Loading .ang files using orix.io.load() can produce duplicate phase names

[gsparks3]

In _get_phases_from_header() under orix/io/plugins/ang.py, there is a section of code (permalink) that replaces the phase names loaded from the .ang file header with phase formulas (also loaded from said file header) if a formula exists for each phase. This does tend to reduce the verbosity of the automatically loaded phase names. However, it is possible for multiple distinct phases to have the same formula - in my case, alpha titanium and beta titanium both had a formula of "Ti". The automatic replacement causes multiple phases to have the same name when loading such a file using orix.io.load(), which breaks the ability to select data from individual phases in a CrystalMap by phase name.

Thankfully, the phases in the loaded CrystalMap are still distinct outside of the identical names, so the phase names can be manually overridden after loading, at which point everything works properly again. However, this seems like a sufficiently common situation that it should be addressed more directly, especially since the default phases in APEX (the most recent EDAX software for EBSD data acquisition) are "Generic FCC" and "Generic BCC", both with a formula of "Fe", on top of the issue with dual-phase Ti scans.

A few possible options:

• Remove the name-replacement code entirely
• Add a check so that the replacement is only done if each phase has a unique formula
• Detect duplicates phase names and raise a warning so the user knows they need to manually override the duplicates
  (they will probably find out once their code breaks, but the warning will make the cause a little more obvious)