change import workflow to do field mapping in reverse of what currently exists

[takinbo]

When doing an import (e.g. participants or locations), we have internal data fields that are then mapped to columns in the import document. This isn't flexible and resorts to using things like column prefixes for fields that have multiple columns in the import document. This is a proposal to switch the import workflow to map column fields to internal data fields; this will allow us to do the following:

phone 1 -> phone number
phone 2 -> phone number
national -> sample
state -> sample

And this way we can tell that phone1 and phone 2 are columns of phone number type. We would also be able to import special data fields like this as well.

[takinbo]

It will also be important to ensure that during the mapping, certain fields can only be mapped once. Most fields will be this way except for ones that allow multiple values like phone numbers, groups and samples.

[takinbo]

A version of this has been implemented for location data imports.

[takinbo]

@dodumosu I have done a review of this for location imports and it works as expected. We would need this done for participants and also used for sample imports as well. I think once an option has been selected, it should not be available for selection again. E.g. once I have mapped Country Name, I shouldn't be able to map it to another field; there are exceptions like Phone Number, Group and Sample.

[dodumosu]

Participant import has been changed to this, a good while ago. Is it possible we got some feedback from the users on whether this is more intuitive, or the old method is better?