Creating plotting.py with pandas-flavor

[FloraSauerbronn]

decorating the plot transect method.

[ocefpaf]

Suggested change

	ax.set_ylabel("Depth (m)")
	ax.set_ylabel("pressure")

[ocefpaf]

@FloraSauerbronn do you mind adding a description above, you can edit your empty box, so we have something for future reference? You can go as simple as "decorating the plot transect method."

BTW, one place where you oceanography background will com in handy, is in implementing some improvements to this function. For example, you can add an if-clause to error out if the variable chosen for plotting is not in an expected list of vars that would work. At the moment I believe that we can only do temperature and salinity.

[ocefpaf]

We will need to import the plotting routine in the __init__.py to make it available as a DataFrame method. You can add something like:

from .plotting import plot_transect
from .fetchers import GliderDataFetcher


__all__ = [
    "GliderDataFetcher",
    "plot_transect",
]

to the __init__.py file. The GliderDataFetcher is not required for this PR but it is something we should've done a while back.

[FloraSauerbronn]

About the if clause @ocefpaf

--> When you say 'not in an expected list of vars,' do you mean they have a lot of NaNs? Or just outliers? Because if it's just NaNs causing the problem, it's simple, but if it's outliers, then we would need to consider the values of the limit more carefully.

[ocefpaf]

About the if clause @ocefpaf

--> When you say 'not in an expected list of vars,' do you mean they have a lot of NaNs? Or just outliers? Because if it's just NaNs causing the problem, it's simple, but if it's outliers, then we would need to consider the values of the limit more carefully.

I mean the variables names. In the plot_transect we have one mandatory argument:

:param var: variable to colour the scatter plot

That can be either salinity or temperature, based on the list of variables we have are fetching:

server_vars = {
    "https://gliders.ioos.us/erddap": [
        "latitude",
        "longitude",
        "pressure",
        "profile_id",
        "salinity",
        "temperature",
        "time",
    ],
}

None of the others variables make sense to be in a section plot, right? However, let's keep this first PR simple and we can address that in a follow up one. reducing the context of the PRs will help us keep the momentum going and code changes/reviews easier to manage.

[ocefpaf]

pre-commit.ci autofix