Closed steven-murray closed 2 months ago
Attention: Patch coverage is 88.57143%
with 8 lines
in your changes are missing coverage. Please review.
Project coverage is 97.18%. Comparing base (
93f6fa3
) to head (bbb76f5
).
Files | Patch % | Lines |
---|---|---|
hera_cal/io.py | 88.23% | 8 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
@jsdillon I'm inclined to merge even though not all lines are covered -- the remaining lines are very difficult to cover right now (most of them will be covered once pyuvdata 3 is out by default, and then we can remove some other code branches)
Fine by me.
On Wed, May 22, 2024 at 12:45 PM Steven Murray @.***> wrote:
@jsdillon https://github.com/jsdillon I'm inclined to merge even though not all lines are covered -- the remaining lines are very difficult to cover right now (most of them will be covered once pyuvdata 3 is out by default, and then we can remove some other code branches)
— Reply to this email directly, view it on GitHub https://github.com/HERA-Team/hera_cal/pull/931#issuecomment-2125623205, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABIJK424Q54RAJCQFM45GHLZDTYVTAVCNFSM6AAAAABCTNBYC6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRVGYZDGMRQGU . You are receiving this because you were mentioned.Message ID: @.***>
Fixes #928
This fixes issues related to the
antpair2ind
update in pyuvdata v3.0. The biggest issue was that the output ofantpair2ind
is now no longer guaranteed to be an array of indices, but instead could be eitherNone
or aslice
object. Furthermore, the results of calls toantpair2ind
are now cached on the UVData object, and this cache needs to be refreshed on any update to the blt axis of the data. This is done inUVData.read()
but this method is over-written inHERAData
so it needed to be added manually.Bryna's comment: It also fixes things for some other changes in pyuvdata v3.0. There's an update to the telescope metadata handling, which has deprecation but I went ahead and fixed some of the places this comes up. Also,
write_vis
andwrite_cal
broke when we actually removed support for the current array shapes. In the process of fixing these, I decided that it would be easier and better to just change them to use the.new
methods on those objects. That reduces the amount of code that needs to be maintained in this repo. Note that this makes #851 unnecessary.