Discuss fmu-dataio scope - also _read_ results from storage?

[perolavsvendsen]

For discussion:
Should scope of fmu-dataio be expanded from "just" exporting data, to also reading data?

Positives:

• Alignment between data exporting and data reading, e.g. we better ensure that data exported can also be read. A risk today is that exported metadata does not make it possible to obtain uniqueness and/or does not make it easy/possible to read the data afterwards.
• Better conditions for developing end-to-end solutions for FMU results being exported.
• More involvement from other teams, e.g. Sudo. The fmu.sumo.explorer library is used for reading FMU results, but developing require a lot of knowledge about both consumption patterns and metadata. Every development in fmu-sumo creates new contracts on metadata, and the two must be perfectly aligned. This is tricky to obtain in practice. However, if these were part of the same PR's, it would perhaps be better.
• Incentive for considering data reading a bit more during development of fmu-dataio.

Negatives:

• Scope is already big
• Getting data into FMU (input data) is also a logical scope candidate for fmu-dataio

Conceptual example

# export FMU results (existing scope)
from fmu.dataio.export import InplaceVolumesRMS
myexport = InplaceVolumesRMS(...).export()

# read FMU results (expanded scope)
from fmu.dataio.read import InplaceVolumes
df = InplaceVolumes(...).to_pandas()