Sample Logs

The following information will be useful to you if you want to write an Algorithm that manipulates sample log data.

Sample logs

When you load a nexus file using Mantid, a set of sample logs relevant to the experimental run are loaded into the workspace’s run object, which can be accessed through:

#include "MantidAPI/Run.h"

// Declare property with default settings
// IndexType::WorkspaceIndex is default
MatrixWorkspace_sptr workspace;
....
Run &run = workspace->mutableRun()
// get all log data
auto logData = run.getLogData()
// access individual log
auto protonCharge = run.getLogData("proton_charge")

Log data can be added and removed from the run object using the addLogData and removeLogData methods.

Multiperiod workspace sample logs

When you load multiperiod data using LoadISISNexus v2 or LoadEventNexus v1 a set of periods logs, describing the period information for each workspace are created using the ISISRunLogs class. This class adds three logs, which are summarised as:

period n

This log contains times and boolean flags describing when a period started, but not necessarily when data collection for that period occurred. For example, a typical sample log entry for this log would read:

Time	Boolean flag value	Description
2020-Oct-20 12:19:40	0
2020-Oct-20 12:19:41	1	nth period started
2020-Oct-20 12:19:48	0	nth period ended

running

This log contains a series of times and boolean describing when a data collection was occurring, for example:

Time	Boolean flag value	Description
2020-Oct-20 12:19:40	0
2020-Oct-20 12:19:41	1	Data collection started
2020-Oct-20 12:19:46	0	Data collection ended

periods

This log contains times and period numbers describing when each period started, e.g.

Time	Period number	Description
2020-Oct-20 12:19:40	1	Period 1 started
2020-Oct-20 12:19:49	2	Period 2 started
2020-Oct-20 12:19:56	3	Period 3 started

These three logs are constructed from the ICP_EVENT which tells you when various “events” of significance to the software happen, such as beginning or end of a run or a change of period number.

Filtering

For multiperiod workspaces, time-series data, such as Theta will be filtered by combining the period n and running logs, to create a filter describing when data collection for each period occurred. For example, a filter for the nth period would be described as follows, where we note that each period is described entirely be two boolean values, the first describing when the data collection started for that period and the second stating when data collection ended.

Time	Boolean flag	Description
2020-Oct-20 12:19:40	0
2020-Oct-20 12:19:41	1	Data collection for period n started
2020-Oct-20 12:19:46	0	Data collection for period n finished

This filtering is performed by the static method ISISRunLogs::applyLogFiltering.

Event data

For event data, an additional period_log is present in the workspace, which is created from the framelog/period_log entry. This entry contains the value of various items on a frame by frame (pulse by pulse) basis. So it is showing you what the neutron acquisition electronics believe the period number to be at the point that bit of neutron data was recorded.

The framelog/period_log entry will therefore contain a list of m times and the corresponding period number which was active during that frame:

Time	Period number	Description
2020-Oct-20 12:19:40	1	Period 1 data
2020-Oct-20 12:19:41	1	Period 1 data
2020-Oct-20 12:19:44	1	Period 1 data
2020-Oct-20 12:19:50	2	Period 2 data
2020-Oct-20 12:19:51	2	Period 2 data
2020-Oct-20 12:19:52	3	Period 3 data
2020-Oct-20 12:20:01	3	Period 3 data
2020-Oct-20 12:20:02	3	Period 3 data

This log is therefore a combination of the period n and running logs defined above, whereby it records the times and periods during data collection. However, rather than recording a single value describing when the data collection for that period started, the framelog data contains a discrete number of time recordings corresponding to the period that each “bit” of neutron data was collected in.