QuantLib › quantlib-users

Exposing GeneralizedHullWhite in QuantLibXL

Classic

List

Threaded

6 messages Options

Hyungseok Hahm

Exposing GeneralizedHullWhite in QuantLibXL

Dear all,

I am working on exposing GeneralizedHullWhite in QuantLibXL.

It was not terribly challenging until this step. Here is the .xml part. The problem is in the red.

<libraryFunction>GeneralizedHullWhite</libraryFunction>

</SupportedPlatforms>

<type>QuantLib::YieldTermStructure</type>

<superType>libToHandle</superType>

<tensorRank>scalar</tensorRank>

<description>YieldTermStructure object ID.</description>

</Parameter>

<type>QuantLib::Date</type>

<tensorRank>vector</tensorRank>

<description>Speed Structure</description>

</Parameter>

<type>QuantLib::Date</type>

<tensorRank>vector</tensorRank>

<description>Vol Structure</description>

</Parameter>

<Parameter name='Speed' default='""'>

<type>QuantLib::Real</type>

<tensorRank>vector</tensorRank>

<description>Speed</description>

</Parameter>

<Parameter name='Vol' default='""'>

<type>QuantLib::Real</type>

<tensorRank>vector</tensorRank>

<description>Vol</description>

</Parameter>

</Parameters>

</ParameterList>

</Constructor>

The corresponding part of GeneralizedHullWhite.hpp in QuantLib is

boost::function<Real (Time)> speed() const;

boost::function<Real (Time)> vol() const;

As a result, when compiling, the following error message pops up.

error C2440: 'initializing' : cannot convert from 'boost::function<Signature>' to 'std::vector<_Ty>'

At this point, what has caused the error looks obvious. In .xml, the type is defined wrong.

<type>QuantLib::Real</type>

Then my question is what is the type definition in QuantLibXL that corresponds to boost::function<Signature> in QuantLib?

If anyone has a clue, please enlighten me. It would be tremendous help!

Thanks!

Hyungseok

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Peter Caspers-4

Re: Exposing GeneralizedHullWhite in QuantLibXL

> The corresponding part of GeneralizedHullWhite.hpp in QuantLib is
>
> boost::function<Real (Time)> speed() const;
> boost::function<Real (Time)> vol() const;

no, it isn't. These are private methods of the class, if I am not
mistaken, not part of the constructor. The
latter reads

GeneralizedHullWhite(
const Handle<YieldTermStructure>& yieldtermStructure,
const std::vector<Date>& speedstructure,
const std::vector<Date>& volstructure,
const std::vector<Real>& speed,
const std::vector<Real>& vol);

so you should check your wrapper code (or post it here for further
clarification).

Depending on what you are planning to do, a look into the sister
folder ql/experimental/models might also
be interesting (well at the moment probably only if you connect to the
current git master, not on a released
version).

Best regards
Peter

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Hyung-Seok Hahm

Re: Exposing GeneralizedHullWhite in QuantLibXL

Hi Peter,

In fact, I moved the methods from private to public. I know it is not the smartest thing to do, but for the sake of practicality. After being sure that the code works, I will make the code prettier.

Thanks for the heads up. I'll look into that then.

Cheers,

Hyungseok

On Thu, Jan 9, 2014 at 7:44 PM, Peter Caspers <[hidden email]> wrote:

> The corresponding part of GeneralizedHullWhite.hpp in QuantLib is
>
> boost::function<Real (Time)> speed() const;
> boost::function<Real (Time)> vol() const;

no, it isn't. These are private methods of the class, if I am not
mistaken, not part of the constructor. The
latter reads

GeneralizedHullWhite(
const Handle<YieldTermStructure>& yieldtermStructure,
const std::vector<Date>& speedstructure,
const std::vector<Date>& volstructure,
const std::vector<Real>& speed,
const std::vector<Real>& vol);

so you should check your wrapper code (or post it here for further
clarification).

Depending on what you are planning to do, a look into the sister
folder ql/experimental/models might also
be interesting (well at the moment probably only if you connect to the
current git master, not on a released
version).

Best regards
Peter

Hyungseok Hahm, PhD, FRM

Quantitative Analyst

Industrial Bank of Korea

Trading Department

Tel : 82-2-729-7081

Fax : 82-2-729-6084

E-mail : [hidden email]

Yahoo Msn : hyungseok.hahm

Google Talk : hs.hahm.ibk[hidden email]

LinkedIn: http://kr.linkedin.com/pub/dir/Hyung-seok/Hahm

Peter Caspers-4

Fwd: Exposing GeneralizedHullWhite in QuantLibXL

Hi Hyungseok,

ok, then the question is how to expose methods to excel with
non-native return type, right ? This is not straightforward, I had the
same question some time ago:

http://quantlib.10058.n7.nabble.com/Return-object-in-QLXL-td8942.html

Probably you have to wrap boost::function into a ql object to follow
the path described in the thread, though (not sure about that, it's
been a time since I worked
with the addin).

Peter

On 10 January 2014 00:48, Hyung-Seok Hahm <[hidden email]> wrote:

> Hi Peter,
>
> In fact, I moved the methods from private to public. I know it is not the
> smartest thing to do, but for the sake of practicality. After being sure
> that the code works, I will make the code prettier.
>
> Thanks for the heads up. I'll look into that then.
>
> Cheers,
>
> Hyungseok
>
>
> On Thu, Jan 9, 2014 at 7:44 PM, Peter Caspers <[hidden email]>
> wrote:
>>
>> > The corresponding part of GeneralizedHullWhite.hpp in QuantLib is
>> >
>> > boost::function<Real (Time)> speed() const;
>> > boost::function<Real (Time)> vol() const;
>>
>> no, it isn't. These are private methods of the class, if I am not
>> mistaken, not part of the constructor. The
>> latter reads
>>
>> GeneralizedHullWhite(
>> const Handle<YieldTermStructure>& yieldtermStructure,
>> const std::vector<Date>& speedstructure,
>> const std::vector<Date>& volstructure,
>> const std::vector<Real>& speed,
>> const std::vector<Real>& vol);
>>
>> so you should check your wrapper code (or post it here for further
>> clarification).
>>
>> Depending on what you are planning to do, a look into the sister
>> folder ql/experimental/models might also
>> be interesting (well at the moment probably only if you connect to the
>> current git master, not on a released
>> version).
>>
>> Best regards
>> Peter
>
>
>
>
> --
>
> Hyungseok Hahm, PhD, FRM
>
> Quantitative Analyst
>
> Industrial Bank of Korea
>
> Trading Department
>
> Tel : 82-2-729-7081
>
> Fax : 82-2-729-6084
>
> E-mail : [hidden email]
>
> Yahoo Msn : hyungseok.hahm
>
> Google Talk : [hidden email]
>
> LinkedIn: http://kr.linkedin.com/pub/dir/Hyung-seok/Hahm

Luigi Ballabio

Re: Exposing GeneralizedHullWhite in QuantLibXL

In reply to this post by Hyung-Seok Hahm

Hi Hyungseok,
it's not that practical if it blocks you :)

If you want to expose the information quickly, you'll be better off
keeping the two methods private and adding something like:

private:
boost::function<Real (Time)> speed_() const;
boost::function<Real (Time)> vol_() const;
public:
Real speed(Time t) const { return speed_()(t); }
Real vol(Time t) const { return vol_()(t); }

The two public methods above should be straightforward to export.
(They won't be fast, though, as they'll recreate the function every
time. As a further step, you could modify the class so that it caches
the two boost::function instances.)

Luigi

On Fri, Jan 10, 2014 at 12:48 AM, Hyung-Seok Hahm
<[hidden email]> wrote:

--
<https://implementingquantlib.blogspot.com>
<https://twitter.com/lballabio>

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Hyung-Seok Hahm

Re: Exposing GeneralizedHullWhite in QuantLibXL

Luigi and Peter,

I am not totally a stranger to expose user-defined data types in QuantLibXL, so I think I can handle it from here. Adding public methods is also a wonderful idea in the sense that it can keep its integrity.

Thank you very much for the insight that you guys have shared.

Cheers!

- Hyungseok

On Fri, Jan 10, 2014 at 5:19 PM, Luigi Ballabio <[hidden email]> wrote:

Hi Hyungseok,
it's not that practical if it blocks you :)

If you want to expose the information quickly, you'll be better off
keeping the two methods private and adding something like:

private:
boost::function<Real (Time)> speed_() const;
boost::function<Real (Time)> vol_() const;
public:
Real speed(Time t) const { return speed_()(t); }
Real vol(Time t) const { return vol_()(t); }

The two public methods above should be straightforward to export.
(They won't be fast, though, as they'll recreate the function every
time. As a further step, you could modify the class so that it caches
the two boost::function instances.)

Luigi

On Fri, Jan 10, 2014 at 12:48 AM, Hyung-Seok Hahm
<[hidden email]> wrote:
> Hi Peter,
>
> In fact, I moved the methods from private to public. I know it is not the
> smartest thing to do, but for the sake of practicality. After being sure
> that the code works, I will make the code prettier.
>
> Thanks for the heads up. I'll look into that then.
>
> Cheers,
>
> Hyungseok
>
>
> On Thu, Jan 9, 2014 at 7:44 PM, Peter Caspers <[hidden email]>
> wrote:
>>
>> > The corresponding part of GeneralizedHullWhite.hpp in QuantLib is
>> >
>> > boost::function<Real (Time)> speed() const;
>> > boost::function<Real (Time)> vol() const;
>>
>> no, it isn't. These are private methods of the class, if I am not
>> mistaken, not part of the constructor. The
>> latter reads
>>
>> GeneralizedHullWhite(
>> const Handle<YieldTermStructure>& yieldtermStructure,
>> const std::vector<Date>& speedstructure,
>> const std::vector<Date>& volstructure,
>> const std::vector<Real>& speed,
>> const std::vector<Real>& vol);
>>
>> so you should check your wrapper code (or post it here for further
>> clarification).
>>
>> Depending on what you are planning to do, a look into the sister
>> folder ql/experimental/models might also
>> be interesting (well at the moment probably only if you connect to the
>> current git master, not on a released
>> version).
>>
>> Best regards
>> Peter
>
>
>
>
> --
>
> Hyungseok Hahm, PhD, FRM
>
> Quantitative Analyst
>
> Industrial Bank of Korea
>
> Trading Department
>
> Tel : 82-2-729-7081
>
> Fax : 82-2-729-6084
>
> E-mail : [hidden email]
>
> Yahoo Msn : hyungseok.hahm
>
> Google Talk : [hidden email]
>
> LinkedIn: http://kr.linkedin.com/pub/dir/Hyung-seok/Hahm
>
>

> ------------------------------------------------------------------------------
> CenturyLink Cloud: The Leader in Enterprise Cloud Services.
> Learn Why More Businesses Are Choosing CenturyLink Cloud For
> Critical Workloads, Development Environments & Everything In Between.
> Get a Quote or Start a Free Trial Today.
> http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
> _______________________________________________
> QuantLib-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>

--
<https://implementingquantlib.blogspot.com>
<https://twitter.com/lballabio>

Hyungseok Hahm, PhD, FRM

Quantitative Analyst

Industrial Bank of Korea

Trading Department

Tel : 82-2-729-7081

Fax : 82-2-729-6084

E-mail : [hidden email]

Yahoo Msn : hyungseok.hahm

Google Talk : hs.hahm.ibk[hidden email]

LinkedIn: http://kr.linkedin.com/pub/dir/Hyung-seok/Hahm