version 1, including all changes.
.
Rev |
Author |
# |
Line |
1 |
perry |
1 |
PROCMAILSC |
|
|
2 |
!!!PROCMAILSC |
|
|
3 |
NAME |
|
|
4 |
SYNOPSIS |
|
|
5 |
DESCRIPTION |
|
|
6 |
Weighted regular expression conditions |
|
|
7 |
Weighted program conditions |
|
|
8 |
Weighted length conditions |
|
|
9 |
MISCELLANEOUS |
|
|
10 |
EXAMPLES |
|
|
11 |
CAVEATS |
|
|
12 |
SEE ALSO |
|
|
13 |
BUGS |
|
|
14 |
MISCELLANEOUS |
|
|
15 |
NOTES |
|
|
16 |
AUTHORS |
|
|
17 |
---- |
|
|
18 |
!!NAME |
|
|
19 |
|
|
|
20 |
|
|
|
21 |
procmailsc - procmail weighted scoring technique |
|
|
22 |
!!SYNOPSIS |
|
|
23 |
|
|
|
24 |
|
|
|
25 |
[[__*__] __w^x condition__ |
|
|
26 |
!!DESCRIPTION |
|
|
27 |
|
|
|
28 |
|
|
|
29 |
In addition to the traditional true or false conditions you |
|
|
30 |
can specify on a recipe, you can use a weighted scoring |
|
|
31 |
technique to decide if a certain recipe matches or not. When |
|
|
32 |
weighted scoring is used in a recipe, then the final score |
|
|
33 |
for that recipe must be positive for it to |
|
|
34 |
match. |
|
|
35 |
|
|
|
36 |
|
|
|
37 |
A certain condition can contribute to the score if you |
|
|
38 |
allocate it a `weight' (__w__) and an `exponent' |
|
|
39 |
(__x__). You do this by preceding the condition (on the |
|
|
40 |
same line) with: |
|
|
41 |
|
|
|
42 |
|
|
|
43 |
__ w^x |
|
|
44 |
__Whereas both __w__ and __x__ are real numbers between -2147483647.0 and 2147483647.0 inclusive. |
|
|
45 |
!!Weighted regular expression conditions |
|
|
46 |
|
|
|
47 |
|
|
|
48 |
The first time the regular expression is found, it will add |
|
|
49 |
''w'' to the score. The second time it is found, |
|
|
50 |
''w*x'' will be added. The third time it is found, |
|
|
51 |
''w*x*x'' will be added. The fourth time ''w*x*x*x'' |
|
|
52 |
will be added. And so forth. |
|
|
53 |
|
|
|
54 |
|
|
|
55 |
This can be described by the following concise |
|
|
56 |
formula: |
|
|
57 |
|
|
|
58 |
|
|
|
59 |
n |
|
|
60 |
n k-1 x - 1 |
|
|
61 |
w * Sum x = w * ------- |
|
|
62 |
k=1 x - 1 |
|
|
63 |
It represents the total added score for this condition if __n__ matches are found. |
|
|
64 |
|
|
|
65 |
|
|
|
66 |
Note that the following case distinctions can be |
|
|
67 |
made: |
|
|
68 |
|
|
|
69 |
|
|
|
70 |
x=0 |
|
|
71 |
|
|
|
72 |
|
|
|
73 |
Only the first match will contribute w to the score. Any |
|
|
74 |
subsequent matches are ignored. |
|
|
75 |
|
|
|
76 |
|
|
|
77 |
x=1 |
|
|
78 |
|
|
|
79 |
|
|
|
80 |
Every match will contribute the same w to the score. The |
|
|
81 |
score grows linearly with the number of matches |
|
|
82 |
found. |
|
|
83 |
|
|
|
84 |
|
|
|
85 |
0 |
|
|
86 |
|
|
|
87 |
|
|
|
88 |
Every match will contribute less to the score than the |
|
|
89 |
previous one. The score will asymptotically approach a |
|
|
90 |
certain value (see the __NOTES__ section |
|
|
91 |
below). |
|
|
92 |
|
|
|
93 |
|
|
|
94 |
1 |
|
|
95 |
|
|
|
96 |
|
|
|
97 |
Every match will contribute more to the score than the |
|
|
98 |
previous one. The score will grow |
|
|
99 |
exponentially. |
|
|
100 |
|
|
|
101 |
|
|
|
102 |
x |
|
|
103 |
|
|
|
104 |
|
|
|
105 |
Can be utilised to favour odd or even number of |
|
|
106 |
matches. |
|
|
107 |
|
|
|
108 |
|
|
|
109 |
If the regular expression is negated (i.e., matches if it |
|
|
110 |
isn't found), then __n__ obviously can either be zero or |
|
|
111 |
one. |
|
|
112 |
!!Weighted program conditions |
|
|
113 |
|
|
|
114 |
|
|
|
115 |
If the program returns an exitcode of EXIT_SUCCESS (=0), |
|
|
116 |
then the total added score will be __w__. If it returns |
|
|
117 |
any other exitcode (indicating failure), the total added |
|
|
118 |
score will be __x__. |
|
|
119 |
|
|
|
120 |
|
|
|
121 |
If the exitcode of the program is negated, then, the |
|
|
122 |
exitcode will be considered as if it were a virtual number |
|
|
123 |
of matches. Calculation of the added score then proceeds as |
|
|
124 |
if it had been a normal regular expression with |
|
|
125 |
__n=`exitcode'__ matches. |
|
|
126 |
!!Weighted length conditions |
|
|
127 |
|
|
|
128 |
|
|
|
129 |
If the length of the actual mail is __M__ |
|
|
130 |
then: |
|
|
131 |
|
|
|
132 |
|
|
|
133 |
* w^x |
|
|
134 |
will generate an additional score of: |
|
|
135 |
|
|
|
136 |
|
|
|
137 |
x |
|
|
138 |
/ M \ |
|
|
139 |
w * | --- | |
|
|
140 |
\ L / |
|
|
141 |
And: |
|
|
142 |
|
|
|
143 |
|
|
|
144 |
* w^x |
|
|
145 |
will generate an additional score of: |
|
|
146 |
|
|
|
147 |
|
|
|
148 |
x |
|
|
149 |
/ L \ |
|
|
150 |
w * | --- | |
|
|
151 |
\ M / |
|
|
152 |
|
|
|
153 |
|
|
|
154 |
In both cases, if L=M, this will add w to the score. In the |
|
|
155 |
former case however, larger mails will be favoured, in the |
|
|
156 |
latter case, smaller mails will be favoured. Although x can |
|
|
157 |
be varied to fine-tune the steepness of the function, |
|
|
158 |
typical usage sets x=1. |
|
|
159 |
!!MISCELLANEOUS |
|
|
160 |
|
|
|
161 |
|
|
|
162 |
You can query the final score of all the conditions on a |
|
|
163 |
recipe from the environment variable __$=__. This |
|
|
164 |
variable is set ''every'' time just after procmail has |
|
|
165 |
parsed all conditions on a recipe (even if the recipe is not |
|
|
166 |
being executed). |
|
|
167 |
!!EXAMPLES |
|
|
168 |
|
|
|
169 |
|
|
|
170 |
The following recipe will ditch all mails having more than |
|
|
171 |
150 lines in the body. The first condition contains an empty |
|
|
172 |
regular expression which, because it always matches, is used |
|
|
173 |
to give our score a negative offset. The second condition |
|
|
174 |
then matches every line in the mail, and consumes up the |
|
|
175 |
previous negative offset we gave (one point per line). In |
|
|
176 |
the end, the score will only be positive if the mail |
|
|
177 |
contained more than 150 lines. |
|
|
178 |
|
|
|
179 |
|
|
|
180 |
:0 Bh |
|
|
181 |
* -150^0 |
|
|
182 |
* 1^1 ^.*$ |
|
|
183 |
/dev/null |
|
|
184 |
Suppose you have a priority folder which you always read first. The next recipe picks out the priority mail and files them in this special folder. The first condition is a regular one, i.e., it doesn't contribute to the score, but simply has to be satisfied. The other conditions describe things like: john and claire usually have something important to say, meetings are usually important, replies are favoured a bit, mails about Elvis (this is merely an example :-) are favoured (the more he is mentioned, the more the mail is favoured, but the maximum extra score due to Elvis will be 4000, no matter how often he is mentioned), lots of quoted lines are disliked, smileys are appreciated (the score for those will reach a maximum of 3500), those three people usually don't send interesting mails, the mails should preferably be small (e.g., 2000 bytes long mails will score -100, 4000 bytes long mails do -800). As you see, if some of the uninteresting people send mail, then the mail still has a chance of landing in the priority folder, e.g., if it is about a meeting, or if it contains at least two smileys. |
|
|
185 |
|
|
|
186 |
|
|
|
187 |
:0 HB |
|
|
188 |
* !^Precedence:.*(junk|bulk) |
|
|
189 |
* 2000^0 ^From:.*(john@home|claire@work) |
|
|
190 |
* 2000^0 ^Subject:.*meeting |
|
|
191 |
* 300^0 ^Subject:.*Re: |
|
|
192 |
* 1000^.75 elvis|presley |
|
|
193 |
* -100^1 ^ |
|
|
194 |
If you are subscribed to a mailinglist, and just would like to read the quality mails, then the following recipes could do the trick. First we make sure that the mail is coming from the mailinglist. Then we check if it is from certain persons of whom we value the opinion, or about a subject we absolutely want to know everything about. If it is, file it. Otherwise, check if the ratio of quoted lines to original lines is at most 1:2. If it exceeds that, ditch the mail. Everything that survived the previous test, is filed. |
|
|
195 |
|
|
|
196 |
|
|
|
197 |
:0 |
|
|
198 |
^From mailinglist-request@some.where |
|
|
199 |
{ |
|
|
200 |
:0: |
|
|
201 |
* ^(From:.*(paula|bill)|Subject:.*skiing) |
|
|
202 |
mailinglist |
|
|
203 |
:0 Bh |
|
|
204 |
* 20^1 ^ |
|
|
205 |
For further examples you should look in the procmailex(5) man page. |
|
|
206 |
!!CAVEATS |
|
|
207 |
|
|
|
208 |
|
|
|
209 |
Because this speeds up the search by an order of magnitude, |
|
|
210 |
the procmail internal egrep will always search for the |
|
|
211 |
leftmost ''shortest'' match, unless it is determining |
|
|
212 |
what to assign to __MATCH__, in which case it searches |
|
|
213 |
the leftmost ''longest'' match. E.g. for the leftmost |
|
|
214 |
''shortest'' match, by itself, the regular |
|
|
215 |
expression: |
|
|
216 |
|
|
|
217 |
|
|
|
218 |
__.*__ |
|
|
219 |
|
|
|
220 |
|
|
|
221 |
will always match a zero length string at the same |
|
|
222 |
spot. |
|
|
223 |
|
|
|
224 |
|
|
|
225 |
__.+__ |
|
|
226 |
|
|
|
227 |
|
|
|
228 |
will always match one character (except newlines of |
|
|
229 |
course). |
|
|
230 |
!!SEE ALSO |
|
|
231 |
|
|
|
232 |
|
|
|
233 |
procmail(1), procmailrc(5), |
|
|
234 |
procmailex(5), sh(1), csh(1), |
|
|
235 |
egrep(1), grep(1), |
|
|
236 |
!!BUGS |
|
|
237 |
|
|
|
238 |
|
|
|
239 |
If, in a length condition, you specify an __x__ that |
|
|
240 |
causes an overflow, procmail is at the mercy of the |
|
|
241 |
pow(3) function in your mathematical |
|
|
242 |
library. |
|
|
243 |
|
|
|
244 |
|
|
|
245 |
Floating point numbers in `engineering' format (e.g., 12e5) |
|
|
246 |
are not accepted. |
|
|
247 |
!!MISCELLANEOUS |
|
|
248 |
|
|
|
249 |
|
|
|
250 |
As soon as `plus infinity' (2147483647) is reached, any |
|
|
251 |
subsequent ''weighted'' conditions will simply be |
|
|
252 |
skipped. |
|
|
253 |
|
|
|
254 |
|
|
|
255 |
As soon as `minus infinity' (-2147483647) is reached, the |
|
|
256 |
condition will be considered as `no match' and the recipe |
|
|
257 |
will terminate early. |
|
|
258 |
!!NOTES |
|
|
259 |
|
|
|
260 |
|
|
|
261 |
If in a regular expression weighted formula |
|
|
262 |
__0__, the to- tal added score for this |
|
|
263 |
condition will asymptotically ap- proach: |
|
|
264 |
|
|
|
265 |
|
|
|
266 |
w |
|
|
267 |
------- |
|
|
268 |
1 - x |
|
|
269 |
In order to reach half the maximum value you need |
|
|
270 |
|
|
|
271 |
|
|
|
272 |
- ln 2 |
|
|
273 |
n = -------- |
|
|
274 |
ln x |
|
|
275 |
matches. |
|
|
276 |
!!AUTHORS |
|
|
277 |
|
|
|
278 |
|
|
|
279 |
Stephen R. van den Berg |
|
|
280 |
|
|
|
281 |
|
|
|
282 |
|
|
|
283 |
Philip A. Guenther |
|
|
284 |
|
|
|
285 |
|
|
|
286 |
|
|
|
287 |
---- |