TIPSTER Architecture Change Request
Title:  Pattern Specification Language									Page 1 of ?
Date Prepared: 16 March 1998										CR No.  10
Priority: Routine														Date Logged
Document Affected: Design
Document Version: 2.3
Paragraphs Affected: Section 8.3 (new) and Appendices B & C
References: 	None
Change Required: New Section under Extraction
Specific Recommendations: Modify Architecture Design document pages by adding/replacing 
affected sections and appendices with new material as provided.
Reason for the Proposed Change:  Under the current Architecture, Extraction has had minimal 
discussion and specification.  One of the problems that had been identified was the lack of a 
method to allow developers to express and exchange rules which identify patterns necessary for 
extraction.  An initial solution to this problem has been the design of a Pattern Specification 
Language.  This Language will become part of the Architecture.
----------------------------------------------------------------------------------------------------
Pending final material from the TWG for the Pattern Specification Language, a prototype YACC 
grammar developed by Mr. Cowie and Mr. Appelt has been placed herein as a placeholder.

%{

/*  TIPSTER Common Pattern Specification Language Grammar
    Syntax by Jim Cowie, CRL, NMSU
    Semantics and minor mods by Doug Appelt, SRI International
    
*/

%}

%union {
  Int32 intval ;
  char* string ;
  void* pointer ;
  short* iarray ;
  struct MatchSpec* mspec ;
  struct KleeneOpGroup* kspec ;
  struct RuleRecord* rspec ;
}

%token <string> SYMBOL
%token RIGHT_ARROW
%token DOUBLE_ARROW
%token LEFT_ANGLE_BRA
%token RIGHT_ANGLE_BRA
%token DOUBLE_LEFT_BRA
%token DOUBLE_RIGHT_BRA
%token LEFT_SET_BRA
%token RIGHT_SET_BRA
%token LEFT_BRA
%token RIGHT_BRA
%token LEFT_SQUARE_BRA
%token RIGHT_SQUARE_BRA
%token AMPERSAND
%token BAR
%token <string> NUMBER
%token MINUS
%token <string> QUOTED_STRING
%token ANY
%token TEMP
%token NORM
%token IF
%token THEN
%token ELSE
%token STAR
%token PLUS
%token QUESTION
%token COLON
%token PLUS_COLON
%token SEMICOLON
%token DOUBLE_SEMICOLON
%token ASSIGN
%token ADD_ASSIGN
%token COMMA
%token STOP
%token EQUAL
%token NOT_EQUAL
%token LESS_OR_EQUAL
%token GT_OR_EQUAL
%token CARAT
%token RULE
%token PHASE
%token PRIORITY
%token TK_INPUT
%token TK_TRUE
%token TK_FALSE
%token ATSIGN

%type <intval> a_c_expression
%type <intval> a_constraint
%type <intval> action
%type <intval> actions
%type <intval> action_exp
%type <intval> anno
%type <intval> anno_type
%type <intval> arbitrary
%type <intval> arglist
%type <intval> assignment
%type <intval> attr_name
%type <intval> basic_pattern_element
%type <mspec> binding
%type <intval> boolean_op
%type <intval> c_expression
%type <intval> constraint
%type <intval> constraint_group
%type <rspec> constraints
%type <intval> decl
%type <intval> declarations
%type <intval> field
%type <intval> function_call
%type <intval> function_name
%type <intval> index
%type <intval> index_expression
%type <intval> index_op
%type <intval> input_token
%type <kspec> kleene_op
%type <intval> macro
%type <intval> macros
%type <intval> macro_header
%type <intval> more_actions
%type <intval> namedecl
%type <intval> param_list
%type <intval> pattern_element
%type <intval> pattern_elements
%type <intval> phase
%type <intval> post_condition
%type <intval> pre_condition
%type <intval> prioritydecl
%type <rspec> rule
%type <intval> rules
%type <intval> something_or_other
%type <intval> test_op
%type <intval> value

%%
phase:
     macros declarations rules

macros:

     | macros macro

macro:
   macro_header DOUBLE_ARROW arbitrary RIGHT_ARROW  arbitrary 
DOUBLE_SEMICOLON

macro_header:
   SYMBOL DOUBLE_LEFT_BRA param_list DOUBLE_RIGHT_BRA

arbitrary:

     | arbitrary something_or_other

something_or_other:
       SYMBOL
     | LEFT_ANGLE_BRA
     | RIGHT_ANGLE_BRA
     | DOUBLE_LEFT_BRA
     | DOUBLE_RIGHT_BRA
     | LEFT_SET_BRA
     | RIGHT_SET_BRA
     | LEFT_BRA
     | RIGHT_BRA
     | LEFT_SQUARE_BRA
     | RIGHT_SQUARE_BRA
     | AMPERSAND
     | BAR
     | NUMBER
     | MINUS
     | QUOTED_STRING
     | ANY
     | TEMP
     | NORM
     | IF
     | THEN
     | ELSE
     | STAR
     | PLUS
     | QUESTION
     | COLON
     | SEMICOLON
     | PLUS_COLON
     | ASSIGN
     | ADD_ASSIGN
     | COMMA
     | STOP
     | EQUAL
     | NOT_EQUAL
     | LESS_OR_EQUAL
     | GT_OR_EQUAL
     | CARAT
     | TK_TRUE
     | TK_FALSE
     | ATSIGN

declarations:

     | declarations decl

decl:
     PHASE COLON SYMBOL
     | input_token  COLON symbol_list

input_token:
    TK_INPUT

symbol_list:
     SYMBOL
     | symbol_list COMMA SYMBOL

param_list:
     SYMBOL
     | param_list SEMICOLON SYMBOL

rules:

     | rules rule
     | rules error rule
     
rule:
     namedecl constraints RIGHT_ARROW actions
  | namedecl prioritydecl constraints RIGHT_ARROW actions
			     
namedecl:
    RULE COLON SYMBOL

prioritydecl:
    PRIORITY COLON NUMBER
     
constraints:
     pre_condition constraint_group post_condition
     
pre_condition:

     | LEFT_ANGLE_BRA constraint_group RIGHT_ANGLE_BRA
     
post_condition:
     | LEFT_ANGLE_BRA constraint_group RIGHT_ANGLE_BRA
     
constraint_group:
     pattern_elements BAR constraint_group
     | pattern_elements
     
pattern_elements:
     pattern_element
    | pattern_element pattern_elements
     
pattern_element:
     basic_pattern_element
     | LEFT_BRA constraint_group RIGHT_BRA kleene_op binding
     | LEFT_BRA constraint_group RIGHT_BRA
     | LEFT_BRA constraint_group RIGHT_BRA kleene_op
     | LEFT_BRA constraint_group RIGHT_BRA binding
     
kleene_op:
     STAR
     | PLUS
     | QUESTION
     
binding:
     index_op index
     
basic_pattern_element:
     LEFT_SET_BRA c_expression RIGHT_SET_BRA  
     | QUOTED_STRING
     | SYMBOL
     | function_name LEFT_SQUARE_BRA RIGHT_SQUARE_BRA
     
c_expression:
     constraint
     | constraint COMMA c_expression
     
index_op:
     COLON
   | PLUS_COLON
     
index:
     NUMBER
     | SYMBOL
     
constraint:
    anno test_op value
    | anno_type

    
anno:
     anno_type STOP attr_name
     
anno_type:
     SYMBOL
    | ANY
     
attr_name:
     SYMBOL
     
test_op:
     EQUAL
     | NOT_EQUAL
     | LEFT_ANGLE_BRA
     | RIGHT_ANGLE_BRA
     | LESS_OR_EQUAL
     | GT_OR_EQUAL
     
value:
     NUMBER
     | QUOTED_STRING
     | SYMBOL
     | TK_TRUE
     | TK_FALSE
     
actions:
     action_exp more_actions
     | action_exp
     
more_actions:
     COMMA action_exp more_actions
     | COMMA action_exp
     
action_exp:
     LEFT_BRA IF a_c_expression THEN actions RIGHT_BRA 
     | LEFT_BRA IF a_c_expression THEN actions ELSE actions 
RIGHT_BRA 
     | action {$$ = $1 ;}
     
a_c_expression:
     a_constraint
     | a_c_expression boolean_op a_constraint
         
boolean_op:
     AMPERSAND
     | BAR
     
a_constraint:
     index_expression test_op value

action:
     assignment
     | function_call
     
assignment:
     index_expression ASSIGN ATSIGN
     | index_expression ASSIGN value
     | index_expression ASSIGN index_expression
     | index_expression ASSIGN function_call
     | index_expression ADD_ASSIGN value
     | index_expression ADD_ASSIGN index_expression
     | index_expression ADD_ASSIGN function_call

index_expression:
     COLON index field
     
field:
     STOP anno_type STOP attr_name
    | STOP anno_type
     
function_call:
     function_name LEFT_SQUARE_BRA arglist RIGHT_SQUARE_BRA

function_name:
     SYMBOL
     
arglist:
     index_expression
     |  CARAT index_expression
     | value
     | value COMMA arglist
     | index_expression COMMA arglist
     | CARAT index_expression COMMA arglist
				       
     
%%