File Writers#

The file writer callbacks are:

FileWriterCallbackBase(*args, **kwargs)

Base class for filewriter callbacks.

NXWriterAPS(*args, **kwargs)

Customize NXWriter with APS-specific content.

NXWriter(*args, **kwargs)

General class for writing HDF5/NeXus file (using only NeXus base classes).

SpecWriterCallback([filename, auto_write, ...])

Collect data from Bluesky RunEngine documents to write as SPEC data.

Overview#

Each filewriter can be used as a callback to the Bluesky RunEngine for live data acquisition or later as a handler for a document set from the databroker. Methods are provided to handle each document type. The callback method, receiver(document_type, document), receives the set of documents, one-by-one, and sends them to the appropriate handler. Once the stop document is received, the writer() method is called to write the output file.

Examples#

Write SPEC file automatically from data acquisition:

specwriter = apstools.callbacks.SpecWriterCallback()
RE.subscribe(specwriter.receiver)

Write NeXus file automatically from data acquisition:

nxwriter = apstools.callbacks.NXWriter()
RE.subscribe(nxwriter.receiver)

Write APS-specific NeXus file automatically from data acquisition:

nxwriteraps = apstools.callbacks.NXWriterAPS()
RE.subscribe(nxwriteraps.receiver)

Programmer’s Note#

Subclassing from object (or no superclass) avoids the need to import bluesky.callbacks.core.CallbackBase. One less import when only accessing the Databroker. The only advantage to subclassing from CallbackBase seems to be a simpler setup call to RE.subscribe().

superclass

subscription code

object

RE.subscribe(specwriter.receiver)

CallbackBase

RE.subscribe(specwriter)

HDF5/NeXus File Writers#

FileWriterCallbackBase#

Base class for filewriter callbacks.

Applications should subclass and rewrite the writer() method.

The local buffers are cleared when a start document is received. Content is collected here from each document until the stop document. The content is written once the stop document is received.

Output File Name and Path#

The output file will be written to the file named in self.file_name. (Include the output directory if different from the current working directory.)

When self.file_name is None (the default), the make_file_name() method will construct the file name using a combination of the date and time (from the start document), the start document uid, and the scan_id. The default file extension (given in NEXUS_FILE_EXTENSION) is used in make_file_name(). The directory will be self.file_path (or the current working directory if self.file_path is None which is the default).

Either specify self.file_name or override make_file_name() in a subclass to change the procedure for default output file names.

Metadata#

Almost all metadata keys (additional attributes added to the run’s start document) are completely optional. Certain keys are specified by the RunEngine, some keys are specified by the plan (or plan support methods), and other keys are supplied by the user or the instrument team.

These are the keys used by this callback to help guide how information is stored in a NeXus HDF5 file structure.

key

creator

how is it used

detectors

inconsistent

name(s) of the signals used as detectors

motors

inconsistent

synonym for positioners

plan_args

inconsistent

parameters (arguments) given

plan_name

inconsistent

name of the plan used to collect data

positioners

inconsistent

name(s) of the signals used as positioners

scan_id

RunEngine

incrementing number of the run, user can reset

uid

RunEngine

unique identifier of the run

versions

instrument

documents the software versions used to collect data

For more information about Bluesky events and document types, see https://blueskyproject.io/event-model/data-model.html.

NXWriter#

General class for writing HDF5/NeXus file (using only NeXus base classes).

One scan is written to one HDF5/NeXus file.

Output File Name and Path#

The output file will be written to the file named in self.file_name. (Include the output directory if different from the current working directory.)

When self.file_name is None (the default), the make_file_name() method will construct the file name using a combination of the date and time (from the start document), the start document uid, and the scan_id. The default file extension (given in NEXUS_FILE_EXTENSION) is used in make_file_name(). The directory will be self.file_path (or the current working directory if self.file_path is None which is the default).

Either specify self.file_name of override make_file_name() in a subclass to change the procedure for default output file names.

Metadata#

Almost all metadata keys (additional attributes added to the run’s start document) are completely optional. Certain keys are specified by the RunEngine, some keys are specified by the plan (or plan support methods), and other keys are supplied by the user or the instrument team.

These are the keys used by this callback to help guide how information is stored in a NeXus HDF5 file structure.

key

creator

how is it used

detectors

inconsistent

name(s) of the signals used as plottable values

motors

inconsistent

synonym for positioners

plan_args

inconsistent

parameters (arguments) given

plan_name

inconsistent

name of the plan used to collect data

positioners

inconsistent

name(s) of the positioners used for plotting

scan_id

RunEngine

incrementing number of the run, user can reset

subtitle

user

-tba-

title

user

/entry/title

uid

RunEngine

unique identifier of the run

versions

instrument

documents the software versions used to collect data

Notes:

  1. detectors[0] will be used as the /entry/data@signal attribute

  2. the complete list in positioners will be used as the /entry/data@axes attribute

Templates#

Use templates to build NeXus base classes or application definitions with data already collected by the nxwriter and saved in other places in the NeXus/HDF5 file (such as /entry/instrument/bluesky). Templates are used to:

  • make links from existing fields or groups to new locations

  • create new groups as directed

  • create constants for attributes or fields

A template is a [source, target] list, where source is a string and target varies depending on the type of template. A list of templates is stored as a JSON string in the run’s metadata under a pre-arranged key.

For reasons of expediency, always use absolute HDF5 addresses. Relative addresses are not supported now.

A link template makes a pointer from an existing field or group to another location. In a link template, the source is an existing HDF5 address. The target is a NeXus class path. If the path contains a group which is not yet defined, the addtional component names the NeXus class to be used. See the examples below.

A constant template defines a new NeXus field. The source string, which can be a class path (see above), ends with an =. The target can be a text, a number, or an array. Anything that can be converted into a JSON document and then written to an HDF5 dataset.

An attribute template adds a new attribute to a field or group. Use the @ symbol in the source string as shown in the examples below. The target is the value of the attribute.

EXAMPLE:

 1template_examples = [
 2    # *constant* template
 3    # define a constant array in a new NXdata group
 4    ["/entry/example:NXdata/array=", [1, 2, 3]],
 5
 6    # *attribute* template
 7    # set the example group "signal" attribute
 8    #   (new array is the default plottable data)
 9    ["/entry/example/@signal", "array"],
10
11    # *link* template
12    # link the new array into a new NXnote group as field "x"
13    ["/entry/example/array", "/entry/example/note:NXnote/x"],
14]
15
16md = {
17    "title": "NeXus/HDF5 template support example",
18
19    # encode the Python dictionary as a JSON string
20    nxwriter.template_key: json.dumps(template_examples),
21}

The templates in this example add this structure to the /entry group in the HDF5 file:

 1/entry:NXentry
 2    @NX_class = "NXentry"
 3    ...
 4    example:NXdata
 5        @NX_class = "NXdata"
 6        @signal = "array"
 7        @target = "/entry/example"
 8        array:NX_INT64[3] = [1, 2, 3]
 9            @target = "/entry/example/array"
10        note:NXnote
11            @NX_class = "NXnote"
12            @target = "/entry/example/note"
13            x --> /entry/example/array

NXWriterAPS#

Customize NXWriter with APS-specific content.

  • Adds /entry/instrument/undulator group if metadata exists.

  • Adds APS information to /entry/instrument/source group.

APS instruments should subclass NXWriterAPS to make customizations for specific plans or other considerations.

HDF5/NeXus File Structures#

Bluesky stores a wealth of information about a measurement in a run. Raw data from the Bluesky run is stored in the HDF5/NeXus structure under the /entry/instrument/bluesky group as shown in this example:

1  bluesky:NXnote
2    @NX_class = NXnote
3    @target = /entry/instrument/bluesky
4    plan_name --> /entry/instrument/bluesky/metadata/plan_name
5    uid --> /entry/instrument/bluesky/metadata/run_start_uid

The NeXus structure is built using links to the raw data in /entry/instrument/bluesky.

Metadata#

Metadata from the start document is stored in the metadata subgroup as shown in this example:

 1  metadata:NXnote
 2    @NX_class = NXnote
 3    @target = /entry/instrument/bluesky/metadata
 4    beamline_id:NX_CHAR = b'APS USAXS 9-ID-C'
 5      @target = /entry/instrument/bluesky/metadata/beamline_id
 6    datetime:NX_CHAR = b'2019-05-02 17:45:33.904824'
 7      @target = /entry/instrument/bluesky/metadata/datetime
 8    detectors:NX_CHAR = b'- I0_USAXS\n'
 9      @target = /entry/instrument/bluesky/metadata/detectors
10      @text_format = yaml
11    hints:NX_CHAR = b'dimensions:\n- - - m_stage_r\n  - primary\n'
12      @target = /entry/instrument/bluesky/metadata/hints
13      @text_format = yaml
14    login_id:NX_CHAR = b'usaxs@usaxscontrol.xray.aps.anl.gov'
15      @target = /entry/instrument/bluesky/metadata/login_id
16    motors:NX_CHAR = b'- m_stage_r\n'
17      @target = /entry/instrument/bluesky/metadata/motors
18      @text_format = yaml
19    pid:NX_INT64[] = 
20      @target = /entry/instrument/bluesky/metadata/pid
21    plan_name:NX_CHAR = b'tune_mr'
22      @target = /entry/instrument/bluesky/metadata/plan_name
23    plan_type:NX_CHAR = b'generator'
24      @target = /entry/instrument/bluesky/metadata/plan_type
25    proposal_id:NX_CHAR = b'testing Bluesky installation'
26      @target = /entry/instrument/bluesky/metadata/proposal_id
27    purpose:NX_CHAR = b'tuner'
28      @target = /entry/instrument/bluesky/metadata/purpose
29    run_start_uid:NX_CHAR = 2ffe4d87-9f0c-464a-9d14-213ec71afaf7
30      @long_name = bluesky run uid
31      @target = /entry/instrument/bluesky/metadata/run_start_uid
32    tune_md:NX_CHAR = b"initial_position: 8.824977\ntime_iso8601: '2019-05-02 17:45:33.923544'\nwidth: -0.004\n"
33      @target = /entry/instrument/bluesky/metadata/tune_md
34      @text_format = yaml
35    tune_parameters:NX_CHAR = b'initial_position: 8.824977\nnum: 31\npeak_choice: com\nwidth: -0.004\nx_axis: m_stage_r\ny_axis: I0_USAXS\n'
36      @target = /entry/instrument/bluesky/metadata/tune_parameters
37      @text_format = yaml
38    uid --> /entry/instrument/bluesky/run_start_uid

Note that complex structures (lists and dictionaries) are written as YAML. YAML is easily converted back into the python structure using yaml.load(yaml_text) where yaml_text is the YAML text from the HDF5 file.

Streams#

Data from each ophyd signal in a document stream is stored as datasets within a NXdata [1] subgroup of that stream. Bluesky collects value(s) and timestamp(s) which are stored in the datasets value and EPOCH, respectively. Since EPOCH is the absolute number of seconds from some time in the deep past, an additional time dataset repeats this data as time relative to the first time. (This makes it much easier for some visualization programs to display value v. time plots.) Additional information, as available (such as units), is added as NeXus attributes. The @target attribute is automatically added to simplify linking any item into the NeXus tree structure.

For example, the main data in a run is usually stored in the primary stream. Here, we show the tree structure for one signal (I0_USAXS) from the primary stream:

 1  primary:NXnote
 2    @NX_class = NXnote
 3    @target = /entry/instrument/bluesky/streams/primary
 4    @uid = 90489e9b-d66e-4753-8c4f-849e7a809aeb
 5    I0_USAXS:NXdata
 6      @NX_class = NXdata
 7      @axes = time
 8      @signal = value
 9      @signal_type = detector
10      @target = /entry/instrument/bluesky/streams/primary/I0_USAXS
11      EPOCH:NX_FLOAT64[31] = [1556837134.972796, 1556837135.589462, 1556837135.989462, '...', 1556837147.656129]
12        @long_name = epoch time (s)
13        @target = /entry/instrument/bluesky/streams/primary/I0_USAXS/EPOCH
14        @units = s
15      time:NX_FLOAT64[31] = [0.0, 0.6166660785675049, 1.0166659355163574, '...', 12.683332920074463]
16        @long_name = time since first data (s)
17        @start_time = 1556837134.972796
18        @start_time_iso = 2019-05-02T17:45:34.972796
19        @target = /entry/instrument/bluesky/streams/primary/I0_USAXS/time
20        @units = s
21      value:NX_FLOAT64[31] = [127.0, 126.0, 127.0, '...', 127.0]
22        @long_name = I0_USAXS
23        @lower_ctrl_limit = 0.0
24        @precision = 0
25        @signal_type = detector
26        @source = PV:9idcLAX:vsc:c0.S2
27        @target = /entry/instrument/bluesky/streams/primary/I0_USAXS/value
28        @units = 
29        @upper_ctrl_limit = 0.0

Baseline#

In Bluesky, baseline streams record the value (and timestamp) of a signal at the start and end of the run. Similar to the handling for streams (above), a subgroup is created for each baseline stream. The datasets include value, EPOCH, time (as above) and value_start and value_end. Here’s an example:

 1  baseline:NXnote
 2    @NX_class = NXnote
 3    @target = /entry/instrument/bluesky/streams/baseline
 4    @uid = f5fce6ac-f3fa-4c34-b11d-9e33c263cd20
 5    aps_current:NXdata
 6      @NX_class = NXdata
 7      @axes = time
 8      @signal = value
 9      @target = /entry/instrument/bluesky/streams/baseline/aps_current
10      EPOCH:NX_FLOAT64[2] = [1556837133.753769, 1556837147.753723]
11        @long_name = epoch time (s)
12        @target = /entry/instrument/bluesky/streams/baseline/aps_current/EPOCH
13        @units = s
14      time:NX_FLOAT64[2] = [0.0, 13.999953985214233]
15        @long_name = time since first data (s)
16        @start_time = 1556837133.753769
17        @start_time_iso = 2019-05-02T17:45:33.753769
18        @target = /entry/instrument/bluesky/streams/baseline/aps_current/time
19        @units = s
20      value:NX_FLOAT64[2] = [0.004512578244000032, 0.003944485484000011]
21        @long_name = aps_current
22        @lower_ctrl_limit = 0.0
23        @precision = 1
24        @source = PV:S:SRcurrentAI
25        @target = /entry/instrument/bluesky/streams/baseline/aps_current/value
26        @units = mA
27        @upper_ctrl_limit = 310.0
28      value_end:NX_FLOAT64[] = 
29        @long_name = aps_current
30        @lower_ctrl_limit = 0.0
31        @precision = 1
32        @source = PV:S:SRcurrentAI
33        @target = /entry/instrument/bluesky/streams/baseline/aps_current/value_end
34        @units = mA
35        @upper_ctrl_limit = 310.0
36      value_start:NX_FLOAT64[] = 
37        @long_name = aps_current
38        @lower_ctrl_limit = 0.0
39        @precision = 1
40        @source = PV:S:SRcurrentAI
41        @target = /entry/instrument/bluesky/streams/baseline/aps_current/value_start
42        @units = mA
43        @upper_ctrl_limit = 310.0

Full Structure#

The full structure of the example HDF5/NeXus file (omitting the attributes and array data for brevity) is shown next. You can see that most of the NeXus structure is completed by making links to data from either /entry/instrument/bluesky/metadata or /entry/instrument/bluesky/streams:

  120190502-174533-S00108-2ffe4d8.hdf : NeXus data file
  2  entry:NXentry
  3    duration:NX_FLOAT64[] = [ ... ]
  4    end_time:NX_CHAR = 2019-05-02T17:45:48.078618
  5    entry_identifier --> /entry/instrument/bluesky/uid
  6    plan_name --> /entry/instrument/bluesky/metadata/plan_name
  7    program_name:NX_CHAR = bluesky
  8    start_time:NX_CHAR = 2019-05-02T17:45:33.937294
  9    title:NX_CHAR = tune_mr-S0108-2ffe4d8
 10    contact:NXuser
 11      affiliation --> /entry/instrument/bluesky/streams/baseline/bss_user_info_institution/value_start
 12      email --> /entry/instrument/bluesky/streams/baseline/bss_user_info_email/value_start
 13      facility_user_id --> /entry/instrument/bluesky/streams/baseline/bss_user_info_badge/value_start
 14      name --> /entry/instrument/bluesky/streams/baseline/bss_user_info_contact/value_start
 15      role:NX_CHAR = contact
 16    data:NXdata
 17      EPOCH --> /entry/instrument/bluesky/streams/primary/scaler0_time/time
 18      I0_USAXS --> /entry/instrument/bluesky/streams/primary/I0_USAXS/value
 19      m_stage_r --> /entry/instrument/bluesky/streams/primary/m_stage_r/value
 20      m_stage_r_soft_limit_hi --> /entry/instrument/bluesky/streams/primary/m_stage_r_soft_limit_hi/value
 21      m_stage_r_soft_limit_lo --> /entry/instrument/bluesky/streams/primary/m_stage_r_soft_limit_lo/value
 22      m_stage_r_user_setpoint --> /entry/instrument/bluesky/streams/primary/m_stage_r_user_setpoint/value
 23      scaler0_display_rate --> /entry/instrument/bluesky/streams/primary/scaler0_display_rate/value
 24      scaler0_time --> /entry/instrument/bluesky/streams/primary/scaler0_time/value
 25    instrument:NXinstrument
 26      bluesky:NXnote
 27        plan_name --> /entry/instrument/bluesky/metadata/plan_name
 28        uid --> /entry/instrument/bluesky/metadata/run_start_uid
 29        metadata:NXnote
 30          APSTOOLS_VERSION:NX_CHAR = b'1.1.0'
 31          BLUESKY_VERSION:NX_CHAR = b'1.5.2'
 32          EPICS_CA_MAX_ARRAY_BYTES:NX_CHAR = b'1280000'
 33          EPICS_HOST_ARCH:NX_CHAR = b'linux-x86_64'
 34          OPHYD_VERSION:NX_CHAR = b'1.3.3'
 35          beamline_id:NX_CHAR = b'APS USAXS 9-ID-C'
 36          datetime:NX_CHAR = b'2019-05-02 17:45:33.904824'
 37          detectors:NX_CHAR = b'- I0_USAXS\n'
 38          hints:NX_CHAR = b'dimensions:\n- - - m_stage_r\n  - primary\n'
 39          login_id:NX_CHAR = b'usaxs@usaxscontrol.xray.aps.anl.gov'
 40          motors:NX_CHAR = b'- m_stage_r\n'
 41          pid:NX_INT64[] = [ ... ]
 42          plan_name:NX_CHAR = b'tune_mr'
 43          plan_type:NX_CHAR = b'generator'
 44          proposal_id:NX_CHAR = b'testing Bluesky installation'
 45          purpose:NX_CHAR = b'tuner'
 46          run_start_uid:NX_CHAR = 2ffe4d87-9f0c-464a-9d14-213ec71afaf7
 47          tune_md:NX_CHAR = b"initial_position: 8.824977\ntime_iso8601: '2019-05-02 17:45:33.923544'\nwidth: -0.004\n"
 48          tune_parameters:NX_CHAR = b'initial_position: 8.824977\nnum: 31\npeak_choice: com\nwidth: -0.004\nx_axis: m_stage_r\ny_axis: I0_USAXS\n'
 49          uid --> /entry/instrument/bluesky/run_start_uid
 50        streams:NXnote
 51          baseline:NXnote
 52            aps_current:NXdata
 53              EPOCH:NX_FLOAT64[2] = [ ... ]
 54              time:NX_FLOAT64[2] = [ ... ]
 55              value:NX_FLOAT64[2] = [ ... ]
 56              value_end:NX_FLOAT64[] = [ ... ]
 57              value_start:NX_FLOAT64[] = [ ... ]
 58            aps_fill_number:NXdata
 59              EPOCH:NX_FLOAT64[2] = [ ... ]
 60              time:NX_FLOAT64[2] = [ ... ]
 61              value:NX_FLOAT64[2] = [ ... ]
 62              value_end:NX_FLOAT64[] = [ ... ]
 63              value_start:NX_FLOAT64[] = [ ... ]
 64            aps_global_feedback:NXdata
 65              EPOCH:NX_FLOAT64[2] = [ ... ]
 66              time:NX_FLOAT64[2] = [ ... ]
 67              value:NX_CHAR[3,3] = ["Off", "Off"]
 68              value_end:NX_CHAR = b'Off'
 69              value_start:NX_CHAR = b'Off'
 70            # many baseline groups omitted for brevity
 71          primary:NXnote
 72            I0_USAXS:NXdata
 73              EPOCH:NX_FLOAT64[31] = [ ... ]
 74              time:NX_FLOAT64[31] = [ ... ]
 75              value:NX_FLOAT64[31] = [ ... ]
 76            m_stage_r:NXdata
 77              EPOCH:NX_FLOAT64[31] = [ ... ]
 78              time:NX_FLOAT64[31] = [ ... ]
 79              value:NX_FLOAT64[31] = [ ... ]
 80            m_stage_r_soft_limit_hi:NXdata
 81              EPOCH:NX_FLOAT64[31] = [ ... ]
 82              time:NX_FLOAT64[31] = [ ... ]
 83              value:NX_FLOAT64[31] = [ ... ]
 84            m_stage_r_soft_limit_lo:NXdata
 85              EPOCH:NX_FLOAT64[31] = [ ... ]
 86              time:NX_FLOAT64[31] = [ ... ]
 87              value:NX_FLOAT64[31] = [ ... ]
 88            m_stage_r_user_setpoint:NXdata
 89              EPOCH:NX_FLOAT64[31] = [ ... ]
 90              time:NX_FLOAT64[31] = [ ... ]
 91              value:NX_FLOAT64[31] = [ ... ]
 92            scaler0_display_rate:NXdata
 93              EPOCH:NX_FLOAT64[31] = [ ... ]
 94              time:NX_FLOAT64[31] = [ ... ]
 95              value:NX_FLOAT64[31] = [ ... ]
 96            scaler0_time:NXdata
 97              EPOCH:NX_FLOAT64[31] = [ ... ]
 98              time:NX_FLOAT64[31] = [ ... ]
 99              value:NX_FLOAT64[31] = [ ... ]
100      detectors:NXnote
101        I0_USAXS:NXdetector
102          data --> /entry/instrument/bluesky/streams/primary/I0_USAXS
103      monochromator:NXmonochromator
104        energy --> /entry/instrument/bluesky/streams/baseline/monochromator_dcm_energy/value_start
105        feedback_on --> /entry/instrument/bluesky/streams/baseline/monochromator_feedback_on/value_start
106        mode --> /entry/instrument/bluesky/streams/baseline/monochromator_dcm_mode/value_start
107        theta --> /entry/instrument/bluesky/streams/baseline/monochromator_dcm_theta/value_start
108        wavelength --> /entry/instrument/bluesky/streams/baseline/monochromator_dcm_wavelength/value_start
109        y_offset --> /entry/instrument/bluesky/streams/baseline/monochromator_dcm_y_offset/value_start
110      positioners:NXnote
111        m_stage_r:NXpositioner
112          value --> /entry/instrument/bluesky/streams/primary/m_stage_r
113      source:NXsource
114        name:NX_CHAR = Bluesky framework
115        probe:NX_CHAR = x-ray
116        type:NX_CHAR = Synchrotron X-ray Source

APS-specific HDF5/NeXus File Structures#

Examples of additional structure in NeXus file added by NXWriterAPS():

 1  source:NXsource
 2    @NX_class = NXsource
 3    @target = /entry/instrument/source
 4    current --> /entry/instrument/bluesky/streams/baseline/aps_current/value_start
 5    energy:NX_INT64[] = 
 6      @units = GeV
 7    fill_number --> /entry/instrument/bluesky/streams/baseline/aps_fill_number/value_start
 8    name:NX_CHAR = Advanced Photon Source
 9      @short_name = APS
10    probe:NX_CHAR = x-ray
11    type:NX_CHAR = Synchrotron X-ray Source
12  undulator:NXinsertion_device
13    @NX_class = NXinsertion_device
14    @target = /entry/instrument/undulator
15    device --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_device/value_start
16    energy --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_energy/value_start
17    energy_taper --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_energy_taper/value_start
18    gap --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_gap/value_start
19    gap_taper --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_gap_taper/value_start
20    harmonic_value --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_harmonic_value/value_start
21    location --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_location/value_start
22    total_power --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_total_power/value_start
23    type:NX_CHAR = undulator
24    version --> /entry/instrument/bluesky/streams/baseline/undulator_downstream_version/value_start

SPEC File Structure#

EXAMPLE : use as Bluesky callback:

from apstools import SpecWriterCallback
specwriter = SpecWriterCallback()
RE.subscribe(specwriter.receiver)

EXAMPLE : use as writer from Databroker:

from apstools import SpecWriterCallback
specwriter = SpecWriterCallback()
for key, doc in db.get_documents(db["a123456"]):
    specwriter.receiver(key, doc)
print("Look at SPEC data file: "+specwriter.spec_filename)

EXAMPLE : use as writer from Databroker with customizations:

from apstools import SpecWriterCallback

# write into file: /tmp/cerium.spec
specwriter = SpecWriterCallback(filename="/tmp/cerium.spec")
for key, doc in db.get_documents(db["abcd123"]):
    specwriter.receiver(key, doc)

# write into file: /tmp/barium.dat
specwriter.newfile("/tmp/barium.dat")
for key, doc in db.get_documents(db["b46b63d4"]):
    specwriter.receiver(key, doc)

Example output from SpecWriterCallback():

 1#F test_specdata.txt
 2#E 1510948301
 3#D Fri Nov 17 13:51:41 2017
 4#C BlueSky  user = mintadmin  host = mint-vm
 5
 6#S 233  scan(detectors=['synthetic_pseudovoigt'], num=20, motor=['m1'], start=-1.65, stop=-1.25, per_step=None)
 7#D Fri Nov 17 11:58:56 2017
 8#C Fri Nov 17 11:58:56 2017.  plan_type = generator
 9#C Fri Nov 17 11:58:56 2017.  uid = ddb81ac5-f3ee-4219-b047-c1196d08a5c1
10#MD beamline_id = developer__YOUR_BEAMLINE_HERE
11#MD login_id = mintadmin@mint-vm
12#MD motors = ['m1']
13#MD num_intervals = 19
14#MD num_points = 20
15#MD pid = 7133
16#MD plan_pattern = linspace
17#MD plan_pattern_args = {'start': -1.65, 'stop': -1.25, 'num': 20}
18#MD plan_pattern_module = numpy
19#MD proposal_id = None
20#N 20
21#L m1  m1_user_setpoint  Epoch_float  Epoch  synthetic_pseudovoigt
22-1.6500000000000001 -1.65 8.27465009689331 8 2155.6249784809206
23-1.6288 -1.6289473684210525 8.46523666381836 8 2629.5229081466964
24-1.608 -1.6078947368421053 8.665581226348877 9 3277.4074328018964
25-1.5868 -1.5868421052631578 8.865738153457642 9 4246.145049452576
26-1.5656 -1.5657894736842104 9.066259145736694 9 5825.186516381953
27-1.5448000000000002 -1.5447368421052632 9.266754627227783 9 8803.414029867528
28-1.5236 -1.5236842105263158 9.467074871063232 9 15501.419687691103
29-1.5028000000000001 -1.5026315789473683 9.667330741882324 10 29570.38936784884
30-1.4816 -1.4815789473684209 9.867793798446655 10 55562.3437459487
31-1.4604000000000001 -1.4605263157894737 10.067811012268066 10 89519.64275090238
32-1.4396 -1.4394736842105262 10.268356084823608 10 97008.97190269837
33-1.4184 -1.418421052631579 10.470621824264526 10 65917.29757650592
34-1.3972 -1.3973684210526316 10.669955730438232 11 36203.46726798266
35-1.3764 -1.3763157894736842 10.870310306549072 11 18897.64061096024
36-1.3552 -1.3552631578947367 11.070487976074219 11 10316.223844200193
37-1.3344 -1.3342105263157895 11.271018743515015 11 6540.179615556269
38-1.3132000000000001 -1.313157894736842 11.4724280834198 11 4643.555421314616
39-1.292 -1.2921052631578946 11.673305034637451 12 3533.8582404216445
40-1.2712 -1.2710526315789474 11.874176025390625 12 2809.1872596809008
41-1.25 -1.25 12.074703216552734 12 2285.9226305883626
42#C Fri Nov 17 11:59:08 2017.  num_events_primary = 20
43#C Fri Nov 17 11:59:08 2017.  time = 2017-11-17 11:59:08.324011
44#C Fri Nov 17 11:59:08 2017.  exit_status = success

Source Code#

Base Class for File Writer Callbacks#

FileWriterCallbackBase(*args, **kwargs)

Base class for filewriter callbacks.

class apstools.callbacks.callback_base.FileWriterCallbackBase(*args, **kwargs)[source]#

Base class for filewriter callbacks.

New with apstools release 1.3.0.

Applications should subclass and rewrite the writer() method.

The local buffers are cleared when a start document is received. Content is collected here from each document until the stop document. The content is written once the stop document is received.

User Interface methods

receiver(key, doc)

bluesky callback (handles a stream of documents)

Internal methods

clear()

delete any saved data from the cache and reinitialize

make_file_name()

generate a file name to be used as default

writer()

print summary of run as diagnostic

Document Handler methods

bulk_events(doc)

Deprecated.

datum(doc)

Like an event, but for data recorded outside of bluesky.

descriptor(doc)

description of the data stream to be acquired

event(doc)

a single "row" of data

resource(doc)

like a descriptor, but for data recorded outside of bluesky

start(doc)

beginning of a run, clear cache and collect metadata

stop(doc)

end of the run, end collection and initiate the writer() method

bulk_events(doc)[source]#

Deprecated. Use EventPage instead.

clear()[source]#

delete any saved data from the cache and reinitialize

datum(doc)[source]#

Like an event, but for data recorded outside of bluesky.

Example:

Datum
=====
datum_id        : 621caa0f-70f1-4e3d-8718-b5123d434502/0
datum_kwargs    :
HDF5_file_name  : /mnt/usaxscontrol/USAXS_data/2020-06/06_10_Minjee_waxs/AGIX3N1_0699.hdf
point_number    : 0
resource        : 621caa0f-70f1-4e3d-8718-b5123d434502
descriptor(doc)[source]#

description of the data stream to be acquired

event(doc)[source]#

a single “row” of data

make_file_name()[source]#

generate a file name to be used as default

default format: {ymd}-{hms}-S{scan_id}-{short_uid}.{ext} where the time (the run start time):

  • ymd = {year:4d}{month:02d}{day:02d}

  • hms = {hour:02d}{minute:02d}{second:02d}

override in subclass to change

receiver(key, doc)[source]#

bluesky callback (handles a stream of documents)

resource(doc)[source]#

like a descriptor, but for data recorded outside of bluesky

start(doc)[source]#

beginning of a run, clear cache and collect metadata

stop(doc)[source]#

end of the run, end collection and initiate the writer() method

writer()[source]#

print summary of run as diagnostic

override this method in subclass to write a file

NeXus File Writer Callbacks#

NXWriter(*args, **kwargs)

General class for writing HDF5/NeXus file (using only NeXus base classes).

NXWriterAPS(*args, **kwargs)

Customize NXWriter with APS-specific content.

class apstools.callbacks.nexus_writer.NXWriter(*args, **kwargs)[source]#

General class for writing HDF5/NeXus file (using only NeXus base classes).

One scan is written to one HDF5/NeXus file.

Note

If you use NXWriter, you must wait for the writer() method to finish before proceeding with the next acquisition or processing. (The writer() method is launched in a background thread to complete once all readable assets are available, potentially even after the run ends.)

See the table below for which wait method to call.

EXAMPLES:

Interactive use (outside of a plan):

# ...
nxwriter = NXWriter()  # create the callback instance
RE.subscribe(nxwriter.receiver)  # subscribe to the RunEngine
# ...
RE(bp.count([camera]))
nxwriter.wait_writer()
# ...
run = cat.v2[-1]  # for additional processing

Note

There are two methods to wait for the callback. One is for interactive use when not using the RunEngine. The other is for use in a plan that is executed by the bluesky RunEngine.

When

Method

Which uses

interactive

wait_writer()

time.sleep()

in a plan

wait_writer_plan_stub()

yield from bps.sleep()

The wait_writer() method calls time.sleep() and this would block the RunEngine from its routine processing of other background tasks. The wait_writer_plan_stub() method replaces that call with yield from bps.sleep() which does not block the RunEngine from processing other background tasks.

In a custom plan, use the wait_writer_plan_stub() method instead:

# ...
nxwriter = NXWriter()  # create the callback instance
RE.subscribe(nxwriter.receiver)  # subscribe to the RunEngine
# ...

def my_plan(dets, n=5):
    for i in range(n):
        yield from bp.count(dets)
        yield from nxwriter.wait_writer_plan_stub()

METHODS

writer()

Write collected data to HDF5/NeXus data file.

h5string(text)

Format string for h5py interface.

add_dataset_attributes(ds, v[, long_name])

add attributes from v dictionary to dataset ds

assign_signal_type()

decide if a signal in the primary stream is a detector or a positioner

create_NX_group(parent, specification)

create an h5 group with named NeXus class (specification)

get_sample_title()

return the title for this sample

get_stream_link(signal[, stream, ref])

return the h5 object for signal

resolve_class_path(class_path)

Parse the class path, make any groups, return the HDF5 address.

wait_writer()

Wait for the writer to finish.

wait_writer_plan_stub()

Wait for the writer to finish.

write_data(parent)

group: /entry/data:NXdata

write_detector(parent)

group: /entry/instrument/detectors:NXnote/DETECTOR:NXdetector

write_entry()

group: /entry/data:NXentry

write_instrument(parent)

group: /entry/instrument:NXinstrument

write_metadata(parent)

group: /entry/instrument/bluesky/metadata:NXnote

write_monochromator(parent)

group: /entry/instrument/monochromator:NXmonochromator

write_positioner(parent)

group: /entry/instrument/positioners:NXnote/POSITIONER:NXpositioner

write_root(filename)

root of the HDF5 file

write_sample(parent)

group: /entry/sample:NXsample

write_slits(parent)

group: /entry/instrument/slits:NXnote/SLIT:NXslit

write_source(parent)

group: /entry/instrument/source:NXsource

write_streams(parent)

group: /entry/instrument/bluesky/streams:NXnote

write_templates()

Process any link templates provided as run metadata.

write_user(parent)

group: /entry/contact:NXuser

New with apstools release 1.3.0. Update in release 1.6.11 to wait for area detector HDF5 files.

add_dataset_attributes(ds, v, long_name=None)[source]#

add attributes from v dictionary to dataset ds

assign_signal_type()[source]#

decide if a signal in the primary stream is a detector or a positioner

create_NX_group(parent, specification)[source]#

create an h5 group with named NeXus class (specification)

getResourceFile(resource_id)[source]#

full path to the resource file specified by uid resource_id

override in subclass as needed

get_sample_title()[source]#

return the title for this sample

default title: S{scan_id}-{plan_name}-{short_uid}

return the h5 object for signal

DEFAULTS

stream : baseline key : value_start

h5string(text)[source]#

Format string for h5py interface.

resolve_class_path(class_path)[source]#

Parse the class path, make any groups, return the HDF5 address.

New with apstools release 1.6.18.

template_key = 'nxwriter_template'#

The template (dict) is written as a JSON string to this metadata key.

wait_writer()[source]#

Wait for the writer to finish. For interactive use (Not in a plan).

If you use NXWriter interactively, you must call NXWriter.wait_writer() which waits for all data processing to finish before proceeding with the next acquisition or processing.

wait_writer_plan_stub()[source]#

Wait for the writer to finish. Use in a plan (with RunEngine).

If you use NXWriter in a plan, you must call NXWriter.wait_writer() which waits for all data processing to finish before proceeding with the next acquisition or processing.

write_data(parent)[source]#

group: /entry/data:NXdata

write_detector(parent)[source]#

group: /entry/instrument/detectors:NXnote/DETECTOR:NXdetector

write_entry()[source]#

group: /entry/data:NXentry

write_instrument(parent)[source]#

group: /entry/instrument:NXinstrument

write_metadata(parent)[source]#

group: /entry/instrument/bluesky/metadata:NXnote

metadata from the bluesky start document

write_monochromator(parent)[source]#

group: /entry/instrument/monochromator:NXmonochromator

write_positioner(parent)[source]#

group: /entry/instrument/positioners:NXnote/POSITIONER:NXpositioner

write_root(filename)[source]#

root of the HDF5 file

write_sample(parent)[source]#

group: /entry/sample:NXsample

write_slits(parent)[source]#

group: /entry/instrument/slits:NXnote/SLIT:NXslit

override in subclass to store content, name patterns vary with each instrument

write_source(parent)[source]#

group: /entry/instrument/source:NXsource

Note: this is (somewhat) generic, override for a different source

write_streams(parent)[source]#

group: /entry/instrument/bluesky/streams:NXnote

data from all the bluesky streams

write_templates()[source]#

Process any link templates provided as run metadata.

New in v1.6.18.

write_user(parent)[source]#

group: /entry/contact:NXuser

writer()[source]#

Write collected data to HDF5/NeXus data file.

The callback launches _threaded_writer() and returns. In the thread, write_root() (or methods within) can wait on certain items (such as an external HDF5 file written by an area detector IOC) to become readable or a timeout period has expired.

class apstools.callbacks.nexus_writer.NXWriterAPS(*args, **kwargs)[source]#

Customize NXWriter with APS-specific content.

New with apstools release 1.3.0.

  • Adds /entry/instrument/undulator group if metadata exists.

  • Adds APS information to /entry/instrument/source group.

write_instrument(parent)

group: /entry/instrument:NXinstrument

write_source(parent)

group: /entry/instrument/source:NXsource

write_undulator(parent)

group: /entry/instrument/undulator:NXinsertion_device

write_instrument(parent)[source]#

group: /entry/instrument:NXinstrument

write_source(parent)[source]#

group: /entry/instrument/source:NXsource

Note: this is specific to the APS, override for a different source

write_undulator(parent)[source]#

group: /entry/instrument/undulator:NXinsertion_device

SPEC Data File Writer Callback#

EXAMPLE:

Execution of this plan (with RE(myPlan())):

def myPlan():
    yield from bps.open_run()
    spec_comment("this is a start document comment", "start")
    spec_comment("this is a descriptor document comment", "descriptor")
    yield bps.Msg('checkpoint')
    yield from bps.trigger_and_read([scaler])
    spec_comment("this is an event document comment after the first read")
    yield from bps.sleep(2)
    yield bps.Msg('checkpoint')
    yield from bps.trigger_and_read([scaler])
    spec_comment("this is an event document comment after the second read")
    spec_comment("this is a stop document comment", "stop")
    yield from bps.close_run()

results in this SPEC file output:

#S 1145  myPlan()
#D Mon Jan 28 12:48:09 2019
#C Mon Jan 28 12:48:09 2019.  plan_type = generator
#C Mon Jan 28 12:48:09 2019.  uid = ef98648a-8e3a-4e7e-ac99-3290c9b5fca7
#C Mon Jan 28 12:48:09 2019.  this is a start document comment
#C Mon Jan 28 12:48:09 2019.  this is a descriptor document comment
#MD APSTOOLS_VERSION = 2019.0103.0+5.g0f4e8b2
#MD BLUESKY_VERSION = 1.4.1
#MD OPHYD_VERSION = 1.3.0
#MD SESSION_START = 2019-01-28 12:19:25.446836
#MD beamline_id = developer
#MD ipython_session_start = 2018-02-14 12:54:06.447450
#MD login_id = mintadmin@mint-vm
#MD pid = 21784
#MD proposal_id = None
#N 2
#L Epoch_float  scaler_time  Epoch
1.4297869205474854 1.1 1
4.596935987472534 1.1 5
#C Mon Jan 28 12:48:11 2019.  this is an event document comment after the first read
#C Mon Jan 28 12:48:14 2019.  this is an event document comment after the second read
#C Mon Jan 28 12:48:14 2019.  this is a stop document comment
#C Mon Jan 28 12:48:14 2019.  num_events_primary = 2
#C Mon Jan 28 12:48:14 2019.  exit_status = success

SpecWriterCallback([filename, auto_write, ...])

Collect data from Bluesky RunEngine documents to write as SPEC data.

spec_comment(comment[, doc, writer])

make it easy to add spec-style comments in a custom plan

class apstools.callbacks.spec_file_writer.SpecWriterCallback(filename=None, auto_write=True, RE=None, reset_scan_id=False)[source]#

Collect data from Bluesky RunEngine documents to write as SPEC data.

This gathers data from all documents in a scan and appends scan to the file when the stop document is received. One or more scans can be written to the same file. The file format is text.

Note

SpecWriterCallback() does not inherit from FileWriterCallbackBase().

PARAMETERS

filename

string : (optional) Local, relative or absolute name of SPEC data file to be used. If filename=None, defaults to format of YYYmmdd-HHMMSS.dat derived from the current system time.

auto_write

boolean : (optional) If True (default), write_scan() is called when stop document is received. If False, the caller is responsible for calling write_scan() before the next start document is received.

RE

object : Instance of bluesky.RunEngine or None.

reset_scan_id

boolean : (optional) If True, and filename exists, then sets RE.md.scan_id to highest scan number in existing SPEC data file. default: False

User Interface methods

receiver(key, document)

Bluesky callback: receive all documents for handling

newfile([filename, scan_id, RE])

prepare to use a new SPEC data file

usefile(filename)

read from existing SPEC data file

make_default_filename()

generate a file name to be used as default

clear()

reset all scan data defaults

prepare_scan_contents()

format the scan for a SPEC data file

write_scan()

write the most recent (completed) scan to the file

Internal methods

write_header()

Write the (initial) header section of a SPEC data file.

start(doc)

handle start documents

descriptor(doc)

handle descriptor documents

event(doc)

handle event documents

bulk_events(doc)

handle bulk_events documents

datum(doc)

handle datum documents

resource(doc)

handle resource documents

stop(doc)

handle stop documents

bulk_events(doc)[source]#

handle bulk_events documents

clear()[source]#

reset all scan data defaults

datum(doc)[source]#

handle datum documents

descriptor(doc)[source]#

handle descriptor documents

prepare for primary scan data, ignore any other data stream

event(doc)[source]#

handle event documents

make_default_filename()[source]#

generate a file name to be used as default

newfile(filename=None, scan_id=None, RE=None)[source]#

prepare to use a new SPEC data file

but don’t create it until we have data

prepare_scan_contents()[source]#

format the scan for a SPEC data file

Returns:

[str] a list of lines to append to the data file

receiver(key, document)[source]#

Bluesky callback: receive all documents for handling

resource(doc)[source]#

handle resource documents

start(doc)[source]#

handle start documents

stop(doc)[source]#

handle stop documents

usefile(filename)[source]#

read from existing SPEC data file

write_header()[source]#

Write the (initial) header section of a SPEC data file.

write_scan()[source]#

write the most recent (completed) scan to the file

  • creates file if not existing

  • writes header if needed

  • appends scan data

note: does nothing if there are no lines to be written

apstools.callbacks.spec_file_writer.spec_comment(comment, doc=None, writer=None)[source]#

make it easy to add spec-style comments in a custom plan

These comments only go into the SPEC data file.

PARAMETERS

comment string :

(optional) Comment text to be written. SPEC expects it to be only one line!

doc string :

(optional) Bluesky RunEngine document type. One of: start descriptor event resource datum stop (default: event)

writer obj :

(optional) Instance of SpecWriterCallback(), typically: specwriter = SpecWriterCallback()