Error writing bctides.in when using add_Z0

[VHernaman]

Hi, since v0.0.10 there has been support for adding Z0 in bctides.in, but now using 0.1.15 I get an error when I write bctides.in having added Z0. The error is:
KeyError: 'Z0'

From the code I can't see 0.1.15 requires a different way of handling it, so not sure if this is a bug? I see Z0 is still accounted for in tides.py, so maybe it is a bug with bctides.write?

#For example, I instantiate an instance of bctides and then use add_Z0 to add Z0 to my list of constituents
bctides.tides.add_Z0(0.9)
#Check that Z0 has been added:
bctides.tides.get_active_forcing_constituents()
#returns
['M2', 'S2', 'N2', 'K2', 'K1', 'O1', 'P1', 'Q1', 'Z0']

#write it out
bctides.write(
outdir,
start_date=start_date,
rnday=30,
overwrite=True,
)

#But I get this error

KeyError Traceback (most recent call last)
Cell In[36], line 1
----> 1 bctides.write(
2 outdir,
3 start_date=start_date,
4 rnday=30,
5 overwrite=True,
6 )

File /srv/envs/cmextremes/.venv/lib/python3.11/site-packages/pyschism/forcing/bctides/bctides.py:265, in Bctides.write(self, output_directory, start_date, rnday, overwrite)
263 raise IOError("path exists and overwrite is False")
264 with open(bctides, "w") as f:
--> 265 f.write(str(self))

File /srv/envs/cmextremes/.venv/lib/python3.11/site-packages/pyschism/forcing/bctides/bctides.py:93, in Bctides.str(self)
91 if self.nbfr > 0:
92 for constituent in self.tides.get_active_forcing_constituents():
---> 93 forcing = self.tides(self.start_date, self.rnday, constituent)
94 f.append(
95 " ". join([
96 f"{constituent}\n",
(...)
101 ])
102 )
104 #get amplitude and phase for each open boundary

File /srv/envs/cmextremes/.venv/lib/python3.11/site-packages/pyschism/forcing/bctides/tides.py:119, in Tides.call(self, start_date, rnday, constituent)
111 """Returns the tidal characteristic values for a given time period.
112 """
113 _logger.debug(
114 f'Computing tidal factors for start_date={start_date} and '
115 f'constituent {constituent}')
116 return (self.get_tidal_species_type(constituent),
117 self.get_tidal_potential_amplitude(constituent), # TPK
118 self.get_orbital_frequency(constituent), # FF*
--> 119 self.get_nodal_factor(start_date, rnday, constituent), # Amig*
120 self.get_greenwich_factor(start_date, rnday, constituent))

File /srv/envs/cmextremes/.venv/lib/python3.11/site-packages/pyschism/forcing/bctides/tides.py:263, in Tides._manage_dates..decorator(self, start_date, rnday, constituent)
262 def decorator(self, start_date, rnday, constituent):
--> 263 val = f(self, start_date, rnday, constituent)
264 if hasattr(self, 'start_date_utc'):
265 del(self.start_date_utc)

File /srv/envs/cmextremes/.venv/lib/python3.11/site-packages/pyschism/forcing/bctides/tides.py:276, in Tides.get_nodal_factor(self, start_date, rnday, constituent)
271 @_manage_dates
272 def get_nodal_factor(self, start_date: datetime,
273 rnday: Union[float, timedelta],
274 constituent: str):
275 if constituent not in ALL_CONSTITUENTS:
--> 276 return self._nodal_factors[constituent]
277 if start_date.tzinfo is not None and
278 start_date.tzinfo.utcoffset(start_date) is not None:
279 self.start_date_utc = start_date.astimezone(timezone(timedelta(0)))

KeyError: 'Z0'

[SorooshMani-NOAA]

This looks like an issue introduced in d815c9e. I'm not sure if just adding Z0 to ALL_CONSTITUENTS is the right fix or not, but alternatively one can use

pyschism/pyschism/forcing/bctides/tides.py

Line 192 in fc41b51

def add_constituent(

in the current version of the code instead.

[mymyrere]

Hi,
I think it is similar problem, I see in the newer version a line has been added:
https://github.com/schism-dev/pyschism/blob/main/pyschism/forcing/bctides/tides.py#L275
but if the constituent is L2 for exemple, it is not in ALL_CONSTITUENTS but it is still define in:
https://github.com/schism-dev/pyschism/blob/main/pyschism/forcing/bctides/tides.py#L351

should line 275 be deleted?