writing obs sequences after calculating stats

[hkershaw-brown]

Drop columns currently

pyDARTdiags/src/pydartdiags/obs_sequence/obs_sequence.py

Lines 369 to 370 in 8660d99

	if "bias" in df_copy.columns:
	df_copy = df_copy.drop(columns=["bias", "sq_err"])

So if you have binned by level, you have a vlevels and a midpoint
Should this be drop anything that is not in the header or:
start_required_columns = ["obs_num", "observation"]
end_required_columns = [
"linked_list",
"longitude",
"latitude",
"vertical",
"vert_unit",
"type",
"metadata",
"external_FO",
"seconds",
"days",
"time",
"obs_err_var",

Or explicit, midpoint, vlevels?

[hkershaw-brown]

You can bin a obs_seq.out (no assimilation info), which I think is something you might want to look at, e.g you might want to remove obs above/below a certain level.

[hkershaw-brown]

Also this drop is wrong now since we calculate for prior and post
df_copy = df_copy.drop(columns=["bias", "sq_err"])