Sample Logs should store units, besides value

[jmborr]

TODO:

• Read log unit from the Nexus events file and save into hydraworkspace (see NexusConvertingApp.split_sample_logs() and HidraWorkspace.set_sample_log())
• saving to project file should save both value and unit. (see HidraWorkspace.save_experimental_data())
• reading from project file will read both value and unit (see HidraProjectFile.read_sample_logs())
• backwards compatibility: old hydra project files will contain only log values. Thus, when loaded each log entry should be assigned an empty string for units attribute.

Possible locations in the code base impacted by the introduction of LogEntry:

• reading to and writing from the project file
• Peak fitting UI
• Sample logs when creating the point list

Blocks #274

Notes in the current implementation of the sample logs:

SampleLogs Class

• class SampleLogs is instantiated in HidraWorkspace.__init__ as attribute _sample_logs, with an empty state
• The state of HidraWorkspace._sample_logs is initialized in HidraWorkspace._load_sample_logs(HidraProjectFile), which
  calls HidraProjectFile.read_sample_logs()

Project File

The hierarchy of a HDF5 project file:

(Subruns begin at 1, not 0. )

Log entries contain a single float for each subrun number. The attribute is empty. The idea is to extend it to contain a value and a unit.

Writing the logs in a projectfile with class HidraProjectFile:

• HidraProjectFile._init_project() creates datasets for logs and subruns, but don't fill them
• HidraProjectFile.append_experiment_log(log_name, log_value_array) add one log entry in the logs dataset
• HidraProjectFile.write_sub_runs(sub_runs) add data into the subruns dataset

Writing the logs to a project file is done from a HidraWorkspace:

• HidraWorkspace.save_experimental_data() calls HidraProjectFile.get_sample_log_values() and immediately after calls HidraProjectFile.append_experiment_log()