Closed bennahugo closed 5 years ago
In the future lets not merge until we're sure it runs on actual data.
@IanHeywood @o-smirnov please note to go from the qualified release tag 0.1.1 when you use this in production.
@sjperkins lets use the data set above as a qualification testing dataset and compare against the previous best chi^2. its a 4 antenna dataset
Why run master in production?
Anyway, this is an issue with dask and python3.5.
Could we make dask 1.2.0 an <= requirement in this case?
or python version >=3.6 in the setup.py
Regrettably, since this a private repo, we can't use travis to run tests https://github.com/ska-sa/tricolour/pull/27. I wonder if we could use the SDP jenkins...
Closing since there aren't any qualification tests anyway
We need to change the dask version. 3.5 is the last version of python on 16.04 systems
We can put the qualification test on a git lfs and run it manually?
Closed by #34