Home
Manual
Packages
Global Index
Keywords
Quick Reference
|
/*
STRING.I
String and related convenience functions.
$Id$
(Based on routines contributed by Eric Theibaut.)
*/
/* Copyright (c) 1996. The Regents of the University of California.
All rights reserved. */
/*---------------------------------------------------------------------------
* @(#) string.i: string manipulation and miscellaneous functions for Yorick
* @(#) by Eric THIEBAUT.
*----------------------------------------------------------------------------
* History:
* 02/11/95 by Eric THIEBAUT: added definitions of `scalar()' and
* `is_vector()'.
*/
/* ------------------------------------------------------------------------ */
func gettime (&time)
/* DOCUMENT gettime -- get current time in the form "HH:MM:SS"
*
* SYNOPSIS: time= gettime();
* gettime, time;
*
* HISTORY: October 30, 1995 by Eric THIEBAUT.
*
* SEE ALSO: getdate, parsedate, timestamp.
*/
{ return (time= strpart(timestamp(), 12:19)); }
func getdate (&date)
/* DOCUMENT getdate -- get date of the day in the form "DD/MM/YY"
*
* SYNOPSIS: date= getdate();
* getdate, date;
*
* HISTORY: October 30, 1995 by Eric THIEBAUT.
*
* SEE ALSO: gettime, parsedate, timestamp.
*/
{
local day, month, year;
parsedate, timestamp(), day, month, year;
year-= (year>=2000)? 2000 : 1900;
return (date= swrite(format="%02d/%02d/%02d", day, month, year));
}
func parsedate (timestamp, &day, &month, &year, &hour, &minute, &second)
/* DOCUMENT parsedate -- get numerical version of a timestamp
*
* SYNOPSIS: parsedate, timestamp, day,month,year, hour,minute,second;
* parsedate(timestamp)
*
* HISTORY: October 30, 1995 by Eric THIEBAUT.
*
* SEE ALSO: gettime, getdate, timestamp.
*/
{
dayName= "";
monthName= "";
day= year= hour= minute= second= 0;
sread, timestamp, format="%s%s%d%d:%d:%d%d", dayName, monthName,
day, hour, minute, second, year;
month= (monthName == ["Jan", "Feb", "Mar", "Apr",
"May", "Jun", "Jul", "Aug",
"Sep", "Oct", "Nov", "Dec"]);
month= 13 - sum(month(psum));
return [day, month, year, hour, minute, second];
}
/* ------------------------------------------------------------------------ */
func strtoupper (s)
/* DOCUMENT strtoupper -- convert a string to upper case letters
*
* SYNOPSIS: s2 = strtoupper(s)
*
* HISTORY: October 10, 1995 by Eric THIEBAUT.
*
* SEE ALSO: strtolower
*/
{
c1= 'a'; c2= 'z'; dir= -1; /* parameters for _strcase */
return _strcloop(s, _strcase);
}
func strtolower (s)
/* DOCUMENT strtolower -- convert a string to lower case letters
*
* SYNOPSIS: s2 = strtolower(s)
*
* HISTORY: October 10, 1995 by Eric THIEBAUT.
*
* SEE ALSO: strtoupper
*/
{
c1= 'A'; c2= 'Z'; dir= +1; /* parameters for _strcase */
return _strcloop(s, _strcase);
}
func _strcase (s)
{
/* (DHM) s is string to have case changed */
if (!s) return s;
c= *pointer(s);
if (numberof((i=where(c >= c1))) && numberof((j=where(c(i) <= c2))))
c(i(j))+= dir*('a'-'A');
return string(&c);
}
func strtrim (s, which, blank=)
/* DOCUMENT strtrim(string)
or strtrim(string, which)
or strtrim(string, which, blank=blank)
returns STRING without leading and/or trailing blanks. If STRING is
only made of blanks, return "". If STRING is 0x0, return 0x0.
If WHICH is 1, trim leading blanks (least expensive). If WHICH is 2,
trim trailing blanks (a more costly operation). If WHICH is 3, (the
default) trim both leading and trailing blanks.
If STRING is an array of strings, result has same dimensions.
The BLANK keyword is a string constituted by characters considered
as blanks; by default, BLANK is " \t\n" meaning that spaces,
tabs, and newlines are discarded.
In the BLANK string, "^", "]", and "-" are treated specially:
"]" and "-", if present, should come first in the list to avoid
special treatment, while "^" should not come first.
*/
{
if (is_void(blank) || !strlen(blank)) blank= " \t\n";
if (is_void(which)) which= 3;
blank= "%["+blank+"]";
return _strcloop(s, _strtrim);
}
func _strtrim (s)
{
if (!s) return s;
b= e= "";
if (which&1) sread, s, format=blank, b;
if (which&2) {
c= *pointer(s);
if (numberof(c)<2) return s;
c(1:-1)= c(-1:1:-1);
sread, string(&c), format=blank, e;
}
return strpart(s, 1+strlen(b):-strlen(e));
}
func strchr (s, c, last=)
/* DOCUMENT strchr -- get first/last index of a character in a string
*
* SYNOPSIS: i = strchr(s, c)
* i = strchr(s, c, last=1)
*
* DIAGNOSTIC: returns 0 if character C is not found in string S.
*
* HISTORY: October 27, 1995 by Eric THIEBAUT.
*
* SEE ALSO: strmatch
*/
{
return _strcloop(s, _strchr);
}
func _strchr (ss)
{
/* alter result array data type in _strcloop */
if (structof(s)==string) s= array(0, dimsof(s));
ss= *pointer(ss);
if (numberof((i=where(ss == char(c))))) return (last? i(0) : i(1));
return 0;
}
/* (DHM) On the principal of hiding ugliness, do possible loop
* on string arrays for several operations here. */
func _strcloop (ss, operation)
{
vector= dimsof(ss);
s= array(string, vector); /* avoid clobbering original if array */
vector= vector(1);
for (i=1 ; i<=numberof(ss) ; ++i) {
r= operation(ss(i));
if (vector) s(i)= r;
else s= r; /* work around Yorick bug */
}
return s;
}
/* ------------------------------------------------------------------------ */
func reform (x, ..)
/* DOCUMENT reform(x, dimlist)
* returns array X reshaped according to dimension list DIMLIST.
* In most cases, prefer this to reshape.
* SEE ALSO: array, dimsof
*/
{
dims= [0];
while (more_args()) {
y= next_arg();
if (is_void(y)) continue;
if (!dimsof(y)(1)) y= [1, y];
n= y(1);
grow, dims, y(2:1+n);
dims(1)+= n;
}
if (dims(1)) {
y= array(structof(x), dims);
y(*)= x(*); /* will blow up if lengths differ */
} else {
if (numberof(x)>1) error, "X longer than specified DIMLIST";
y= x(1);
}
return y;
}
func is_scalar (x)
/* DOCUMENT is_scalar(object)
* returns 1 if OBJECT is a scalar, else 0.
* SEE ALSO: is_array, is_func, is_void, is_range, is_struct, is_stream
*/
{ return is_array(x) && !dimsof(x)(1); }
func is_vector (x)
/* DOCUMENT is_vector(object)
* returns 1 if OBJECT is a vector (i.e., OBJECT has a single
* dimension), else 0.
* SEE ALSO: is_array, is_func, is_void, is_range, is_struct, is_stream
*/
{ return is_array(x) && dimsof(x)(1)==1; }
/* ------------------------------------------------------------------------ */
func scalar (x, def, lt=, le=, gt=, ge=, type=, arg=, fn=)
/* DOCUMENT scalar -- get optional scalar parameter
*
* PROTOTYPE
* x = scalar(xarg, xdef, lt=, le=, gt=, ge=, type=, arg=, fn=);
*
* ARGUMENTS
* XARG argument passed to the function.
* XDEF default value for the scalar argument (optional, if not
* specified, then it is guessed that the caller must supply the
* argument).
* KEYWORDS
* GE= to be valid, XARG must be >= GE (optional, only one of GT or GE
* can be used).
* GT= to be valid, XARG must be > GT (optional, only one of GT or GE
* can be used).
* LE= to be valid, XARG must be <= LE (optional, only one of LT or LE
* can be used).
* LT= to be valid, XARG must be < LT (optional, only one of LT or LE
* can be used).
* TYPE= data type of the scalar (optional).
* FN= function name for error messages (optional string).
* ARG= argument name for error messages (optional string).
*
* DESCRIPTION
* Check XARG and return a scalar value (i.e., either XARG converted to TYPE
* if it is not void or XDEF otherwise). If XARG is not within any specified
* bound or if it is not a scalar or if it is void (e.g., not specified) and
* there is no default value XDEF, an error message is written out.
*
* EXAMPLE
* The following function has 2 scalar arguments X and Y, the 1st one is an
* integer (of type long) which must be specified and be strictly greater
* than 22 while the 2nd default to .5 and must be in [0., 1.]:
* func foo(x,y) {
* x= scalar(x, gt=22, type=long, fn="foo", arg="X");
* y= scalar(y, .5, ge=0., le=1., type=double, fn="foo", arg="Y");
* ...
* }
*
* WARNING
* There is no checking of consistency of options.
*
* HISTORY: 29 Sept. 1995 by Eric THIEBAUT. (Modified slightly by DHM)
*/
{
/* Efficiency note (DHM):
This is pretty slow no matter what because of the long argument list.
A faster implementation might be:
check_range(default_value(x, def), lower, upper, flags)
since you could optionally perform the various checks. Of course,
the total number of arguments for a complete test isn't any smaller,
and there would be extra overhead in multiple function calls.
Furthermore, it would be difficult to pass in the "user friendly"
function and argument name options. (The names of the routines in
the current call chain would be a handy thing to make available by
means of a Yorick builtin function call, as would the ability to
have the error function "pop up" some number of levels so it left
the person in dbug mode at the level of the caller of functions
like this one... That still leaves the argument name, though in
principal Yorick can figure that out at runtime, too.)
*/
/* get default x if necessary */
if (is_void(x)) {
if (is_void(def)) _scalar_err, 5;
x= def;
}
/* check that x is indeed scalar */
dims= dimsof(x);
if (is_void(dims) || dims(1)) _scalar_err, 6;
/* convert data type if required (note type could be function too) */
if (!is_void(type)) x= type(x);
/* check that x is in range */
if (!is_void(lt) && x>=lt) _scalar_err, 1, lt;
if (!is_void(le) && x>le) _scalar_err, 2, le;
if (!is_void(gt) && x<=gt) _scalar_err, 3, gt;
if (!is_void(ge) && x<ge) _scalar_err, 4, ge;
return x;
}
func _scalar_err (oops, value)
{
if (is_void(fn)) fn= "";
else fn= fn+": ";
if (is_void(arg)) arg= "argument";
if (oops==5) {
error, fn+"no default value for "+arg;
} else if (oops==6) {
error, fn+arg+" not a scalar value";
} else {
error, fn+arg+" must be "+["<","<=",">",">="](oops)+pr1(value);
}
}
/* ------------------------------------------------------------------------ */
|