Fixing small bug in mWater api pull

[SamStruthers]

mWater decided to change the allowed forms of datetime which broke our parsing of the datetime columns (start_dt, end_dt and malfunction_end_dt). Work:

• Simple fix in clean_mwater_notes to correctly parse DT
• When this was joined with old field notes, end_dt became numeric values. Adding end_dt = as.POSIXct(NA, tz = "MST")) in clean_field_notes seemed to have resolved that and end_dt is a datetime object again

Review Req: Make sure the field notes seem to get pulled in correctly on your side and thoughts on whether we should move to using parse_datetime to give multiple formats in the event that this type of change occurs again

[SamStruthers]

@juandlt-csu I added the datetime parsing if you want to test using that instead of the changes I made before

[SamStruthers]

Neglected to merge to main before adding realistic range commits. @kathryn-willi check out the last couple commits for my recent changes. Work:

• created a scratch function grab_raw_data.R to quickly pull raw data. I needed to share data with a grad student and now there is a relatively easy way to query what you want from the flagged dataset and the pre 2023 dataset.
• created add_realistic_flag.R and its associated yml. This is essentially a copy of add_spec_flag but using realistic ranges rather than sensor ranges. the yml could be trimmed down to only the parameters of interst (ph, sc, temp) but I left everything else from sensor_spec
• small updates to figures for 2023, going to chat with MR about it next week to share with FC Review: Check out realistic ranges

[kathryn-willi]

Going to merge and start incorporating your new flag as well as a few others. Please stay tuned for another PR from me sometime next week! :)