Handle non-alphanumeric characters in terms.

[ShawnMilo]

Similar to #367, except other characters, including - and @ will cause the same problem.
#367 was fixed by #382, but we should handle in a more universal manner. The current solution uses string.replace(), because Django's slugify adds hypens, which have the same problem.

• Find out which characters have the issue (hopefully from Elasticsearch's documentation.
• Create a function similar to slugify, but which handles this case.

Preferably have a reverse function for displaying the term in the template as well; the current solution just replaces all _ characters with spaces. It would be nice to return the "special" character instead. This would mean somehow encoding the original value. Perhaps even something as simple as base64 encoding and decoding. If base64 encoding is selected, that eliminates the need to research and replace specific characters, as long as = isn't a problem for Elasticsearch.