Remove lambda-dependency from all phenomenological background models

[luisfabib]

No description provided.

[luisfabib]

The lambda dependency has been removed from all phenomenological background models. However, they still can take lambda as a third input. It has no effect on the math, so it is basically useless in the phenomenological models.

Now we have two alternatives:

• Keep the lambda input for the phenomenological models even though it has no effect whatsoever. This would profit of a cleaner code everywhere else.
• Remove the lambda input for the phenomenological models. This would require all functions which use these models, e.g. fitsignal, to have a check of hard-coded model-names to recognize which ones take lambda and which ones not.

At the moment I'm slightly in favor of the first one, provided that the documentation clearly shows that phenomenological models do not depend on lambda. This would solve the issue, albeit I would agree not cleanly. It would give us though time to find a cleaner and more elegant solution while the "math" is already correct.

[stestoll]

I think the second alternative is better. A function shouldn't take a positional argument if it doesn't use it. But I understand that this creates some other issues. Maybe signature from the inspect module, or something similar, could help.

[luisfabib]

Did not know about the inspect.signature function. Yeah, that should take care of most potential issues.