feature/DiskDataChanges

[m053m716]

Changes to nigeLab.libs.DiskData

Key Changes

• nigeLab.libs.DiskData re-work
  • Now, clear reason for three obj.type_ categories:
    • 'MatFile' : Files that should not grow or shrink once they have been allocated on the disk (such as from the already-known number of samples in the record)
    • 'Hybrid': "Vector" style file that can be expanded along columns. This should be used for data streams such as RawData or DigIO.
    • 'Event': "Matrix" style file that can be expanded along rows. This should be used for things like 'Spikes' or 'ScoredEvents'.
  • Split subsref and subsasgn methods of nigeLab.libs.DiskData into more modular design to make it clearer what logic is in those methods.
  • Initialization for all three types of input arguments (single, two inputs, or three inputs) should all work now (previously, it didn't seem the single-input case worked).
  • Now no longer creates a matfile (.diskfile_); instead, uses hdf5 handling for everything.
  • Still (initially) saves using matfile (save(__,'-v7.3')) so that the proper Matlab header is created.
• Bug with doRawExtraction should be resolved
  • May need to re-run extraction for STREAMS (it was incorrectly using 8-bit values that resulted in empty streams which otherwise contained data)
  • Added parsing of FSM DAC values:
    • amplifier channel
    • window start sample
    • window stop sample
    • voltage threshold
    • window type
  • Re-worked how Stim is parsed, in general.

---

Other Changes

• nigeLab.defaults

  • Add flag to skip UI Prompt for videos in +defaults/Videos
  • Update +defaults/doActions so each struct field that corresponds to a doMethod has a third field: batch
    • This field is checked for example if it is expected that many important Blocks may need to have a particular field (such as .Videos) but that not every Block in a Tank is expected to have it. Instead of throwing an error, checkActionIsValid will return a flag as false if the .Status of .Fields listed in .batch are not completed.

• nigeLab.libs

  • Update .empty() methods of .VidStreamsType and .VideosFieldType
  • nigeLab.libs.VidStreamsType and nigeLab.libs.VideosFieldType should now initialize correctly.
    • Unfortunately, once they started initializing correctly, it caused the _Obj.mat files to expand in size by an order of magnitude. Re-worked how data pointers are handled for these classes, so they should no longer be as inefficient

• nigeLab.Block

  • Update a bunch of the Block methods that referenced deprecated parameters fields (VideoPars and SortPars).
  • Remove deprecated methods:
    • nigeLab.Block/getFileType()
    • nigeLab.Block/getFieldType()
    • nigeLab.Block/getFieldTypeIndex()
  • Corresponding methods are implemented in nigeLab.nigelObj
  • nigelab.Block/doVidInfoExtraction should now work more-or-less as intended.
    • Fixed bad problem causing extracted VidStreams to make all nigelFiles grow by order of magnitude (oops).
  • nigeLab.Block/getStatus() this method is a mess and needs to be moved to nigelObj soon. Basically, we want it to behave the same way when we call it at all levels otherwise it gets too confusing.
    • For now, just put some shitty code in there to make it work for the time-being.
  • Fix initEvents
  • Fix doEventDetection and doEventHeaderExtraction

• nigeLab.Tank

  • Add varargin to nigeLab.Tank/runFun() method, which allows it to supply input

Other Stuff

• nigeLab.nigelObj/getFieldInfo() and nigeLab.libs.nigelButtonArray
  • Were supposed to work in combination to make mouse-over and clickable things happen for DashBoard but I think it's too messy and didn't want to waste time on it. Regardless, they exist (and that's why).
• Added .empty methods to a few other Classes so their validation doesn't screw up other working classes.