dotnaught/spellerizer

An extension for Google Chrome that allows you to replace words with US spellings in The Register articles with UK spellings.

spellerizer's Issues

Switch to regex-based matching to allow for context matches

As stated in your article, currently you just replace "check" with "cheque" because you are only operating on individual words.
If you replaced the check match with ((a|the)\s)?check then you would match "a check" or "the check", and you would just have to replace it with '\1cheque(although you might want to replace the central section manually to case match it). You could also add acheck( please)` match.
You could alternatively specify it in the JSON as a context prefix/suffix where you would specify it as a regex (and when building the actual regex, add the case insensitive flag and probably lookahead/look behind so the context matched could also be replaced if needed.

I am using check as an example, I'm sure that there are other words that would benefit from context matching.

Edit: I have just read your article further, and noticed that the spelling_data.json file was not written by you and this would require changes to that file / format