CSV import: bring relationships over with their type

Exports from Follow Up Boss, Lofty and similar systems carry the people around
a contact in repeating columns: Relationship 1 Name, Relationship 1 Email,
Relationship 1 Phone, Relationship 1 Type, and so on. The importer used to
ignore them, so a spouse, a child or a referral partner arrived either as
nothing at all or as a note somebody had to re-link by hand.

What you may notice

• The column mapper has a new Relationships group holding three sets of
  Name / Email / Phone / Type fields. Columns named the way those exports name
  them are matched for you when the file is uploaded.
• For each relationship on a row, the import finds the person if they are
  already in your account and creates them if they are not, then links the two
  contacts with the type from your file. Type is matched without regard to
  capitalisation, so spouse, Spouse and SPOUSE all arrive as Spouse.
• A type Sure Send cannot link — Fiancée, say — is imported as Other
  rather than failing the row, and the completion screen says which value it
  was and what it became.
• Contacts created for a relationship appear in the import’s own contact list
  and carry the import’s lead source, the same as contacts created from a row
  of their own.
• When a relationship names somebody who also has a row of their own in the
  same file, both end up on the one contact whichever order the rows are in.
• The name in a relationship column is matched forgivingly: Doe, Jane,
  Jane Doe and Jane M Doe are all read as the Jane Doe already on the
  record, rather than each producing a contact of their own.

Contacts already in your account are never rewritten

Relationship columns describe someone else; they are not an edit to a contact
they happen to match. A contact the import matches for a relationship keeps
everything it already had.

One case gets special care. If a relationship’s email address is stored as a
secondary address on some other contact — a spouse’s address kept on their
partner’s record, which is common — matching on it would attach the
relationship to the wrong person. When the name in your file does not agree
with that contact, the import creates the related contact separately and says
so on the completion screen, naming the contact it declined to use.

A primary email address is treated as an identity claim, so a relationship
is linked to the contact holding it even when the name in your file disagrees —
but the completion screen now tells you when that happened and whose name was
used. A relationship with an email address or phone number and no name at all
is still imported, and is now called out on the completion screen too, since
that contact has nothing to show in your lists.

What this doesn’t change

Re-importing the same file does not duplicate anything: a relationship that
already exists is left alone, including its type, so a type someone corrected
by hand stays corrected. Follow Up Boss’s own integration does not sync
relationships yet — this is the CSV path.