fix Philter regex

[katie-ta]

summary

Changes include:

• a few fixes so philter.py would run properly, without errors
• fix all instances of letter-matching regular expressions to include accented characters
• use utf-8 encoding by default
  • ran into issues using the detected encoding on datasets that included accented characters; utf-8 works just as fine