the desired information from stdout/stderr - which feels quite a
>> bit wiggly, especially as I don't necessarily consider "console
>> output" as being an "API stable" format, right?
>
> I have since looked at the email conversation you had with Josh back in
> June last year:
>
> http://lists.openmicroscopy.org.uk/pipermail/ome-devel/2015-June/003336.html
>
> and yes, there should be some API support for this.

right, that's exactly my starting point. Thanks for referring to it, I 
should have done this myself.

>> Plus, the referenced code is using the "capfd" fixture for the tests,
>> which probably isn't a good idea in a non-testing environment. Meaning
>> again custom code which has to be written "client side" to capture the
>> outputs, not unlikely resulting in bugs, from time to time.
>
> Absolutely, cafd was an illustration but other ways of doing this are
> not ideal.
>
>> Basically, I'm wondering if there is no proper (TM) way using all the
>> Ice'ish magic to "cleanly" retrieve this information ;-)
>
> So at the moment the answer is no (as far as I know!) but I have created
> an, as yet internal, Trello card for discussion with Josh and others.
> I'll push that card to a public board once we have had a chance to chat.
> Hopefully we can then prioritise the work and come up with a solution.

That's great news, thanks. I *really* appreciate this!

Cheers,
~Niko