the instruments_used group is still present outside the multilayer group
but currently a new instruments_used sub-group is created in the
layer-specific group
instruments used to deposit a single layer are in
/sample/multilayer/layer_N/instruments_used and there's only one value
for each category (rheed, laser, chamber)
in /instruments_used (root) for each category there's a list of every
(unique) instrument involved in the full deposition process
now keys with numeric values are sub-dictionaries with a "value" and a
"units" key - unitS not unit to comply directly with NeXus format, which
turned out to be a good idea to avoid confusion since eLabFTW uses the
word "units" for the list of accepted units and "unit" for the selected
one...
NOTE: UoM = Unit of Measurement
Target.description defaults to "" (empty str)
Substrate.thickness defaults to "" (empty str)
Substrate.thickness_unit is now hardcoded to "μm"
did you know? apparently h5py does NOT like null values
get_instruments returns a dictionary with the names of every system used
during the deposition
unfortunately, NeXus standard allows for a single value of all three
keys per every sample - not every layer
this means that every layer has its own data for laser, rheed system and
depo chamber which IDEALLY is the same for every layer, but in practice
they COULD be different and I still don't know how to deal with this
only lacks the instrument_used data and units of measurement
NOTE: units of measurement are hard to collect, but could be assumed
considering our instruments are standard
takes API key and elabid of the "entrypoint" sample as input
returns indented JSON with the reconstructed dataset!
currently lacks instruments_used data (matter of minutes)
and all the layer data (already present in Layer-class objects)
the fix changes "sample_data" - which was residual from copy-pasting the
call_entrypoint... function and therefore undefined - with
"material_data" which is very well defined in the previous line
