Dedup the Perl_Companies files

[vmbrasseur]

There are a lot of dupes in these things. This needs to be cleaned up.

[thaljef]

I think the trick is to dedup the data whilst still being able to re-run the script to extract additional data from the mail or add future mail.

One possibility it to maintain a separate list of regexes that sanitize the data. For example Yahoo, Inc => /yahoo/i would mean to treat any company that matches m/yahoo/i as "Yahoo, Inc."

That still list has to be maintained by hand, which kinda sucks. But it is probably better than manually editing each mail.

[vmbrasseur]

Agreed, that would be a much better way to do it. But I didn't have the cycles to start that process (again: perfect was becoming enemy of the good; plus: other things to do ATM). Hopefully once the project is announced someone will be able to start the process with a pull request.