http://bugzilla.spamassassin.org/show_bug.cgi?id=2853
------- Additional Comments From [EMAIL PROTECTED] 2004-05-24 13:34 ------- BTW, I *do* like the idea of distributing and improving usability of these scripts by default, to lower the barrier-to-entry for mass-checking, running perceptron etc. I'm +1 on the patch contents. (I also am -1 on the idea of rewriting the perceptron. can't see the point in doing that, but I don't think that's what Duncan was really suggesting, fwiw ;) I'd suggest a further tweak, BTW: for the "masses" scripts that DAF is talking about putting into a -tools pkg, I suggest we rename them to use an "sa-" prefix. e.g. sa-mass-check, sa-hit-frequencies, et al. This helps to keep them logically together when installed, and is consistent with sa-learn. (we should probably do similar for the "tools" scripts too eventually). I'd also suggest some renaming to be clearer about what they do in reality, e.g.: mass-check sa-mass-scan extract-message-from-mbox sa-mbox-get hit-frequencies sa-msr-to-freqs lint-rules-from-freqs sa-rules-lint logs-to-c sa-msr-compile fp-fn-statistics sa-accuracy-report perceptron sa-perceptron rewrite-cf-with-new-scores sa-scores-rewrite also these scripts are very useful in rule dev, QA, and scoring: mass-check-results-to-mbox sa-msr-to-mbox overlap sa-msr-rule-overlap corpora/remove-tests-from-logs sa-msr-strip-rules (Note: "sa-rules-lint" should be modified to do a basic lint (without reading logs from STDIN) unless a (new) cmdline flag is supplied. Most of its linting behaviour does not require freqs data at all.) Also, would it be a good idea to come up with a file format name instead of "mass-check log file"? e.g. "MSR" for "mass-scan results"? (that's what sa-msr-to-mbox et al refers to.) ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
