Filters can be set for spots, announcements and WWV. You will find the -directories for these under /spider/filter. You will find some examples in -the directories with the suffix .issue. There are two types of -filter, one for incoming information and one for outgoing information. -Outgoing filters are in the form CALLSIGN.pl and incoming filters -are in the form in_CALLSIGN.pl. Filters can be set for both nodes -and users. -
-
All filters work in basically the same way. There are several elements -delimited by commas. There can be many lines in the filter and they are -read from the top by the program. When writing a filter you need to think -carefully about just what you want to achieve. You are either going to write -a filter to accept or to reject. Think of a filter as -having 2 main elements. For a reject filter, you would have a line or multiple -lines rejecting the things you do not wish to receive and then a default line -accepting everything else that is not included in the filter. Likewise, for an -accept filter, you would have a line or multiple lines accepting the things you -wish to receive and a default line rejecting everthing else. -
-
In the example below, a user requires a filter that would only return SSB spots -posted in Europe on the HF bands. This is achieved by first rejecting the CW -section of each HF band and rejecting all of VHF, UHF etc based on frequency. -Secondly, a filter rule is set based on CQ zones to only accept spots posted in -Europe. Lastly, a default filter rule is set to reject anything outside the filter. -
-
-
-$in = [
- [ 0, 0, 'r', # reject all CW spots
- [
- 1800.0, 1850.0,
- 3500.0, 3600.0,
- 7000.0, 7040.0,
- 14000.0, 14100.0,
- 18068.0, 18110.0,
- 21000.0, 21150.0,
- 24890.0, 24930.0,
- 28000.0, 28180.0,
- 30000.0, 49000000000.0,
- ] ,1 ],
- [ 1, 11, 'n', [ 14, 15, 16, 20, 33, ], 15 ], #accept EU
- [ 0, 0, 'd', 0, 1 ], # 1 = want, 'd' = everything else
-];
-
--
The actual elements of each filter are described more fully in the following -sections. -
-
The elements of the Spot filter are .... +
In the /spider/msg directory you will find a file called badmsg.pl.issue. Rename +this to badmsg.pl and edit the file. The original looks something like this ....
-[action, field_no, sort, possible_values, hops]
-
--
There are 3 elements here to look at. Firstly, the action element. This is -very simple and only 2 possible states exist, accept (1) or drop (0). -
-
The second element is the field_no. There are 13 possiblities to choose from -here .... -
-
-
- 0 = frequency
- 1 = call
- 2 = date in unix format
- 3 = comment
- 4 = spotter
- 5 = spotted dxcc country
- 6 = spotter's dxcc country
- 7 = origin
- 8 = spotted itu
- 9 = spotted cq
- 10 = spotter's itu
- 11 = spotter's cq
- 12 = callsign of the channel on which the spot has appeared
-
--
The third element tells us what to expect in the fourth element. There are -4 possibilities .... -
-
-
- n - numeric list of numbers e.g. [ 1,2,3 ]
- r - ranges of pairs of numbers e.g. between 2 and 4 or 10 to 17 - [ 2,4, 10,17 ]
- a - an alphanumeric regex
- d - the default rule
+
+# the list of regexes for messages that we won't store having
+# received them (bear in mind that we must receive them fully before
+# we can bin them)
+
+
+# The format of each line is as follows
+
+# type source pattern
+# P/B/F T/F/O/S regex
+
+# type: P - private, B - bulletin (msg), F - file (ak1a bull)
+# source: T - to field, F - from field, O - origin, S - subject
+# pattern: a perl regex on the field requested
+
+# Currently only type B and P msgs are affected by this code.
+#
+# The list is read from the top down, the first pattern that matches
+# causes the action to be taken.
+
+# The pattern can be undef or 0 in which case it will always be selected
+# for the action specified
+
+
+
+package DXMsg;
+
+@badmsg = (
+'B', 'T', 'SALE',
+'B', 'T', 'WANTED',
+'B', 'S', 'WANTED',
+'B', 'S', 'SALE',
+'B', 'S', 'WTB',
+'B', 'S', 'WTS',
+'B', 'T', 'FS',
+);
-
The fifth element is simply the hops to set in this filter. This would only -be used if the filter was for a node of course and overrides the hop count in -hop_table.pl. +
I think this is fairly self explanatory. It is simply a list of subject +headers that we do not want to pass on to either the users of the cluster or +the other cluster nodes that we are linked to. This is usually because of +rules and regulations pertaining to items for sale etc in a particular country.
-
So, let's look at an example spot filter. It does not matter in the example -who the filter is to be used for. So, what do we need in the filter? We need -to filter the spots the user/node requires and also set a default rule for -anything else outside the filter. Below is a simple filter that stops spots -arriving from outside Europe.
-
-
-$in = [
- [ 0, 4, 'a', '^(K|N|A|W|VE|VA|J)'], # 0 = drop, 'a' = alphanumeric
- [ 1, 0, 'd', 0, 1 ], # 1 = want, 'd' = everything else
- ];
-
-From version 1.48 onwards the interface to this has changed. You can now +use the commands set/badword to add words that you are not prepared +to see on the cluster, unset/badword to allow that word again and +show/badword to list the words that you have set.
-
So the filter is wrapped in between a pair of square brackets. This tells -Spider to look in between these limits. Then each line is contained within -its own square brackets and ends with a comma. Lets look carefully at the first -line. The first element is 0 (drop). Therefore anything we put on this line -will not be accepted. The next element is 4. This means we are filtering by -the spotter. The third element is the letter "a" which tells the program to -expect an alphanumeric expression in the fourth element. The fourth element -is a list of letters separated by the pipe symbol. +
If you have a previous /spider/data/badwords, the first time you start +the node, it will read and convert this file to the new commands. The old style +file will then be removed.
-
What this line does is tell the program to drop any spots posted by anyone in -the USA, Canada or Japan. +
+There are a number of commands that control whether a spot progresses +any further by regarding it as "bad" in some way.
-
The second line is the default rule for anything else. The "d" tells us this -and the line simply reads... accept anything else. +
A DX Spot has a number of fields which can checked to see whether they +contain "bad" values, they are: the DX callsign itself, the Spotter and +the Originating Node.
-
You can add as many lines as you need to complete the filter but if there are -several lines of the same type it is neater to enclose them all as one line. -An example of this is where specific bands are set. We could write this like -this .... +
There are a set of commands which allow the sysop to control whether a +spot continues:-
-[ 0,0,'r',[1800.0, 2000.0], 1],
-[ 0,0,'r',[10100.0, 10150.0], 1],
-[ 0,0,'r',[14000.0, 14350.0], 1],
-[ 0,0,'r',[18000.0, 18200.0], 1],
+set/baddx
+set/badspotter
+set/badnode
-
But the line below achieves the same thing and is more efficient .... +
These work in the same as the set/badword command, you can add +any words or callsigns or whatever to the appropriate database. For +example, to stop a spot from a particular node you do:
- [ 0, 0, 'r',
- [
- 1800.0, 2000.0, # top band
- 10100.0, 10150.0, # WARC
- 14000.0, 14350.0, # 20m
- 18000.0, 18200.0, # WARC
- [ ,1 ],
+set/badnode gb7djk gb7dxc
-
-
a bad spotter:
-
-# This is an example announce or filter allowing only West EU announces
-#
-# The element list is:-
-# 0 - callsign of announcer
-# 1 - destination * = all, <callsign> = routed to the node
-# 2 - text
-# 3 - * - sysop, <some text> - special list eg 6MUK, ' ', normal announce
-# 4 - origin
-# 5 - 0 - announce, 1 - wx
-# 6 - channel callsign (the interface from which this spot came)
-
-$in = [
- [ 1, 0, 'a', '^(P[ABCDE]|DK0WCY|G|M|2|EI|F|ON)' ],
- [ 0, 0, 'd', 0 ]
-];
+set/badspotter b0mb p1rat nocall
In this example, only the prefixes listed will be allowed. It is possible to -be quite specific. The Dutch prefix "P" is followed by several secondary -identifiers which are allowed. So, in the example, "PA" or "PE" would be ok -but not "PG". It is even possible to allow information from a single callsign. -In the example this is DK0WCY, to allow the posting of his Aurora Beacon. -
-
and some bad dx:
-
-# This is an example WWV filter
-#
-# The element list is:-
-# 0 - nominal unix date of spot (ie the day + hour:13)
-# 1 - the hour
-# 2 - SFI
-# 3 - K
-# 4 - I
-# 5 - text
-# 6 - spotter
-# 7 - origin
-# 8 - incoming interface callsign
-
-# this one doesn't filter, it just sets the hop count to 6 and is
-# used mainly just to override any isolation from WWV coming from
-# the internet.
-
-$in = [
- [ 1, 0, 'd', 0, 6 ]
-];
+set/baddx video wsjt
-
It should be noted that the filter will start to be used only once a user/node -has logged out and back in again. -
I am not going to spend any more time on these filters now as they will become -more "comprehensive" in the near future. +
You can remove a word using the appropriate unset command +(unset/baddx, unset/badspotter, unset/badnode) or list them +using one of show/baddx, show/badspotter and +show/badnode.