In Rust regex, is there a similar function to filtered_re.h in RE2

[yangwentong]

In RE2, filtered_re.h provides a prefilter mechanism that helps in cutting down the number of regexps that need to be actually searched.
Like:

// filtered_re.h

  // Prepares the regexps added by Add for filtering.  Returns a set
  // of strings that the caller should check for in candidate texts.
  // The returned strings are lowercased and distinct. When doing
  // string matching, it should be performed in a case-insensitive
  // way or the search text should be lowercased first.  Call after
  // all Add calls are done.
  void Compile(std::vector<std::string>* strings_to_match);

I would like to ask if there is a function similar to Compile in Rust regex, or other functions provided in filtered_re.h?

[BurntSushi]

No. You can use the regex-syntax to look at the HIR and extract substrings yourself though if you want. And the regex crate does that internally and tries to do at least part of the filtered_re.h optimization on its own, automatically.

[yangwentong]

Thank you for answering my questions

[masklinn]

If anyone swings by and is still interested, I've given the port a shot: https://crates.io/crates/regex-filtered

It seems to work fine so far, though I won't discount having missed something (the underlying model of re2 and regex are not quite the same, and re2 makes use or some C++ things which are not available in rust so had to be worked around), and I've yet to profile it or bench it against re2.

Unlike re2 the prefilter is built-in: since regex depends on aho-corasick which seems like the best prefilter you could ask for, it seems natural to just use it.

[BurntSushi]

This is awesome! Thanks for sharing!

Does regex-filtered have a repository? I don't see it linked in the crate metadata.

[masklinn]

Yeah I kinda fucked up and forgot to fill in those fields before I published it, I tried to re-publish but obviously cargo told me 0.1.0 was already pushed, I figure I'll have to bump to 0.1.1 in order for it to take.

The code is at https://github.com/ua-parser/uap-rust/tree/main/regex-filtered

[BurntSushi]

Yeah you'd need to republish. Crates are immutable once published. (The only exception to that, AFAIK, are legal issues.)