Set EK60/80 Platform and NMEA nan timestamp to first ping_time value [all tests ci]

[emiliom]

Addresses the first two xarray warnings identified in #1153. Implements the solution described there, for EK60.

Setting single-valued nan Platform.time1 cases to the first ping_time value is consistent with the approach currently used with AZFP. EK80 may benefit from this change, too, but I'm not aware of a test case raw file, plus we're short on time.

See discussion in #1153.

[@leewujung edits for future reference]: Saildrone EK80 data by default does not have the lat/lon data since GPS is not plugged into the echosounder itself.

[leewujung]

EK80 may benefit from this change, too, but I'm not aware of a test case raw file, plus we're short on time.

The saildrone EK80 data by default does not have the lat/lon data since GPS is not plugged into the echosounder itself, so those would be the example raw files.

[codecov-commenter]

Codecov Report

Merging #1154 (0649bb4) into dev (5abca87) will increase coverage by 0.02%.
The diff coverage is 94.11%.

@@            Coverage Diff             @@
##              dev    #1154      +/-   ##
==========================================
+ Coverage   82.32%   82.35%   +0.02%     
==========================================
  Files          64       64              
  Lines        5789     5803      +14     
==========================================
+ Hits         4766     4779      +13     
- Misses       1023     1024       +1     

Flag	Coverage Δ
unittests	82.35% <94.11%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
echopype/convert/set_groups_base.py	86.70% <88.88%> (+0.13%)	⬆️
echopype/convert/set_groups_azfp.py	97.67% <100.00%> (+0.05%)	⬆️
echopype/convert/set_groups_ek60.py	100.00% <100.00%> (ø)
echopype/convert/set_groups_ek80.py	97.57% <100.00%> (+0.01%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[leewujung]

Ok, I ended up discovering more cases like this using the EK80 file. To summarize, these are:

• EK60: Platform.time1
• EK80: Platform.time1, Platform.time2
• Platform/NMEA.time (this one did not show up for the OOI files because they happen to have a single, non-GPS NMEA message)

Since they all can be handled in the same way, I added a new private method SetGroupsBase._nan_timestamp_handler for correcting a single nan value timestamp.

@emiliom: I changed your EK60 implementation to use the code I added for EK80, to account for cases where the channels may not ping simultaneously. I also changed the AZFP set_platform code to handle situations with missing timestamp via the _nan_timestamp_handler function.

[leewujung]

I'll self-merge this now to check everything is in good order for releasing v0.8.1.