ACC memZono Release

[jonaswagner2826]

• [ ] Each m file should have a header function comment block in matching format
• [ ] Remove large blocks of commented out code in all files
• [ ] Level of commenting should match what is in transform.m

dimAwareFun.m

• [ ] Remove this from the ACC release

map.m

• [ ] Better comments
• [ ] Seems simpler to move code from linMap and put it into map directly...

plot.m

• [ ] I think we need to require dims to be passed in, otherwise we don't know what order to plot it in.
• [ ] The axis labeling is a nice convenience... but I could imagine scenarios where you are plotting multiple memZonos and their dim labels are not matching, even if they are plotted in the same space...

subsref.m

• [ ] the helper function getKeyIndices - does that do the same thing as a central helper function in memZono?... i.e., could we remove it and replace it with a call to something else?

supportFunc.m

• [ ] You'll have to explain this functionality to me a bit more. But I don't think this belongs in the initial release of memZono for ACC.

transform.m

• [ ] Resolve the @TODO mention... I think it can just be removed.

exampleMemoryReachability_memZono.m

• [ ] Looks good! Maybe just call the two implementations "method1" and "method2"?

reluNN_memZono.m

• [ ] how is this file in the examples directory related to the file in the @memZono directory?

SLAM.m

• [x] line 76 needs the keys specified per time/landmark: + memZono(RH + r_m{k,i},lDims(i),lKeys(k,i)); Looks like you had intended to do this... you had the variable lKeys and never used it :)

memZono.m

• [ ] Document constructor options
• [ ] Do we need the issym and isnumeric for this release?
• [ ] Label (with comment) which functions are not intended for external use, like get.Z_
• [ ] TODO on line 262?
• [ ] Line 291 - this returns keys not creates keys....
• [x] Line 429: "mtimes only overloaded for left multiplication"
• [x] What is needed to make vertcat efficient? (and sum, all)... the warning is not ideal.

[jonaswagner2826]

@jruths

Question answers:

map:

• [ ] Seems simpler to move code from linMap and put it into map directly...

This is nessicary for checks and then the eventual implementation of functional mapping (which is an a minimum a future release tasks if not ideally now)

Plot:

• [ ] I think we need to require dims to be passed in, otherwise we don't know what order to plot it in.

that was the original implementation... I think it is possible to do either but the motivation w/ this approach was to make checks easier (although idk by how much)

• [ ] The axis labeling is a nice convenience... but I could imagine scenarios where you are plotting multiple memZonos and their dim labels are not matching, even if they are plotted in the same space...

This is absolutely true... see SLAM plotting for an example... idk either way...

subsref:

• [ ] the helper function getKeyIndices - does that do the same thing as a central helper function in memZono?... i.e., could we remove it and replace it with a call to something else?

It is possible to reduce some of it (or make it a method or static function) but it is not currently replicated in full

SLAM:

• [ ] how is this file in the examples directory related to the file in the @memZono directory?

Unclear... I modified from what @JChenUTD had so I'm assuming that one is a duplicate... which should we keep?

memZono:

• [ ] Do we need the issym and isnumeric for this release? debatable... this is a fairly standard redefinition... along w/ size()
• [ ] What is needed to make vertcat efficient? (and sum, all)... the warning is not ideal.

We can eliminate the warning... this would require identifying all the shared dims/factors/cons between each combination and doing them in one operation... a bunch of combinatorics and not explicitly defined

Unsure of these requests:

• [ ] TODO on line 262?
• [ ] Line 291 - this returns keys not creates keys....