This BNF strictly shows the syntax of the query language.  While some
semantic checking is done at the parser level, most of the real semantic
checking goes on in the query engine.

Query                : QExpr
                     | QExpr COMMA_TOKEN Query

QExpr                : QTerm
                     | QTerm OR_TOKEN QExpr

QTerm                : (NOT_TOKEN) QProp (W_OPEN_TOKEN Weight W_CLOSE_TOKEN)
                     | (NOT_TOKEN) QProp (W_OPEN_TOKEN Weight W_CLOSE_TOKEN) AND_TOKEN QTerm

QProp                : QFactor
                     | PROP_TOKEN property QFactor
                     | PROP_REGEX_TOKEN property (EQUAL_TOKEN) REGEX
                     | PROP_NATLANG_TOKEN property QPhrase

QFactor              : QGroup
                     | OPEN_TOKEN Query CLOSE_TOKEN
                     | OP_TOKEN phrase

QGroup               : QPhrase
                     | QPhrase PROX_TOKEN QProp
          
QPhrase              : phrase(FUZZY_TOKEN | FUZ2_TOKEN)
                     | REGEX
                     | QUOTES_TOKEN extended_phrase(FUZZY_TOKEN | FUZ2_TOKEN)
         
property             : phrase
                     | C_OPEN_TOKEN ((phrase COMMA_TOKEN)*phrase) C_CLOSE_TOKEN

Weight               : DIGITS

# Note: the branch taken for parsing a "phrase" depends on the data-type
# of the current property.

phrase               : STRING_PHRASE
                     | INTEGER_PHRASE
                     | FLOAT_PHRASE
                     | BOOL_PHRASE
                     | DATE_PHRASE
                     | CURRENCY_PHRASE
                     | GUID_PHRASE

STRING_PHRASE        : ANY

INTEGER_PHRASE       : (-)DIGITS

FLOAT_PHRASE         : (-)DIGITS(.DIGITS)

BOOL_PHRASE          : T
                     | TRUE
                     | FALSE
                     | ANY          # -- assumed to be FALSE

# yyyy/mm/dd hh:mm:ss

DATE_PHRASE          : -[DIGITS PERIOD]+
                     | DIGITS/DIGITS/DIGITS
                     | DIGITS/DIGITS/DIGITS DIGITS:DIGITS:DIGITS
                     | DIGITS/DIGITS/DIGITS DIGITS:DIGITS:DIGITS:MILLISECONDS

PERIOD               : h | n | s | y | m | d | w

CURRENCY_PHRASE      : DIGITS.DIGITS   # dollars.cents

GUID_PHRASE          : %08lX-%04X-%04X-%02X%02X-%02X%02X%02X%02X%02X%02X

DIGITS               : [0-9]+

ANY                  : any sequence excluding {}!&|~*@#()[],=<>\n\"^$

ANYEXCEPTQUOTE       : any sequence excluding "

# note: the real parsing of REGEX expressions is done in the query engine.
# The parsing that the query parser does is just enough to hand off the
# string to the query engine.  The BNF for this is in the Query Language
# Help page.

REGEX                : QUOTES_TOKEN ANYEXCEPTQUOTE QUOTES_TOKEN
                     | any sequence before SPACE_TOKEN
                     | any sequence before a ) not preceeded by a !

# the XOF_TOKEN after OPERATOR here is only meaningful for vector properties

OP_TOKEN             : OPERATOR (XOF_TOKEN)

OPERATOR             : EQUAL_TOKEN
                     | NOT_EQUAL_TOKEN
                     | GREATER_TOKEN
                     | GREATER_EQUAL_TOKEN
                     | LESS_TOKEN
                     | LESS_EQUAL_TOKEN
                     | XOF_TOKEN

XOF_TOKEN            : ALLOF_TOKEN
                     | SOMEOF_TOKEN

C_OPEN_TOKEN         : {

C_CLOSE_TOKEN        : }
                     
OPEN_TOKEN           : (

CLOSE_TOKEN          : )
                     
AND_TOKEN            : &
                     | and
                     
OR_TOKEN             : |
                     | or
                     
NOT_TOKEN            : !
                     | not
                     
NOT_EQUAL_TOKEN      : !=

SPACE_TOKEN          : ' '
                     
GREATER_TOKEN        : >

GREATER_EQUAL_TOKEN  : >=
                   
LESS_TOKEN           : <

LESS_EQUAL_TOKEN     : <=
                     
PROX_TOKEN           : ~
                     | near
                     
QUOTES_TOKEN         : "
                     
PROP_TOKEN           : @
                     
PROP_REGEX_TOKEN     : #
                     
PROP_NATLANG_TOKEN   : $
                     
EQUAL_TOKEN          : =
                     
FUZZY_TOKEN          : *

FUZ2_TOKEY           : **
                     
W_OPEN_TOKEN         : [

W_CLOSE_TOKEN        : ]
                     
COMMA_TOKEN          : ,
                     
ALLOF_TOKEN          : ^a

SOMEOF_TOKEN         : ^s