change timeseries "format" kw to "order"

[orbeckst]

In various issues and discussions we sort-of decided that the Reader.timeseries() method should not use format="afc" as a kwarg but order="afc" (which make it consistent with the MemoryReader, avoids clashes with the builtin format, ... and avoids clashes with the format kwarg of Universe (yes, that one clashes with the builtin, too)).

[orbeckst]

I think we should consider adding this to 0.17 so that we have some deprecation time.

[orbeckst]

Note to myself: While playing around with the memory reader for the Examples for RMSF I found that not all orderings (keyword order=...) work for Universe.load_new(array_like, order="..."). At least I got a cryptic error with fac (something that there were too few dimensions) when I tried a array_like for a single frame. It worked with afc, which I put explicitly in the example (but it is also the default for e.g. DCReader.timeseries).

Need to investigate more and possibly add tests for all permutations when reading with the MemoryReader.

[orbeckst]

Also note that the docs for the MemoryReader need to be update:

• When a array_like is provided then it is automatically passed to the MemoryReader and the format="MEMORY" or format=MemoryReader is not needed anymore.
• Same for Universe.load_new()

In general we need to clarify how file types are recognized.

[orbeckst]

I'll try to do this in my ample spare time ;-) – unless anyone else wants to: feel free to snatch this issue from me.

[kain88-de]

@orbeckst do you have any code for your experiments of order that you did in august.

[orbeckst]

No, sorry. But you can take the Example at https://www.mdanalysis.org/mdanalysis/documentation_pages/analysis/rms.html#MDAnalysis.analysis.rms.RMSF and just change order in the line

# make a reference structure (need to reshape into a 1-frame "trajectory")
reference = mda.Merge(protein).load_new(
            reference_coordinates[:, None, :], order="afc")

and it should fail.

(I am travelling and if you want to work on this please by all means do so and also assign yourself/un-assign me. Thanks!)

[kain88-de]

Your example case works now with my changes.

[orbeckst]

@kain88-de many thanks for finishing the PR, I very much appreciate you jumping in. Sorry I wasn't able to be more helpful.

[kain88-de]

It’s ok. There is a tone of stuff I’m not getting done on MDAnalysis myself right now.

…

On 5. Jan 2018, at 16:19, Oliver Beckstein ***@***.***> wrote: @kain88-de many thanks for finishing the PR, I very much appreciate you jumping in. Sorry I wasn't able to be more helpful. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.