Build: #3 was successful Changes by Pedro Salas <salas.mz@gmail.com>
Code commits
Casa6
-
Pedro Salas <salas.mz@gmail.com> c768e73d68bcd5b6d1a5de065ad9aaade9e635d5
Enable forcing clock1 to be offset from clock0
Adds force_clock_corr_offset option that forces clock1 to be
the same as clock0 with an offset. -
Pedro Salas <salas.mz@gmail.com> 44822b82999f868189936fd0f14004e3a41238a4
Fix missing key error for flagged data
The cost function would try to index data that was none
existent because of the crappy handling of the reference
antenna list. It now avoids things for which there's no
dictionary entry.
Also, run black on the code. -
Pedro Salas <salas.mz@gmail.com> fa4083d4c5d843b4a4ee47eeaae0a1face439117
Add output caltable
-
Pedro Salas <salas.mz@gmail.com> 42ecd44a8a4667fade84e91064b1197d74741296
Support for correlation products
Uses both correlation products when deriving solutions.
Each has its own clock term. It also removes antennas
that had all their solutions flagged. -
Pedro Salas <salas.mz@gmail.com> a1f674bc68f99233c61627f2a9ea5c7aebfc8e78
Use flags
Only unflagged data will be used to determine solutions -
Pedro Salas <salas.mz@gmail.com> 0262ff6aa3ecdb7f5cdd0eb663bc693580e1503e
Preliminary zenithdelaycal implementation
tl/dr: this needs more work; polarization, flags, antenna parsing.
Implements the scipy-based least squares solver for the
zenith delay. It only handles one polarization, it will likely
raise an exception if the antenna parameter is a string, and it
ignores flags in the multi band delay solutions.