cannot access MCAs from Xspress3 when scanned

[tangkong]

Describe the bug Running scan

To Reproduce Steps to reproduce the behavior:

1. Initialize Bluesky ipython profile
2. RE(bp.scan([xsp3], -1, 1, num=3))
3. See lack of files written

Expected behavior files should be written with file numbers appended

Steps Look to compare Dexela behavior, which increments file number correctly during scans.

[tangkong]

Requires {P}:HDF5:FileNumber to be incremented, which occurs every time {P}:HDF5:Capture to be switched on and off. This can be done in the main detector trigger method, but does not allow access to these files quite yet.

Need to communicate these files to the datum documents properly.

[tangkong]

Easier solution is actually to fully leverage the "stream" mode of the HDF5 File store plugin, and write all frames from one scan into one h5 file. This involves some communication that was un-done by previous efforts.
We require:

• {P}:NumImages * (number of points in scan) = {P}:HDF5:NumCapture. The HDF5 plugin keeps a file open until {P}:HDF5:NumCaptured_RBV increments to equal {P}:HDF5:NumCapture_RBV. Setting this value high enough allows acquire to be toggled on and off with additional frames being added to the file.
• {P}:NumImages = 1: multi-spectra scans not supported yet.

HDF5 handler only expects one file, and will pull each frame out and fill tables as needed.

Plans need to update xs_det.total_points to match with the number of points in a scan before the scan is started, requiring some light wrapper plans.