modify take_image() to take N frames per acquisition

[prjemian]

Processing of the take_image() plan might be too slow to push the limits of the pipeline. Add an n frames per acquisition feature.

[prjemian]

Consider compression.

[prjemian]

Note: $(P):cam1:AcquirePeriod becomes important now.

    0 < AcquireTime<= AcquirePeriod

PV	setting
$(P)cam1:AcquireTime	acquisition time per frame
$(P)cam1:AcquirePeriod	time between frame starts
$(P)cam1:NumImages	number of frames per acquisition
$(P)cam1:NumExposures	(? leave at 1 ?) -- probably unused in ADSimDetector
$(P)cam1:ImageMode	Multiple
$(P)cam1:WaitForPlugins	Yes
$(P)HDF1:NumCapture	same as $(P)cam1:NumImages (or less if images are dropped)
$(P)HDF1:Compression	Blosc
$(P)HDF1:FileWriteMode	Capture

[prjemian]

Should add compression as an optional parameter. Default=None

[prjemian]

Why is it that when HDF Capture is moved to the end of the stage_sigs as:

adsimdet.hdf1.stage_sigs.pop("capture")
adsimdet.hdf1.stage_sigs["lazy_open"] = 1
adsimdet.hdf1.stage_sigs["compression"] = "Blosc"
adsimdet.hdf1.stage_sigs["file_template"] = "%s%s_%3.3d.h5"
adsimdet.hdf1.stage_sigs["capture"] = 1

that the resource_path is set with {write_path}/{short_uid}_-001.h5, as in:

DIAGNOSTIC: r = {'spec': 'AD_HDF5', 'root': '/', 'resource_path': 'tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/23ef3149-d0a6-4b32-b162_-001.h5', 'resource_kwargs': {'frame_per_point': 1}, 'path_semantics': 'posix', 'uid': '9abf6561-8b80-45b8-976a-6cb8f52ccc19', 'run_start': '71d0360d-ca64-487d-9f79-b1c567be0bf2'}

When the capture stage_sig is not moved, then is resource_path = {write_path}/{short_uid}_000.h5

[prjemian]

The HDF file that is written is {write_path}/{short_uid}_000.h:

In [1]: !ls /tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/23ef3149-d0a6-4b32-b162_-001.h5
ls: cannot access /tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/23ef3149-d0a6-4b32-b162_-001.h5: No such file or directory

In [2]: !ls /tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/23ef3149-d0a6-4b32-b162_000.h5
/tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/23ef3149-d0a6-4b32-b162_000.h5

[prjemian]

Perhaps the assumption about the resource_path (from the resource document) is not correct. The problem is that the -001 might actually be -1 when the file_template = "%s%s_%3.3d.h5" is applied:

In [9]: "_%3.3d" % -1
Out[9]: '_-001'

Rather, the resource_path is from a template and the actual file name must consider content from the datum document for the image event. See this example:

[('resource',
  {'spec': 'AD_HDF5',
   'root': '/',
   'resource_path': 'tmp/docker_ioc/iocbdpad/tmp/adsimdet/2022/04/07/de0c26a1-d3f1-42a5-a77d_-001.h5',
   'resource_kwargs': {'frame_per_point': 1},
   'path_semantics': 'posix',
   'uid': 'd1b8a908-2a97-49d2-b2b7-2a0ef5acc159',
   'run_start': 'fcab67c5-0e7f-4720-a670-c585d2dc0ba0'}),
 ('datum',
  {'datum_id': 'd1b8a908-2a97-49d2-b2b7-2a0ef5acc159/0',
   'datum_kwargs': {'point_number': 0},
   'resource': 'd1b8a908-2a97-49d2-b2b7-2a0ef5acc159'})]

[prjemian]

Get the file name from the area detector HDF5 plugin:

In [10]: adsimdet.hdf1.full_file_name.get()
Out[10]: '/tmp/adsimdet/2022/04/07/de0c26a1-d3f1-42a5-a77d_000.h5'

and then substitute between the AD IOC and Bluesky mount points.

[prjemian]

This code produces the correct HDF5 file as seen from the bluesky computer:

        path = pathlib.Path(r["root"]) / r["resource_path"]
        fname = pathlib.Path(adsimdet.hdf1.full_file_name.get())
        hdffile = path.parent / fname.name