adding hack for lco bug fix

[havok2063]

This PR adds a hack to remove LCO observations from query results. This tacks on an additional where clause to the append_pipes method which filters on the astra fields obs=apo and release=sdss5. This still includes LCO observations from DR17 processed by Astra. The append_pipes method already includes a filter applying the MJD cutoff for a given release. It currently gets used in the following endpoints:

• main search form (UI)
• cone search
• carton-program search
• target alternate id search
• target sdss id search

This is some but not all of the endpoints. @Sean-Morrison can you take a look at this? Is this a sufficient enough query to filter out the errant data? And are these endpoints sufficient?

[Sean-Morrison]

At first look, the filter seems a bit simple since DR17 reduced LCO spectra are also included for DR19 Astra, and I think the hack would exclude these.

Additionally, those endpoints seem adequate for the Zora interface, but if we release the API, then it seems like people could use functions like get_targets_obs to circumvent the hack.

[havok2063]

ahh yeah you're right about the DR17 LCO observations. This misses those. Yes you're right about the API. I was hoping not to have to worry about it, but I guess it's better to be complete. I'll need to change how this gets implemented then to apply to every endpoint.

[havok2063]

@Sean-Morrison I've update the code so the lco hack now applies to the api target/id, target/spectra, and target/pipelines endpoints. The API still returns LCO DR17 observations. The UI would display something like the attached image. One thing is that source catalog and carton info is still displayed on those tabs. Do you think we need to hide this information as well?

[Sean-Morrison]

This looks good to me.