@davidfor Not sure what your preference is on community contributions, but I was playing around with Calibre plugins and made a few modifications to the count_pages plugin. Namely, a few more sources for page counts (Amazon, AmazonJP, Google) and an additional readability algorithm specifically for Japanese text. It doesn't have the latest update merged in but it is available on my GitHub
https://github.com/sharder996/count_pages if you're interested.