Add tuple as input type in parse_excel.infer_metadata()

[ajeklund]

Description:

parse_excel.infer_metadata() expects its input parameter units as a list, but we want to send a tuple from get(). Suggested solution is to add handling of tuple.

Closes #212.

Type of change:

• Bug fix.
• New feature.
• Documentation update.

Checklist for the reviewer:

This checklist should be used as a help for the reviewer.

• Is the change limited to one issue?
• Does this PR close the issue?
• Is the code easy to read and understand, including clearly named variables?
• Do all new feature have an accompanying new test?
• Has the documentation been updated as necessary?

[CasperWA]

Is there are case where units is actually a list? Or is it always (now) a tuple? If so, the list type should simply be changed to tuple without the Union.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (a7bcd3a) 88.67% compared to head (7a697c8) 88.67%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #213   +/-   ##
=======================================
  Coverage   88.67%   88.67%           
=======================================
  Files          15       15           
  Lines         459      459           
=======================================
  Hits          407      407           
  Misses         52       52           

Flag	Coverage Δ
linux	88.67% <100.00%> (ø)
windows	88.59% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ajeklund]

Is there are case where units is actually a list? Or is it always (now) a tuple? If so, the list type should simply be changed to tuple without the Union.

No, it is only called using a tuple. I've now committed the suggested change.

[CasperWA]

If you want to make it even more accurate, you can pass in the types that will be contained in the tuple, for example if it's two strings: tuple[str, str], or if it's a string first and then a boolean (True/False) then it would be: tuple[str, bool].
If it's an unknown number of the same type that will be contained, ellipsis (...) can be used, for example, if it will be an unknown number of strings it would be: tuple[str, ...].

[jesper-friis]

Since any sequence will work, I will suggest to state that explicitly:

Suggested change

	def infer_metadata(rec: np.recarray, units: tuple) -> dlite.Instance:
	def infer_metadata(rec: np.recarray, units: "Sequence") -> dlite.Instance:

This requires that we load from typing import Sequence.

[ajeklund]

I agree, but this does not work together with @CasperWA's suggestion to also specify an unknown number of strings. (That results in TypeError: Parameters to generic types must be types. Got Ellipsis.). I will hence go for tuple[str, ...].

[ajeklund]

If you want to make it even more accurate, you can pass in the types that will be contained in the tuple, for example if it's two strings: tuple[str, str], or if it's a string first and then a boolean (True/False) then it would be: tuple[str, bool]. If it's an unknown number of the same type that will be contained, ellipsis (...) can be used, for example, if it will be an unknown number of strings it would be: tuple[str, ...].

Thank you! I implemented the tuple[str, ...] suggestion.