Fix for 116

[vjrj]

This PR fix #116. Basically:

• Takes into account other contact fields
• Update also contacts without email
• Prevent to incorrect associate drs to the first contact without email as Contact.findByEmail(empty) was picked in this case.
• Do not fail if an email has a trailing or leading space.

[vjrj]

I think we can refactor this code a bit like this for better results:

First the EMLimportService gets a list a current eml contacts in a dr and tries to find the contacts in the db. 1) If the contact already exist (looking for the email or the full name), update them if we have more info (a name or an email, etc) 2) If a contact is not in the contact table, add it 3) take into consideration if a contact is primary or not and maintain an additional list of primary contacts
return the current updated contact list and the primary contact list

In the IptService we receive that new contact lists and: 4) maintain a list of old contacts in a variable 5) remove all contact_for for that dr 6) relink the new in sync contacts list created in EMLImportService to that dr in contacts_for (as primary or not as we have a list created in 3) 7) loop the old contacts list of step 4) and remove orphan contacts that does not have an entry in contact_for.

With this I think we'll have the contacts exactly with the same info as the EML and the IPT, and contacts not longer linked to a dr will be removed. I think I'm not missing anything, but...

[vjrj]

I added a first part that update existing contacts too taking into consideration if the contact is primary or not.

I can continue with the rest or do it in a new PR in the near future.

This PR solves a lot of issues (the most severe is that currently the first contact without email get associated all drs of contacts without emails, that is a big list for one contact).