Normalize various kinds of punctuation and/or quotation marks in range-finder script

[ryderwishart]

Checking unique characters that follow a backslash in the Spanish alignments (command: grep -o '\\.' spapddpt_alignments.jsonl_final_output.jsonl | awk '{print $NF}' | sort -u ):

\"
\\
\n

And in the French file:

\"

[ryderwishart]

One thing we might try is using uroman to first romanize both sides of the equation, but this will only work if uroman always outputs one character for one input character.