Ruby: do fewer regexp matches in SensitiveActions

[nickrolfe]

Follow-up to #16306, making the same changes for Ruby.

Before:

[2024-04-24 10:42:04] Evaluated non-recursive predicate _Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965_SensitiveDataHeuristics::HeuristicName__#shared@fd52ac02 in 10567ms (size: 991).
Evaluated relational algebra for predicate _Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965_SensitiveDataHeuristics::HeuristicName__#shared@fd52ac02 with tuple counts:
        3539396  ~4%    {4} r1 = JOIN `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76` WITH `Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965` CARTESIAN PRODUCT OUTPUT Lhs.0, Lhs.1, Rhs.0, Rhs.1
            993  ~2%    {4}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.3,Lhs.1
            993  ~4%    {3}    | SCAN OUTPUT In.0, In.2, In.3
                        return r1

[2024-04-24 11:53:40] Evaluated non-recursive predicate SensitiveActions::BasicSensitiveWrite#35b96d84@defc92e9 in 794ms (size: 474).
Evaluated relational algebra for predicate SensitiveActions::BasicSensitiveWrite#35b96d84@defc92e9 with tuple counts:
         216317  ~0%    {4} r1 = SCAN `SensitiveActions::writesProperty/2#249e0c5f` OUTPUT In.0, _, In.1, _
         216317  ~4%    {2}    | REWRITE WITH Tmp.1 := "^@*", Tmp.3 := "", Out.1 := regexpReplaceAll(In.2,Tmp.1,Tmp.3) KEEPING 2
         214165  ~4%    {2}    | AND NOT `_SensitiveActions::writesProperty/2#249e0c5f#antijoin_rhs`(FIRST 2)
        1499155  ~0%    {4}    | JOIN WITH `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76` CARTESIAN PRODUCT OUTPUT Lhs.0, Lhs.1, Rhs.0, Rhs.1
            474  ~0%    {4}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.1,Lhs.3
            474  ~0%    {2}    | SCAN OUTPUT In.0, In.2
                        return r1

[2024-04-24 12:08:52] Evaluated non-recursive predicate SensitiveActions::BasicSensitiveVariableAccess#56ade80c@1dba36vl in 8106ms (size: 1775).
Evaluated relational algebra for predicate SensitiveActions::BasicSensitiveVariableAccess#56ade80c@1dba36vl with tuple counts:
        1152036  ~0%    {4} r1 = SCAN SensitiveActions::SensitiveVariableAccess#bbd64d3d OUTPUT In.0, In.1, _, _
        1152036  ~2%    {3}    | REWRITE WITH Tmp.2 := "^@*", Tmp.3 := "", Out.2 := regexpReplaceAll(InOut.1,Tmp.2,Tmp.3) KEEPING 3
        1144165  ~1%    {3}    | AND NOT _SensitiveActions::SensitiveVariableAccess#bbd64d3d#antijoin_rhs(FIRST 3)
        8009155  ~0%    {5}    | JOIN WITH `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76` CARTESIAN PRODUCT OUTPUT Lhs.0, Lhs.1, Lhs.2, Rhs.0, Rhs.1
           1775  ~4%    {5}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.2,Lhs.4
           1775  ~2%    {3}    | SCAN OUTPUT In.0, In.3, In.1
                        return r1

After:

[2024-04-24 11:05:49] Evaluated non-recursive predicate _Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965_SensitiveDataHeuristics::HeuristicName__#shared@750c34s0 in 6618ms (size: 912).
Evaluated relational algebra for predicate _Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965_SensitiveDataHeuristics::HeuristicName__#shared@750c34s0 with tuple counts:
             1  ~0%    {1} r1 = AGGREGATE `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76_1_#concat_range`, `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76_1_1_#concat_term` ON In.2, In.3 WITH CONCAT<0 ASC> OUTPUT , Agg.0
        505628  ~5%    {5}    | JOIN WITH `Constant::ConstantValue.isStringlikeValue/1#dispred#ff1ad965` CARTESIAN PRODUCT OUTPUT _, Rhs.0, Rhs.1, Lhs.0, _
        505628  ~4%    {3}    | REWRITE WITH Tmp.0 := "(?:", Tmp.4 := ")", Out.0 := (Tmp.0 ++ In.3 ++ Tmp.4) KEEPING 3
           912  ~2%    {3}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.2,Lhs.0
           912  ~2%    {2}    | SCAN OUTPUT In.1, In.2
                       return r1

[2024-04-24 12:00:43] Evaluated non-recursive predicate _SensitiveActions::writesProperty/2#249e0c5f_SensitiveDataHeuristics::HeuristicNames::maybeSensitive__#shared@9f327ffo in 379ms (size: 478).
Evaluated relational algebra for predicate _SensitiveActions::writesProperty/2#249e0c5f_SensitiveDataHeuristics::HeuristicNames::maybeSensitive__#shared@9f327ffo with tuple counts:
             1  ~0%    {1} r1 = AGGREGATE `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76_1_#concat_range`, `SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76_1_1_#concat_term` ON In.2, In.3 WITH CONCAT<0 ASC> OUTPUT , Agg.0
        216317  ~2%    {6}    | JOIN WITH `SensitiveActions::writesProperty/2#249e0c5f` CARTESIAN PRODUCT OUTPUT _, Rhs.0, _, Lhs.0, Rhs.1, _
        216317  ~0%    {3}    | REWRITE WITH Tmp.0 := "(?:", Tmp.2 := ")", Out.0 := (Tmp.0 ++ In.3 ++ Tmp.2), Tmp.2 := "^@*", Tmp.5 := "", Out.2 := regexpReplaceAll(In.4,Tmp.2,Tmp.5) KEEPING 3
           478  ~0%    {3}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.2,Lhs.0
           478  ~2%    {2}    | SCAN OUTPUT In.1, In.2
                       return r1

[2024-04-24 12:15:32] Evaluated non-recursive predicate _SensitiveActions::SensitiveVariableAccess#bbd64d3d__SensitiveDataHeuristics::HeuristicNames::maybeS__#shared@e7090b16 in 11116ms (size: 1784).
Evaluated relational algebra for predicate _SensitiveActions::SensitiveVariableAccess#bbd64d3d__SensitiveDataHeuristics::HeuristicNames::maybeS__#shared@e7090b16 with tuple counts:
        1152036  ~0%    {5} r1 = JOIN `_SensitiveDataHeuristics::HeuristicNames::maybeSensitiveRegexp/1#d87d0a76_1_#concat_range_SensitiveD__#shared` WITH SensitiveActions::SensitiveVariableAccess#bbd64d3d CARTESIAN PRODUCT OUTPUT Lhs.0, Rhs.0, Rhs.1, _, _
        1152036  ~1%    {4}    | REWRITE WITH Tmp.3 := "^@*", Tmp.4 := "", Out.3 := regexpReplaceAll(InOut.2,Tmp.3,Tmp.4) KEEPING 4
           1784  ~0%    {4}    | JOIN WITH PRIMITIVE regexpMatch#bb ON Lhs.3,Lhs.0
           1784  ~0%    {3}    | SCAN OUTPUT In.1, In.2, In.3
                        return r1

[nickrolfe]

DCA shows stage timing improvements for SensitiveActions on a couple of databases.