Python bindings are non-functional

[olepbr]

The Python bindings, merged in #184, are non-functional. Some immediate issues:

• get_energy_consumption_measures() does not actually return any consumption measures; it returns an empty list. This is because it just calls generate_topology() and iterates over the generated Topology's record_buffer, which is empty as the topology was just generated (Topology::new() does record_buffer =vec![]).
• The PR claimed to add Python documentation, but it merely added an empty Sphinx harness.
• The supplied Makefile does not work as intended: make install fails due to $(eval TARGET_WHEEL := $(shell ls ../target/wheels/scaphandre-${PACKAGE_VERSION}-*.whl)) looking for the wheel in the parent directory's target/ instead of its own target/, where maturin build was just run.
• Even the string representation of the class is broken: from scaphandre import Scaphandre; s = Scaphandre(); print(s); results in AttributeError, because the dataclass (which includes all fields in the generated __repr__() unless one specifies repr=False) is defined with sensor_name: str whilst the __init__() sets self.name.
• Furthermore, the bindings deviate from Scaphandre itself: why is the sensor called _scaphandre instead of sensor, for instance? Given that these are "official" bindings, I think they should have a structure similar to the Rust implementation, or at least be more clearly defined and documented.

I guess all of this ties in with #50; since it's not clear how to use Scaphandre as a Rust crate, making bindings to another language is potentially even less clear. Most of these issues are easily fixable, though, but I do question why these bindings were merged in the first place.

Anyway, I hope this is of help. :-)

[TheElectronWill]

Yes, I think that they should be removed until #50 is completely solved.

[bpetit]

Hi !

@fvaleye any clue on a fix ?

I'll merge @TheElectronWill's PR/removal for 1.0 if not fixed before.

Thanks.

[bpetit]

Removed until a new PR that fixes it :)