OnACID and visualization from rolling buffer

[kushalkolar]

try the following:

• Create a circular shared memory rolling buffer SMRB. Maybe wrap this in an object with methods that allow you to "mark that a frame has been read", and "get the least recent unread frame".

• Dump frames into a SMRB for "raw data". Just read from SSD and dump them into the raw-data SMRB as fast as possible.

• From another python kernel, read from the SMRB and use OnACID to fit_frame() on each frame as they come in. Dump the OnACID mcorr result into another SMRB (different from raw data SMRB of course). The other results, A, C, etc. can just be stored in the database or whatever else we end up using

• From another python kernel, use fastplotlib to read from the raw SMRB and mcorr SMRB. Only this kernel should be used for "mark a frame that has been read" for both SMRBs.

[kushalkolar]

For using OnACID outputs, we can use the cnmf extensions that alresdy exist in mesmerize-core. We can probably just subclass the existing CNMFSeriesExtensions and override get_output(). Or we can create a common interface for mesmerize-core CNMF output API and serenity cnmf output. This will need more thought.