1Please read the README file in this directory first. 2.ex 3.Id $Id: procmailsc.man,v 1.15 2001/08/04 06:08:22 guenther Exp $ 4.TH PROCMAILSC 5 \*(Dt BuGless 5.na 6.SH NAME 7procmailsc \- procmail weighted scoring technique 8.SH SYNOPSIS 9.RB [ * ] 10.B "w^x condition" 11.ad 12.SH DESCRIPTION 13In addition to the traditional true or false conditions you can specify 14on a recipe, you can use a weighted scoring technique to decide if 15a certain recipe matches or not. When weighted scoring is used in a 16recipe, then the final score for that recipe must be positive for it 17to match. 18 19A certain condition can contribute to the score if you allocate it 20a `weight' 21.RB ( w ) 22and an `exponent' 23.RB ( x ). 24You do this by preceding the condition (on the same line) with: 25.RS 26.B w^x 27.RE 28Whereas both 29.B w 30and 31.B x 32are real numbers between @MIN32@.0 and @MAX32@.0 inclusive. 33 34.SH "Weighted regular expression conditions" 35The first time the regular expression is found, it will add 36.I w 37to the score. The second time it is found, 38.I w*x 39will be added. The third time it is found, 40.I w*x*x 41will be added. The fourth time 42.I w*x*x*x 43will be added. And so forth. 44 45This can be described by the following concise formula: 46.Sx 4 47 n 48 n k\-1 x \- 1 49w * Sum x = w * \-\-\-\-\-\-\- 50 k=1 x \- 1 51.Ex 52It represents the total added score for this condition if 53.B n 54matches are found. 55 56Note that the following case distinctions can be made: 57.TP 8 58x=0 59Only the first match will contribute w to the score. Any subsequent 60matches are ignored. 61.TP 62x=1 63Every match will contribute the same w to the score. The score grows 64linearly with the number of matches found. 65.TP 660<x<1 67Every match will contribute less to the score than the previous one. 68The score will asymptotically approach a certain value (see the 69.B NOTES 70section below). 71.TP 721<x 73Every match will contribute more to the score than the previous one. 74The score will grow exponentially. 75.TP 76x<0 77Can be utilised to favour odd or even number of matches. 78.PP 79If the regular expression is negated (i.e., matches if it isn't found), 80then 81.B n 82obviously can either be zero or one. 83.SH "Weighted program conditions" 84If the program returns an exitcode of EXIT_SUCCESS (=@EX_OK@), then the total 85added score will be 86.BR w . 87If it returns any other exitcode (indicating failure), the total added 88score will be 89.BR x . 90.PP 91If the exitcode of the program is negated, then, the exitcode will 92be considered as if it were a virtual number of matches. Calculation 93of the added score then proceeds as if it had been a normal regular 94expression with 95.B n=`exitcode' 96matches. 97.SH "Weighted length conditions" 98If the length of the actual mail is 99.B M 100then: 101.Sx 1 102* w^x > L 103.Ex 104will generate an additional score of: 105.Sx 4 106 @POW@ 107 / M \e 108w * | \-\-\- | 109 \e L / 110.Ex 111And: 112.Sx 1 113* w^x < L 114.Ex 115will generate an additional score of: 116.Sx 4 117 @POW@ 118 / L \e 119w * | \-\-\- | 120 \e M / 121.Ex 122.PP 123In both cases, if L=M, this will add w to the score. In the former case 124however, larger mails will be favoured, in the latter case, smaller 125mails will be favoured. Although x can be varied to fine-tune the 126steepness of the function, typical usage sets x=1. 127.SH MISCELLANEOUS 128You can query the final score of all the conditions on a recipe from the 129environment variable 130.BR $= . 131This variable is set 132.I every 133time just after procmail has parsed all conditions on a recipe (even if the 134recipe is not being executed). 135.SH EXAMPLES 136The following recipe will ditch all mails having more than 150 lines in the 137body. 138The first condition contains an empty regular expression which, because 139it always matches, is used to give our score a negative offset. 140The second condition then matches every line in the mail, and consumes 141up the previous negative offset we gave (one point per line). In the end, 142the score will only be positive if the mail contained more than 150 lines. 143.Sx 5 144:0 @BODY_GREP@@PASS_HEAD@ 145* \-150^0 146* 1^1 ^.*$ 147/dev/null 148.Ex 149Suppose you have a priority folder which you always read first. The next 150recipe picks out the priority mail and files them in this special folder. 151The first condition is a regular one, i.e., it doesn't contribute to the 152score, but simply has to be satisfied. The other conditions describe things 153like: john and claire usually have something important to say, meetings 154are usually important, replies are favoured a bit, mails about Elvis 155(this is merely an example :\-) are favoured (the more he is mentioned, the 156more the mail is favoured, but the maximum extra score due to Elvis will 157be 4000, no matter how often he is mentioned), lots of quoted lines are 158disliked, smileys are appreciated (the score for those will reach a maximum 159of 3500), those three people usually don't send 160interesting mails, the mails should preferably be small (e.g., 2000 bytes long 161mails will score \-100, 4000 bytes long mails do \-800). 162As you see, if some of the uninteresting people send mail, then the mail 163still has a chance of landing in the priority folder, e.g., if it is about 164a meeting, or if it contains at least two smileys. 165.Sx 11 166:0 @HEAD_GREP@@BODY_GREP@ 167* !^Precedence:.*(junk|bulk) 168* 2000^0 ^From:.*(john@home|claire@work) 169* 2000^0 ^Subject:.*meeting 170* 300^0 ^Subject:.*Re: 171* 1000^.75 elvis|presley 172* \-100^1 ^> 173* 350^.9 :\-\e) 174* \-500^0 ^From:.*(boss|jane|henry)@work 175* \-100^3 > 2000 176priority_folder 177.Ex 178If you are subscribed to a mailinglist, and just would like to read 179the quality mails, then the following recipes could do the trick. 180First we make sure that the mail is coming from the mailinglist. 181Then we check if it is from certain persons of whom we value 182the opinion, or about a subject we absolutely want to know everything 183about. If it is, file it. Otherwise, check if the ratio of quoted lines 184to original lines is at most 1:2. If it exceeds that, ditch the mail. 185Everything that survived the previous test, is filed. 186.Sx 15 187:0 188^@FROM@mailinglist-request@some.where 189{ 190 :0: 191 * ^(From:.*(paula|bill)|Subject:.*skiing) 192 mailinglist 193 194 :0 @BODY_GREP@@PASS_HEAD@ 195 * 20^1 ^> 196 * \-10^1 ^[^>] 197 /dev/null 198 199 :0: 200 mailinglist 201} 202.Ex 203For further examples you should look in the 204.BR procmailex (5) 205man page. 206.SH CAVEATS 207Because this speeds up the search by an order of magnitude, 208the procmail internal egrep will always search for the leftmost 209.I shortest 210match, unless it is determining what to assign to 211.BR @MATCHVAR@ , 212in which case it searches the leftmost 213.I longest 214match. 215E.g. for the leftmost 216.I shortest 217match, by itself, the regular expression: 218.TP 219.B .* 220will always match a zero length string at the same spot. 221.TP 222.B .+ 223will always match one character (except newlines of course). 224.SH "SEE ALSO" 225.na 226.nh 227.BR procmail (1), 228.BR procmailrc (5), 229.BR procmailex (5), 230.BR sh (1), 231.BR csh (1), 232.BR egrep (1), 233.BR grep (1), 234.hy 235.ad 236.SH BUGS 237If, in a length condition, you specify an 238.B x 239that causes an overflow, procmail is at the mercy of the 240.BR pow (3) 241function in your mathematical library. 242.PP 243Floating point numbers in `engineering' format (e.g., 12e5) are not accepted. 244.SH MISCELLANEOUS 245As soon as `plus infinity' (@MAX32@) is reached, any subsequent 246.I weighted 247conditions will simply be skipped. 248.PP 249As soon as `minus infinity' (@MIN32@) is reached, the condition will 250be considered as `no match' and the recipe will terminate early. 251.SH NOTES 252If in a regular expression weighted formula 253.BR 0<x<1 , 254the total added score for this condition will asymptotically approach: 255.Sx 3 256 w 257\-\-\-\-\-\-\- 258 1 \- x 259.Ex 260In order to reach half the maximum value you need 261.Sx 3 262 \- ln 2 263n = \-\-\-\-\-\-\-\- 264 ln x 265.Ex 266matches. 267