Support Non-Isomorphous Inputs

[kmdalton]

It may be impossible to supply multiple mtz files without --separate-files in the current experimental branch. I should check on this and write a test.

[kmdalton]

It turns out this issue was caused by me providing two mtzs which were not isomorphous. I no longer consider this a bug. However, It is a good idea to come up with a better way to handle this by default. Options:

• Refuse to merge non-isomorphous cells and provide a helpful error message
• Use the first cell in the list to determine the ASU for merging
• Use the average cell to determine the ASU for merging

[JBGreisman]

Just to comment on this -- the default behavior in rs for rs.concat(), DataSet.append(), DataSet.join(), and DataSet.merge() is to use the first cell in the list (or for the DataSet ones, to keep the cell of the calling DataSet). Each of the methods also have a check_isomorphous=True argument, that can be set to False to avoid this check.

Those methods and the check_isomorphous flag may provide a simple way to implement this sort of behavior.

[kmdalton]

Related: #136

[kmdalton]

Having mulled this isue over for quite some time, I'm afraid I have deemed it out of scope. It violates careless's contract with users to try to be clever about this. The mechanism by which users communicate cell info to careless is through their input mtz files, and we have provided a good tool to do that in reciprocalspaceship.