On Tue, Aug 26, 2014 at 11:05 AM, Skip Montanaro <s...@pobox.com> wrote: > I guess I could write a little program that listens to my incoming > email via IMAP. I'll have to see what that generates. Lots of Python > and bike references, no doubt.
I should have something to show the world in a day or two. I have a program which (in a separate thread) connects to a mail server via IMAP, processes all the messages in the "polly" folder, and constructs a word list from the most frequently seen words. It checks every minute for new messages. I believe it classifies words using criteria similar to Chris's D&D Polly - all letters, all lower case, nothing < four letters. In the main thread you can enter commands like "password" to prompt Polly to generate a password and "bad" to mark one or more words as "bad". Gmail's label facility made the construction of a "corpus" particularly easy. I just created a new filter and told it to add the label "polly" to messages with certain properties (stuff sent here, and to a couple bike-related and swimming-related lists). Voila! An instant corpus with over 30,000 emails. As an aside, I found it interesting that one of the "words" it saved was "luofeiyu". Man, that guy posts a lot! Skip -- https://mail.python.org/mailman/listinfo/python-list
