I've written a Java-based lexer/parser that can convert straight quotes to
curly quotes for English prose. It's a one-pass algorithm (O(n)) that uses
neither look-behind nor regex. Here's a list of test cases it handles:

https://raw.githubusercontent.com/DaveJarvis/keenquotes/main/lib/src/test/resources/com/keenwrite/quotes/smartypants.txt

A test harness converted several Project Gutenberg texts quite well. The
folks at PG may be interested in using it themselves to help convert quotes
in older texts en masse. The source code is MIT-licensed:

https://github.com/DaveJarvis/keenquotes/

The code should port to Lua fairly easily, should anyone be interested in
adding a straight/curly quotation mark conversion module to ConTeXt.
(Similar to the LaTeX package, but without using regex.)

Cheers everyone!