Pragmatic solution for calculating the similarity of text / words of database entires in PostgreSQL

psql_text_similaritry_operator.sql

--- If you are missing the pypythonu extension, install it first
--- (Ubuntu): sudo apt-get install python-psycopg2

--- Activate plpython 
CREATE EXTENSION plpythonu;

--- Create the profiling function
CREATE OR REPLACE FUNCTION similarity(text, text) RETURNS numeric AS $$
import difflib
return difflib.SequenceMatcher(None,args[0], args[1]).ratio()
$$ LANGUAGE 'plpythonu' IMMUTABLE STRICT;

--- Optional: Create an operator for this function, in this case '<#>'
--- Otherwise use it like 'similarity(name1,name2)' in your query
CREATE OPERATOR <#>
(PROCEDURE=similarity,
LEFTARG=text,
RIGHTARG=text);

--- Values go from 0.0 to 1.0
--- Execute some example queries to get a hang for the values and their actually meaning of similarity. 
--- Example: SELECT * FROM employees e where similarity('steffen',e.name) > 0.5
--- Queries for Employee Entries with a name similar to 'steffen', like 'stefan' etc...
--- I'd consider > 0.5 as 'very similar'