[ADMB Users] Invoking library qfclib.h

[Rice, Cliff G (DFW)]

I’m trying to use the rounding function described in qfc_sim.cpp. (http://admb-project.org/svn/trunk/contrib/qfclib/qfc_sim.cpp)

I have tried the test file. (http://admb-project.org/svn/trunk/contrib/qfclib/tests/testfunction.tpl)

This file includes the command

#include "qfclib.h" in the GLOBALS_SECTION

 

However, when compiling, I receive the errors:

error: too many arguments to function 'double round(double)' evidently for round(rgamma(2.5,2.5,rnd),3) [line 22]

and

error: cannot convert 'dvector' to 'double' for argument '1' to 'double round(double)' for round(rdirichlet(p,rnd),2) [line 23]

among others.

 

These errors are exactly those produced if the include statement is absent.

 

Judging from statements in qfc_sim.cpp:

  *  5. round your matrix, vector, sclar to some specific decimals for display

  *     in report_section, such as round(dmatrix,4), see round()  \n 

 

Evidently I’m missing something.

Thanks,

Cliff Rice

 

[John Sibert]

What are you trying to do?

Both round(rgamma(2.5,2.5,rnd),3) and round(rdirichlet(p,rnd),2) parse
into two arguments to the function double round(double).


John Sibert
Emeritus Researcher, SOEST
University of Hawaii at Manoa
Honolulu HI (GMT-10)
808-294-3842

Visit the ADMB project http://admb-project.org/

On 05/16/2013 10:43 AM, Rice, Cliff G (DFW) wrote:
>
> I’m trying to use the rounding function described in qfc_sim.cpp

> <http://admb-project.org/svn/trunk/contrib/qfclib/qfc_sim.cpp>.

> (http://admb-project.org/svn/trunk/contrib/qfclib/qfc_sim.cpp)
>
> I have tried the test file

> <http://admb-project.org/svn/trunk/contrib/qfclib/tests/testfunction.tpl>.

> (http://admb-project.org/svn/trunk/contrib/qfclib/tests/testfunction.tpl)
>
> This file includes the command
>
> #include "qfclib.h" in the GLOBALS_SECTION
>
> However, when compiling, I receive the errors:
>
> error: too many arguments to function 'double round(double)' evidently
> for round(rgamma(2.5,2.5,rnd),3) [line 22]
>
> and
>
> error: cannot convert 'dvector' to 'double' for argument '1' to
> 'double round(double)' for round(rdirichlet(p,rnd),2) [line 23]
>
> among others.
>
> These errors are exactly those produced if the include statement is
> absent.
>
> Judging from statements in qfc_sim.cpp:
>
> * 5. round your matrix, vector, sclar to some specific decimals for
> display
>
> * in report_section, such as round(dmatrix,4), see round() \n
>
> Evidently I’m missing something.
>
> Thanks,
>
> Cliff Rice
>
>
>

> _______________________________________________
> Users mailing list
> Us...@admb-project.org
> http://lists.admb-project.org/mailman/listinfo/users


_______________________________________________
Users mailing list
Us...@admb-project.org
http://lists.admb-project.org/mailman/listinfo/users

[CHRIS GRANDIN]

First off, take the #include "qfclib.h" out, it isn't needed. The TPL parser automatically inserts #include <contrib.h> which itself includes qfclib. Once you've done that repost the errors.

Chris

[John Sibert]

Hmm. I'm not sure which OS your are using, but in my world double
round(double) requires only one argument, not two as in your report
section example. It is part of the standard C/C++ math library and is
not an ADMB function.

The " error: cannot convert 'param_init_bounded_number' to 'double' for
argument '1' to 'double round(double)' " can probably be eliminated by

report << "FPSFpD=" << round(value(FPSFpD)) << endl;

The value function returns the double part of the variable, and there is
only one argument to round.

John Sibert
Emeritus Researcher, SOEST
University of Hawaii at Manoa
Honolulu HI (GMT-10)
808-294-3842

Visit the ADMB project http://admb-project.org/

On 05/23/2013 11:10 AM, Rice, Cliff G (DFW) wrote:
> Mssrs. Sibert and Grandin,
>
> I am simply trying to round a parameter value for display in the report section, e.g.,
> Given:
> PARAMETER_SECTION
> init_bounded_number FPSFpD(0.2,0.7)
> .
> .
> .
> REPORT_SECTION
> report<<"FPSFpD="<<round(FPSFpD, 2)<<<<endl;
>
> Produces the error:
> error: cannot convert 'param_init_bounded_number' to 'double' for argument '1' to 'double round(double)'
>
> This is true whether or not I have the include <contrib.h> statement in the TPL.
>
> I discovered you can use round() to round a 'plain' number, but apparently not a bounded number.
> So, I looked for other people's approaches and found the qfclib.h library.
> When I couldn't get the example to work, I quoted it in my question, not my own program, because the example was available for everyone to see. However, the example uses the round function on a nested function - I guess to show how versatile it is, but I'm not interested in what those nested functions.
>
> Thanks,
> Cliff
>
> P.S. I have been trying for the last week to get a message to the Group Owner to make me a member so I can respond directly, but without success (Google keep giving an error). So, I am sending this directly to you as well as cc'ing the group.

>
>
> -----Original Message-----
> From: John Sibert [mailto:sib...@hawaii.edu]
> Sent: Thursday, May 16, 2013 6:08 PM
> To: Rice, Cliff G (DFW)
> Cc: us...@admb-project.org
> Subject: Re: [ADMB Users] Invoking library qfclib.h
>

[John Sibert]

Weird. You are using the horrible old green function reference guide
(aka the dox), last built in 2010.

You should have a look at a more current version at
http://www.admb-project.org/documentation/api/ (which claims to have
been built today).

As far as I know there are no value( ) member functions in class
param_init_bounded_number or any other class that can be called as
FPSFpD.value()

John Sibert
Emeritus Researcher, SOEST
University of Hawaii at Manoa
Honolulu HI (GMT-10)
808-294-3842

Visit the ADMB project http://admb-project.org/

On 05/23/2013 01:43 PM, Rice, Cliff G (DFW) wrote:
>
> Thanks,
>
> round(value(FPSFpD)) worked just fine (except that you have to
> multiply then divide by the number of decimal places you want).
>
> It may be of interest that Chris's suggestion of using FPSFpD.value()
> returns the error:
>
> 'class param_init_bounded_number' has no member named 'value' even
> though value is in the member list for 'class
> param_init_bounded_number'
> <http://admb-foundation.org/documentation/api/default/classparam__init__bounded__number-members.html>
> [http://admb-foundation.org/documentation/api/default/classparam__init__bounded__number-members.html].
>
> Cliff

> > Cc: us...@admb-project.org <mailto:us...@admb-project.org>

> >> Us...@admb-project.org <mailto:Us...@admb-project.org>