[rules-users] Partial Unification / Derivation of Facts?
NOTE: Reposted because I only just subscribed to the mailing list. Hello, I have a set of Fact types which represent factors matched and used throughout the other rules. Each of these Fact types may be derived in a variety of mutually exclusive ways--ie based on the line of business or other factors. The rules are all well-defined for these Facts. Something like this: declare FactorType1 value : BigDecimal end rule FactorType1 when LineOfBusiness( this == LineOfBusiness.LOB1 ) // some other conditions $prms : Params() // ***provisional*** where to specify needed context dependent parameters? see below then BigDecimal value = new BigDecimal(0); // ... determine value based on conditions and $prms insert( new FactorType1(value) ); end // some other mutually exclusive rules to derive FactorType1 here The problem is that in the rules which match these Facts, there are many context dependent (ie *always* unique to that particular rule) parameters which must be passed to the derivation of those Facts. I'm referring to these as Params above. Now, a rule that uses FactorType1: rule a rule that uses FactorType1 when // some conditions FactorType1( $value ) $prms : Params(prm1,prm2,prm3,...) // ***provisional*** how to indicate that FactorType1 should use these parameters in its derivation? then // perform some calculation with $value end The parameters which these Facts use cannot be asserted as Facts themselves. They are unique to the rules which match on the FactorTypes, so that would mean I'd need a duplicate rule for every rule that uses a FactorType (even with rule inheritance, this is excessive). It's almost as if I need to specify a partial derivation or partial unification of the Fact type, which is only asserted once a rule indicates that it is providing the unbound portion of the derivation. I understand queries can be used for some sort of partial unification, but I have no clue how to apply them to my problem and I can't find any good examples or documentation on their usage/behavior. Thanks in advance -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3372546.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
It's almost as if I need to specify a partial derivation or partial unification of the Fact type, which is only asserted once a rule indicates that it is providing the unbound portion of the derivation. I understand queries can be used for some sort of partial unification, but I have no clue how to apply them to my problem and I can't find any good examples or documentation on their usage/behavior. I don't fully understand what you are trying to achieve, but I think what you are asking for is the Trait support that has just been introduced in the latest release? Have a look at the latest release notes and see if that helps you. Thomas ** This message is confidential and intended only for the addressee. If you have received this message in error, please immediately notify the postmas...@nds.com and delete it from your system as well as any copies. The content of e-mails as well as traffic data may be monitored by NDS for employment and security purposes. To protect the environment please do not print this e-mail unless necessary. NDS Limited. Registered Office: One London Road, Staines, Middlesex, TW18 4EX, United Kingdom. A company registered in England and Wales. Registered no. 3080780. VAT no. GB 603 8808 40-00 ** ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
This uses several terms I'm not familiar with in this context, but it seems that 1. a ule uniquely determines a Class to be instantiated on its RHS 2. a set of parameters is required that, in turn, depends on the Class determined according to (1). I do not understand why these parameters cannot be inserted as facts. Each parameter set (!) would be a fact with one field containing the Class, and this could be matched in the rule, passing parameters to the processing to be done in the same rule. $p: Params( clazz == (FactorType1.class), $pSet: paramSet ) -W On 27 September 2011 15:11, JohnnyCaimbridge nbo...@nycm.com wrote: NOTE: Reposted because I only just subscribed to the mailing list. Hello, I have a set of Fact types which represent factors matched and used throughout the other rules. Each of these Fact types may be derived in a variety of mutually exclusive ways--ie based on the line of business or other factors. The rules are all well-defined for these Facts. Something like this: declare FactorType1 value : BigDecimal end rule FactorType1 when LineOfBusiness( this == LineOfBusiness.LOB1 ) // some other conditions $prms : Params() // ***provisional*** where to specify needed context dependent parameters? see below then BigDecimal value = new BigDecimal(0); // ... determine value based on conditions and $prms insert( new FactorType1(value) ); end // some other mutually exclusive rules to derive FactorType1 here The problem is that in the rules which match these Facts, there are many context dependent (ie *always* unique to that particular rule) parameters which must be passed to the derivation of those Facts. I'm referring to these as Params above. Now, a rule that uses FactorType1: rule a rule that uses FactorType1 when // some conditions FactorType1( $value ) $prms : Params(prm1,prm2,prm3,...) // ***provisional*** how to indicate that FactorType1 should use these parameters in its derivation? then // perform some calculation with $value end The parameters which these Facts use cannot be asserted as Facts themselves. They are unique to the rules which match on the FactorTypes, so that would mean I'd need a duplicate rule for every rule that uses a FactorType (even with rule inheritance, this is excessive). It's almost as if I need to specify a partial derivation or partial unification of the Fact type, which is only asserted once a rule indicates that it is providing the unbound portion of the derivation. I understand queries can be used for some sort of partial unification, but I have no clue how to apply them to my problem and I can't find any good examples or documentation on their usage/behavior. Thanks in advance -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3372546.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
Let me start from scratch as I'm realizing how difficult it is to understand my issue. There is a Fact A with some rules use as conditions. A can be derived in several different but mutually exclusive ways. For the most part, the rules which define each of these derivations depend on states/conditions which the rules matching on A are largely agnostic of. rule rule deriving A when //some conditions then insert( new A(/*some values*/) ); end rule another rule driving A when //some other conditions then insert( new A(/*some other values*/) ); end // many other rules deriving A However, there are also other parameters/values needed for the computation of the many derivations of A (they need to be present in the RHSs of rules deriving A). The combination of these values are context-dependent to each matching of A /every time/. That is, each matching of A has a unique set of values binded to a particular set of parameters expected by rules deriving A). Rules deriving A need not and should not match on these parameters because they are different for every matching of A; the derivations should (ideally) take on whatever values are specified in the matching. rule rule matching on A when $a : A() // needs to somehow convey to the derivation of A() that it should use a set of values unique to this matching in its derivation then // use $a end rule another rule matching on A when $a : A() // needs to somehow convey to the derivation of A() that it should use a set of values unique to this matching in its derivation then // use $a end I cannot leave it up to the rules deriving A to handle every single case of different values for these parameters--there would be thousands of different combinations of derivations of A and nobody would understand the code. The rules deriving A cannot match on general Parameter objects because the values in the Parameter objects are unique for each matching of A. This would mean that each rule matching A would need a counterpart rule which asserts the parameters/values needed by the rules deriving A. I mean, I suppose this /is/ a solution, but it's obviously difficult to understand design. I guess what I'm looking for is something inbetween a function (with explicitly binded parameters) and a derivation that is rule based--to be able to say derive this kind of object in any way that you can, but use these particular values in its derivation. Perhaps this is not possible. -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3372828.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
All right: if it is the matching rule that defines the parameter set you can either define the parameters literally in that rule or you can create a lookup-table with the rule name as its key. You can obtain the rule name generically on the RHS, and you can write some code to make sure that all matching rules identified by some code in their name have an entry, etc. The deriving rules don't matter - there is no 1:1-correspondence between them and the matching ones. -W On 27 September 2011 16:53, JohnnyCaimbridge nbo...@nycm.com wrote: Let me start from scratch as I'm realizing how difficult it is to understand my issue. There is a Fact A with some rules use as conditions. A can be derived in several different but mutually exclusive ways. For the most part, the rules which define each of these derivations depend on states/conditions which the rules matching on A are largely agnostic of. rule rule deriving A when //some conditions then insert( new A(/*some values*/) ); end rule another rule driving A when //some other conditions then insert( new A(/*some other values*/) ); end // many other rules deriving A However, there are also other parameters/values needed for the computation of the many derivations of A (they need to be present in the RHSs of rules deriving A). The combination of these values are context-dependent to each matching of A /every time/. That is, each matching of A has a unique set of values binded to a particular set of parameters expected by rules deriving A). Rules deriving A need not and should not match on these parameters because they are different for every matching of A; the derivations should (ideally) take on whatever values are specified in the matching. rule rule matching on A when $a : A() // needs to somehow convey to the derivation of A() that it should use a set of values unique to this matching in its derivation then // use $a end rule another rule matching on A when $a : A() // needs to somehow convey to the derivation of A() that it should use a set of values unique to this matching in its derivation then // use $a end I cannot leave it up to the rules deriving A to handle every single case of different values for these parameters--there would be thousands of different combinations of derivations of A and nobody would understand the code. The rules deriving A cannot match on general Parameter objects because the values in the Parameter objects are unique for each matching of A. This would mean that each rule matching A would need a counterpart rule which asserts the parameters/values needed by the rules deriving A. I mean, I suppose this /is/ a solution, but it's obviously difficult to understand design. I guess what I'm looking for is something inbetween a function (with explicitly binded parameters) and a derivation that is rule based--to be able to say derive this kind of object in any way that you can, but use these particular values in its derivation. Perhaps this is not possible. -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3372828.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
How would I go about implementing the first part to your solution? (where you say if it is the 'matching' rule that defines the parameter set you can define the parameters literally in that rule) I am not aware of any syntax that would permit this type of behavior--that would allow me to (in the matching rule's condition) literally define the values the rules deriving A should use, instead of the deriving rule matching it against working memory (ie to indicate from the matching rule that the derivation rule should always use a certain value as a condition rather than matching against working memory). Given: rule derivation rule when $prm1 : Param1() then insert( new A($prm1) ); end , the following indicates that working memory should be pattern matched for an A with getParam1() == new Param1(val). Of course, this is not my desired functionality: rule matching rule 1 when A( param1 == new Param1(val) ) then // do w/e end What I would like to see is--given the derivation rule above--something like this: rule matching rule 2 when $a : A( param1 := new Param1(val) ) then // do something with $a end ^ Ideally this would tell the rule derivation rule to attempt to derive A(), but under the assumption that param1 should be initialized with new Param1(val) as opposed to strictly matching for it in WM. However, the more I look at this, more it seems like backward chaining is necessary... -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3373242.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
You insert some A(), and you can't match an A unless it has been inserted and is in working memory. I can't follow you here any more - sometimes things appear to depend on the rule alone; sometimes (as below) an inserted A is a function of a matched Param1(). It's impossible to understand things if they aren't fully put on the table. -W On 27 September 2011 18:55, JohnnyCaimbridge nbo...@nycm.com wrote: How would I go about implementing the first part to your solution? (where you say if it is the 'matching' rule that defines the parameter set you can define the parameters literally in that rule) I am not aware of any syntax that would permit this type of behavior--that would allow me to (in the matching rule's condition) literally define the values the rules deriving A should use, instead of the deriving rule matching it against working memory (ie to indicate from the matching rule that the derivation rule should always use a certain value as a condition rather than matching against working memory). Given: rule derivation rule when $prm1 : Param1() then insert( new A($prm1) ); end , the following indicates that working memory should be pattern matched for an A with getParam1() == new Param1(val). Of course, this is not my desired functionality: rule matching rule 1 when A( param1 == new Param1(val) ) then // do w/e end What I would like to see is--given the derivation rule above--something like this: rule matching rule 2 when $a : A( param1 := new Param1(val) ) then // do something with $a end ^ Ideally this would tell the rule derivation rule to attempt to derive A(), but under the assumption that param1 should be initialized with new Param1(val) as opposed to strictly matching for it in WM. However, the more I look at this, more it seems like backward chaining is necessary... -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3373242.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Partial Unification / Derivation of Facts?
Well, just going back and forth and taking the time to explain my issue / the functionality I had in mind has helped significantly in my own understanding, and at the very least I am now confident in what sort of limitations there are. So, I don't consider this time a waste, and thanks for trying to understand me. -- View this message in context: http://drools.46999.n3.nabble.com/Partial-Unification-Derivation-of-Facts-tp3372546p3373529.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
