Support unicode characters by replacing unidecode with new normalize method

[nllong]

Any background context you want to provide?

Back in 2016 we addd the unidecode library to fix unicode issues with the data. That worked well until now, where we need to keep diacritics/accent marks and further support the arabic character set.

What's this PR do?

• Remove unidecode
• Create new method to normalize set of characters that would prevent reasonable matches (e.g., mdash, fancy quotes, etc).
• Use the unicodedata.normalize method to force unicode characters to combine the letter and diacritic together. Using the NFC (Normalization Form Composition) setting which has extended functionality.

How should this be manually tested?

• unit tests
• import unicode data (new test file forthcoming)
• UI testing by inserting unicode characters. A great test is to edit a matching field to insert a unicode character, then import a new dataset with that unicode character in the matching field.

What are the relevant tickets?

4479

Screenshots (if appropriate)

[github-actions[bot]]

Label error. Requires at least 1 of: Feature, Bug, Enhancement, Maintenance, Documentation, Performance, Do not publish. Found: