Pattern Matching :

Caravan language has good support for pattern matching. There are two statements that can be used
to test a string against a given pattern. We will first explain the syntax of the two statements
before explaining the pattern rules.

1. Check if a string matches with pattern/s in an if statement

[else]if <string>=~<pattern>

where string is a variable expression
pattern is a quoted string or variable expression

example 1
if myfile(filename)=~"`*`.pdf"
..
endif

example 2.
var mypattern
mypattern(val)="*.cpp"
mypattern(val)="*.h"
if myfile(filename)=~mypattern(val)
...
endif

In the above example if mypattern(val) has multiple values it will check for a match against
any one of them.

2. Check if a string matches a pattern and create an object by capturing the substrings:

var <objname>={<string>}X{<pattern>}:{pname1,..pnameN}

The above statement creates an object of name 'objname' if the 'string' matches the 'pattern' and
captures the required pname1 .. pnameN fields from the string ; The object is not created if the
string does not match.

example 

var x={"test.data 1230 bytes"}X{"`*`%`*`%*"}:{filename,size}

The " ` " symbol is used to demarcate the substring which is to be captured. The first
substring thus captured is assigned to the first propertyname in the namelist and so on. In the pattern " ` "
is invisible to the pattern matching algorithm. 

Here the object x will be created with two properties
x(filename);// "test.data"
x(size);// "1230"

Here again the 'string' and 'pattern' can be a variable expression , a quantity which is generated
at runtime. Unlike in the previous case, only one pattern and one string is expected.

More about patterns:


A pattern is a string in which some symbols are used to denote some rules much like the wild card
characters used in filenames. For example "*.*" matches any string which has '.' in it.
Similarly "?.*"  matches string with one and only character before '.'.

In caravan the patterns can be constructed to parse very complex strings. 

Pattern matching symbols and their usage is illustrated below:

* ; matches any character multiple times
   example:
     *ing  = "walking", "running" etc.
    
? ; matches any character once
    ?ut = "cut", "put", "but" etc..
%; matches space ' ' and tab
    *%*="hello world", "this is string", "1. Bombay" etc.
$ ; matches any lowercase
    $*="abcd", "a" etc.
# ; matches any digit
    #.## = "1.23","1.66" etc.
@ ; matches any upper case
    @@@ = "GOD"
& ; matches any alphabet
    &&& = "God"
^ ; matches white space including line feeds
    *^="hello world\n"

~ ; matches next chararacter or symbol multiple times
   ~%@~$ = "World", "        World" etc.
    
` ; marks the starting and ending of substring which is to be captured;
    ~%`@~$` ="World", "       World"  and captures "World"
\ ; escapes next character from being treated as a special symbol
   \~* ="~anything", \###="#12" 

Other characters are matched exactly (not case sensitive) 
"A*" = "apple", "Apple" etc.


Now we will show an example of how to use the pattern matching and parsing capabilities of
caravan.

The following are two strings which is given by the "dir" command which we will parse using the
script given below.

" 3-02-99   5:56p       252           0  wireserver.wpj"
"12-07-98   9:02p      1060           0  wiretest.cpp"

Code which you can test by copying into a .html file in the template directory:

<caravan>

var string
string(val)=" 3-02-99   5:56p       252           0  wireserver.wpj"
string(val)="12-07-98   9:02p      1060           0  wiretest.cpp"

loop sl (string(val(00))); // start loop "sl"

"Test String : <b>";string(val(sl(count)));"</b><hr>"

var x={string(val(sl(count)))}X{"~%`#~#`-`##`-`##`~%`#~#`:`##``?`~%`#~#`~%`#~#`~%`*`"}:{date,month,year,hour,minute,ampm,size,ea,name}
// parse each string and create an output
if x
        "Pattern has matched: Following is the reformatted string<br>"
        "<table bgcolor=gray><tr><td>"
        x(name);"</td><td>";x(size);"</td><td>";x(month);"/";x(date);"/";x(year);" ";x(hour);":";x(minute);if x(ampm)="a";"AM";elseif x(ampm)="p";"PM";endif;"</td>"
        "</tr></table>"
else
        "Pattern has not matched<br>"
endif
repeat sl 16;// end loop "sl"

</caravan>

Please comment.