[Feature] Support normalization/placeholders during extraction to deduplicate near-identical strings

[michael-proulx]

Many pages differ only by numbers or small text (e.g., <title>Top news - Page 60 - My Blog</title>), but Rosey will generate a key for every pages.

[
  "key57": "Top news - Page 57 - My Blog",
  "key58": "Top news - Page 58 - My Blog",
  "key59": "Top news - Page 59 - My Blog",
  "key60": "Top news - Page 60 - My Blog",
  ...
]

It would be nice to be able to split that key into small re-usable chunk e.g.

  "key1": "Top news"
  "key2": "Page {{ n }}"
  "key3": "My Blog"

Supporting Custom Params, Regex or maybe re-use a Rosey key e.g.

<title data-rosey-regex="{"/Page\s+(\d+)/": "rosey.key2", "/My Blog/": "rosey.key3"}">
  Top news - Page 60 - My Blog
</title>

Those are suggestions, it doesn't cover every edge cases, if you got any other/better ideas.
Like in the case of my title we could split by -.