quantlib-dev — Discussion of QuantLib development

[Quantlib-dev] QuantLib 1.22 released

[Luigi B. <lui...@gm...>]

QuantLib 1.22 has been released and is available for download at <
https://www.quantlib.org/download.shtml>.

The list of changes for this release is at <
https://www.quantlib.org/reference/history.html>.

Please report any problems you have with this release to the QuantLib
mailing list (<qua...@li...>), or open a GitHub
issue at <https://github.com/lballabio/quantlib/issues>.

Re: [Quantlib-dev] Quantlib-SWIG for Go

[Luigi B. <lui...@gm...>]

Hello,
    yes, from a quick look at the SWIG files it seems there's no support
for shared_ptr yet.  There's an issue open from 2015 if you want to try
asking about it: https://github.com/swig/swig/issues/514

Luigi


On Thu, Mar 11, 2021 at 10:57 PM Francois Botha <ig...@gm...> wrote:

> Hi
>
> I was trying to get Quantlib-SWIG working for Go, but it seems the Boost
> bindings are not available in Go. Is that correct?
>
> Thanks
> Francois
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>

[Quantlib-dev] Quantlib-SWIG for Go

[Francois B. <ig...@gm...>]

Hi

I was trying to get Quantlib-SWIG working for Go, but it seems the Boost
bindings are not available in Go. Is that correct?

Thanks
Francois

[Quantlib-dev] QuantLib 1.21 released

[Luigi B. <lui...@gm...>]

Hello everybody,
    QuantLib 1.21 has been released and is available for download at <
https://www.quantlib.org/download.shtml>.

The list of changes for this release is at <
https://www.quantlib.org/reference/history.html>.

Please report any problems you have with this release to the QuantLib
mailing list (<qua...@li...>), or open a GitHub
issue at <https://github.com/lballabio/quantlib/issues>.

Re: [Quantlib-dev] Retrieve number of legs in QuantLib::Swap

[Luigi B. <lui...@gm...>]

You're right, we're missing a method to return that.  May you please open
an issue on GitHub so it doesn't get forgotten?

Thanks,
    Luigi


On Mon, Nov 30, 2020 at 8:37 AM Francois Botha <ig...@gm...> wrote:

> Hi all,
>
> I'd like to retrieve the number of legs in an arbitrary QuantLib::Swap
> instance. I can't find a public method that gives me this. How would one
> achieve this?
>
> thanks
> Francois Botha
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>

[Quantlib-dev] Retrieve number of legs in QuantLib::Swap

[Francois B. <ig...@gm...>]

Hi all,

I'd like to retrieve the number of legs in an arbitrary QuantLib::Swap
instance. I can't find a public method that gives me this. How would one
achieve this?

thanks
Francois Botha

[Quantlib-dev] QuantLib.natvis

[Francois B. <ig...@gm...>]

Hi all,

If you develop in Visual Studio, you might know that it supports custom
views of C++ objects
<https://docs.microsoft.com/en-us/visualstudio/debugger/create-custom-views-of-native-objects?view=vs-2019>
through.natvis
files which aid in the visualization of variables.

With much older versions of Visual Studio, this was achieved through
additions to one's autoesp.dat file. See past discussions about this at
http://quantlib.10058.n7.nabble.com/Debugging-QuantLib-Array-in-msvc-td7565.html
and http://quantlib.10058.n7.nabble.com/Debugging-dates-td15582.html .

I've created a QuantLib.natvis file (attached) to aid in the visualization
of QuantLib::Date and QuantLib::Array . Maybe there are other types that
can be added later. To use this, the file can be dropped in the root of the
QuantLib project in Visual Studio. If this is valuable, I can add a PR for
the file to be included in the QuantLib repo. Alternatively I can build a
Visual Studio extension that wraps the .natvis file and publish the
extension in the Visual Studio Marketplace.

Your feedback is appreciated. Thanks

For those who are unaware, you can install the
https://marketplace.visualstudio.com/items?itemName=ArkadyShapkin.CDebuggerVisualizersforVS2019
extension to aid in the visualization of boost objects.

Screenshots of effects.
*Arrays*:
Before:
[image: image.png]

After:
[image: image.png]

*Dates*:
Before:
[image: image.png]

After:
[image: image.png]

Francois Botha

Re: [Quantlib-dev] Dropping clean price specific functions in QuantlibXL

[Francois B. <ig...@gm...>]

Ok, I'll do that, thanks.

Francois Botha


On Tue, 10 Nov 2020 at 19:11, Eric Ehlers <eri...@re...> wrote:

> Hi Francois,
>
> What about leaving any existing functions in place, and then creating the
> new functions as desired?  This would ensure backward compatibility, and
> the only cost that I can see would be that a user who does not read the
> release notes might not notice the new functions.  There is no feature to
> mark a function as deprecated but words to that effect could be added to
> the doc string.
>
> Regards,
> Eric
> On 11/10/20 10:09 AM, Francois Botha wrote:
>
> Hi Eric and the rest,
>
> Recently I've made some contributions to QuantLib to generalise the prices
> used in bond functions. A new enumeration, Bond::Price::Type was added
> (with values Clean or Dirty) and that was added as a parameter to most (but
> not all yet) bond functions.
>
> I'd like to take this further to QuantLibXL, which already contains a lot
> of clean price based functions, eg. qlBondCleanPriceFromYieldTermStructure
> .
>
> Instead of creating a dirty price version of each, I'd like to completely
> drop the clean price version and replace them with generic version, i.e.
> replace qlBondCleanPriceFromYieldTermStructure
> with qlBondPriceFromYieldTermStructure , which accepts an addition
> parameter for the price type.
>
> Of course this will be quite a breaking change, but I'm asking you to
> consider it as the maintenance of separate clean and dirty price functions
> in QuantlibXL is going to be significant. I'm not sure if one can mark a
> function in QuantlibXL as obsolete, but even so, that sounds like a lot of
> work and I'm unsure that users will even see that the functions are marked
> as obsolete.
>
> Your thoughts, please?
>
> thanks
> Francois Botha
>
>

Re: [Quantlib-dev] Dropping clean price specific functions in QuantlibXL

[Eric E. <eri...@re...>]

Hi Francois,

What about leaving any existing functions in place, and then creating 
the new functions as desired?  This would ensure backward compatibility, 
and the only cost that I can see would be that a user who does not read 
the release notes might not notice the new functions.  There is no 
feature to mark a function as deprecated but words to that effect could 
be added to the doc string.

Regards,
Eric

On 11/10/20 10:09 AM, Francois Botha wrote:
> Hi Eric and the rest,
>
> Recently I've made some contributions to QuantLib to generalise the 
> prices used in bond functions. A new enumeration, Bond::Price::Type 
> was added (with values Clean or Dirty) and that was added as a 
> parameter to most (but not all yet) bond functions.
>
> I'd like to take this further to QuantLibXL, which already contains a 
> lot of clean price based functions, 
> eg. qlBondCleanPriceFromYieldTermStructure .
>
> Instead of creating a dirty price version of each, I'd like to 
> completely drop the clean price version and replace them with generic 
> version, i.e. replace qlBondCleanPriceFromYieldTermStructure 
> with qlBondPriceFromYieldTermStructure , which accepts an addition 
> parameter for the price type.
>
> Of course this will be quite a breaking change, but I'm asking you to 
> consider it as the maintenance of separate clean and dirty price 
> functions in QuantlibXL is going to be significant. I'm not sure if 
> one can mark a function in QuantlibXL as obsolete, but even so, that 
> sounds like a lot of work and I'm unsure that users will even see that 
> the functions are marked as obsolete.
>
> Your thoughts, please?
>
> thanks
> Francois Botha

[Quantlib-dev] Dropping clean price specific functions in QuantlibXL

[Francois B. <ig...@gm...>]

Hi Eric and the rest,

Recently I've made some contributions to QuantLib to generalise the prices
used in bond functions. A new enumeration, Bond::Price::Type was added
(with values Clean or Dirty) and that was added as a parameter to most (but
not all yet) bond functions.

I'd like to take this further to QuantLibXL, which already contains a lot
of clean price based functions, eg. qlBondCleanPriceFromYieldTermStructure
.

Instead of creating a dirty price version of each, I'd like to completely
drop the clean price version and replace them with generic version, i.e.
replace qlBondCleanPriceFromYieldTermStructure
with qlBondPriceFromYieldTermStructure , which accepts an addition
parameter for the price type.

Of course this will be quite a breaking change, but I'm asking you to
consider it as the maintenance of separate clean and dirty price functions
in QuantlibXL is going to be significant. I'm not sure if one can mark a
function in QuantlibXL as obsolete, but even so, that sounds like a lot of
work and I'm unsure that users will even see that the functions are marked
as obsolete.

Your thoughts, please?

thanks
Francois Botha

[Quantlib-dev] QuantLib 1.20 released

[Luigi B. <lui...@gm...>]

Hello everybody,
    QuantLib 1.20 has been released and is available for download at <
https://www.quantlib.org/download.shtml>.

The list of changes for this release is at <
https://www.quantlib.org/reference/history.html>.

Please report any problems you have with this release to the QuantLib
mailing list (<qua...@li...>), or open a GitHub
issue at <https://github.com/lballabio/quantlib/issues>.

Re: [Quantlib-dev] Regulatory curve for pension funds

[Luigi B. <lui...@gm...>]

Hello Marcin,
    sure, go ahead.

Thanks,
    Luigi


On Tue, Sep 29, 2020 at 11:00 PM Marcin Rybacki <mry...@gm...>
wrote:

> Hi Luigi, all,
>
> Prior to making a pull request I thought it might be good to check with
> you whether such a contribution would be considered useful.
> Basically, I am thinking of adding (to the library) the Dutch regulatory
> term structure - used by pension funds to discount liabilities for the
> purposes of reporting their funding ratios. To a certain extent it
> resembles the regulatory curve prescribed by EIOPA (European Insurance and
> Occupational Pensions Authority) i.e. the risk-free interest rate term
> structure, which is based on the Smith Wilson methodology (see the link
> below for reference):
>
>
> https://www.eiopa.europa.eu/tools-and-data/risk-free-interest-rate-term-structures_en
>
> The idea is that until a certain (last liquid) tenor the term structure
> follows the yields implied from the market instruments and afterwards the
> curve is extended based on a number of parameters.
>
> Let me know what you think and whether it is ok to give it a go.
>
> Looking forward to hearing from you.
>
> Kind regards,
> Marcin
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>

[Quantlib-dev] Regulatory curve for pension funds

[Marcin R. <mry...@gm...>]

Hi Luigi, all,

Prior to making a pull request I thought it might be good to check with you
whether such a contribution would be considered useful.
Basically, I am thinking of adding (to the library) the Dutch regulatory
term structure - used by pension funds to discount liabilities for the
purposes of reporting their funding ratios. To a certain extent it
resembles the regulatory curve prescribed by EIOPA (European Insurance and
Occupational Pensions Authority) i.e. the risk-free interest rate term
structure, which is based on the Smith Wilson methodology (see the link
below for reference):

https://www.eiopa.europa.eu/tools-and-data/risk-free-interest-rate-term-structures_en

The idea is that until a certain (last liquid) tenor the term structure
follows the yields implied from the market instruments and afterwards the
curve is extended based on a number of parameters.

Let me know what you think and whether it is ok to give it a go.

Looking forward to hearing from you.

Kind regards,
Marcin

[Quantlib-dev] QuantLib 1.19 released

[Luigi B. <lui...@gm...>]

Hello everybody,
    QuantLib 1.19 has been released and is available for download at <
https://www.quantlib.org/download.shtml>.

The list of changes for this release is at <
https://www.quantlib.org/reference/history.html>.

Please report any problems you have with this release to the QuantLib
mailing list (<qua...@li...>), or open a GitHub
issue at <https://github.com/lballabio/quantlib/issues>.

[Quantlib-dev] Open Source Risk Engine (ORE) 1.8.5.0 released

[Roland L. <rol...@qu...>]

Dear all,

we have published the 5th release of ORE and ORE-SWIG this week, see  opensourcerisk.org <http://opensourcerisk.org/> and github.com/opensourcerisk <http://github.com/opensourcerisk>.

Changes include
Products: Added Inflation Caps/Floors, Forward Bonds, American Commodity Options, reference data support for Bond products
Markets: Extended currency, calendar and index coverage
Term structures: Basic ESTER/SOFR curve building, fitted Bond curve support, more robust CDS curve building, Equity volatility surface stripping from option premia, Equity forward curve stripping from option premia using put/call parity, enhanced Cap/Floor optionlet stripping, extended CDS volatility configuration, added Commodity basis price and average basis price curves
Tests: Extended range of unit tests across the three libraries
Build systems: Discontinued the automake build system maintenance, after introducing CMake in the 4th release
Python/Java language bindings: Migrated ORE SWIG to work with ORE 1.8.5.0 and QuantLib/QuantLib-SWIG 1.18
and various changes listed in the frelease notes (see News.txt or the user guide’s section 2 at https://www.opensourcerisk.org/documentation <https://www.opensourcerisk.org/documentation>) which have inflated the codebase by about 30%  since the last release.
If you have a chance, please clone the repositories, build and test – all feedback is welcome!

Best regards,
Roland

[Quantlib-dev] QuantLib-SWIG 1.18.1 released

[Luigi B. <lui...@gm...>]

Hi all,
    QuantLib-SWIG 1.18.1 was released today.   Its source code is identical
to version 1.18, but the wrapper code for this version was generated with
the newly-released SWIG 4.0.2, which (among other things) avoids a memory
leak in the Python wrappers when using vectors of shared_ptr.  The new
Python wrappers are available via "pip install".

Luigi

Re: [Quantlib-dev] Clarification about differential operator used in FdmBlackScholesOp

[Klaus S. <kl...@sp...>]

Hi,

it is the Black-Scholes model but formulated in x = ln(S) instead of S. The geometric Brownian motion becomes a Brownian motion if you apply Ito's lemma for x=ln(S) to the original Black-Scholes SDE.  Ito's lemma gives you the extra term sigma^2/2.

best regards
Klaus

On Donnerstag, 7. Mai 2020 17:56:31 CEST Lew Wei Hao wrote:


Hi,


I read a post from https://www.implementingquantlib.com/2015/06/chapter-8-part-1-of-n-finite.html[1] about how the following differential operator is used for Black Scholes vanilla finite difference method in the FdmBlackScholesOp class.  


Why is the S term not there anymore as in the original Black Scholes PDE? And why is there an extra (sigma^2/2)/dx term that i don't see in a classical Black scholes model?


Thanks,
Wei Hao 



--------
[1] https://www.implementingquantlib.com/2015/06/chapter-8-part-1-of-n-finite.html

[Quantlib-dev] Clarification about differential operator used in FdmBlackScholesOp

[Lew W. H. <lew...@ho...>]

Hi,

I read a post from https://www.implementingquantlib.com/2015/06/chapter-8-part-1-of-n-finite.html about how the following differential operator is used for Black Scholes vanilla finite difference method in the FdmBlackScholesOp class.  [cid:250d83d8-71c3-43fd-b866-dc3e2b557ad7]

Why is the S term not there anymore as in the original Black Scholes PDE? And why is there an extra (sigma^2/2)/dx term that i don't see in a classical Black scholes model?

Thanks,
Wei Hao

Re: [Quantlib-dev] Bootstrap inflation term structure from underlying CPI bonds

[Amine I. <ami...@gm...>]

Hi Francois,

Your logic below is the way I understand the bootstrap algorithm (iterative and on yield term structures as it’s the one I am familiar with). You basically would add any new instrument helpers you want, create their corresponding bootstrap error classes which would calculate the difference between their quoted market values and theoretical model ones, and feed those instruments to the bootstrapping class. 

You could choose any interpolation you deem fit, as long as it is specified in the traits.hpp file of the corresponding asset class. In my opinion, this is the way the bootstrapping algorithm was built and intended for future expansion.

Amine Ifri

> On 27 Mar 2020, at 08:53, Francois Botha <ig...@gm...> wrote:
> 
> 
> Hi all,
> 
> I would like to bootstrap an inflation term structure from underlying CPI bonds, so that when these CPI bonds' coupons and redemption are accrued with the solved inflation term structure and discounted at a given nominal discount curve, the price of the CPI bonds reconcile back to quoted market values.
> 
> It seems that the only classes that implement BootstrapHelper<ZeroInflationTermStructure> currently are ZeroCouponInflationSwapHelper and YearOnYearInflationSwapHelper and I think neither of these are suitable for my needs. 
> 
> So I would have to implement a new helper class that implements BootstrapHelper<ZeroInflationTermStructure> and accepts a CPIBond instance.
> 
> Please confirm whether my understanding is correct or whether I'm wrong.
> 
> I eventually want to produce an inflation term structure that assumes flat forward rates, so I will have to implement corresponding classes based on forwards, analogous to the current zero rate classes.
> 
> Would it be possible to refactor PiecewiseZeroInflationCurve to be a generic PiecewiseInflationCurve (similar to PiecewiseYieldCurve) and accept the trait (zero,discount,forward) as a generic parameter?
> 
> thanks
> Francois Botha
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev

[Quantlib-dev] Bootstrap inflation term structure from underlying CPI bonds

[Francois B. <ig...@gm...>]

Hi all,

I would like to bootstrap an inflation term structure from underlying CPI
bonds, so that when these CPI bonds' coupons and redemption are accrued
with the solved inflation term structure and discounted at a given nominal
discount curve, the price of the CPI bonds reconcile back to quoted market
values.

It seems that the only classes that
implement BootstrapHelper<ZeroInflationTermStructure> currently
are ZeroCouponInflationSwapHelper and YearOnYearInflationSwapHelper and I
think neither of these are suitable for my needs.

So I would have to implement a new helper class that implements
BootstrapHelper<ZeroInflationTermStructure> and accepts a CPIBond instance.

Please confirm whether my understanding is correct or whether I'm wrong.

I eventually want to produce an inflation term structure that assumes flat
forward rates, so I will have to implement corresponding classes based on
forwards, analogous to the current zero rate classes.

Would it be possible to refactor PiecewiseZeroInflationCurve to be a
generic PiecewiseInflationCurve (similar to PiecewiseYieldCurve) and accept
the trait (zero,discount,forward) as a generic parameter?

thanks
Francois Botha

[Quantlib-dev] QuantLib 1.18 released

[Luigi B. <lui...@gm...>]

Hello everybody,
    QuantLib 1.18 has been released and is available for download at <
http://quantlib.org/download.shtml>.

The list of changes for this release is at <
http://quantlib.org/reference/history.html>.

Please report any problems you have with this release to the QuantLib
mailing list (<qua...@li...>), or open a GitHub
issue at <https://github.com/lballabio/quantlib/issues>.

[Quantlib-dev] QuantLibXL VS2019 support

[Francois B. <ig...@gm...>]

Hi Eric,

I'm looking at adding VS2019 support for QuantlibXL. Looking at past
commits, I notice that we added a lot of new files for a new msvc release:
new .sln, .vcxproj, .vcxproj.filters files etc.

Is that really necessary to duplicate all those files? Would it not be
possible to consolidate all those files and set the required properties,
e.g. PlatformToolset in a separate .props file?

I'm a novice to these kinds of C++ project features, so I hope you have
time to glance at it.

In the meantime, I've started preparing a PR that updates the auto_link
files, the existing .props files and oddly, #include <string> has to be
included in QuantLibAddin/qlo/date.hpp now. In fact, I'm unsure why it
successfully compiled previously, but for VS2019 the compiler couldn't
resolve std::string.

thanks
Francois Botha

Re: [Quantlib-dev] quantlib-benchmark and quantlib-test-suite no longer built?

[Dirk E. <ed...@de...>]

Hi Luigi,

On 3 December 2019 at 17:18, Luigi Ballabio wrote:
| Dirk,
|     Boost deprecated the header-only timer library we were using, so we
| switched to their newer one, which on the one hand is no-longer header only
| and on the other hand also depend on Boost.System.  The result (see <
| https://github.com/lballabio/QuantLib/pull/692/files>) is that configure
| now also looks for libboost_timer and libboost_system and skips the test
| suite and benchmark if it doesn't find them.

Ahhh.  That would explain what I see given how we build in a chroot cleanroom
with only the prescribed components installed. Will add those two.

I had only glanced at configure et al and could not make sense of it.  Thanks
for waving the cluebat my way!

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | ed...@de...

Re: [Quantlib-dev] quantlib-benchmark and quantlib-test-suite no longer built?

[Luigi B. <lui...@gm...>]

Dirk,
    Boost deprecated the header-only timer library we were using, so we
switched to their newer one, which on the one hand is no-longer header only
and on the other hand also depend on Boost.System.  The result (see <
https://github.com/lballabio/QuantLib/pull/692/files>) is that configure
now also looks for libboost_timer and libboost_system and skips the test
suite and benchmark if it doesn't find them.

Luigi


On Tue, Dec 3, 2019 at 4:25 PM Dirk Eddelbuettel <ed...@de...> wrote:

>
> Shame on me for not testing the rc release, but on the first attempt to
> build
> 1.7 for Debian I got stopped in my tracks because the 'nice to have but not
> essential' quantlib-benchmark and quantlib-test-suite binaries are no
> longer
> made.
>
> No change in setup on my, I configure with
>
>         ./configure     --prefix=/usr                           \
>                         --enable-openmp                         \
>                         --enable-intraday                       \
>                         --enable-examples                       \
>                         --enable-benchmark                      \
>                         --enable-shared                         \
>                         --enable-static                         \
>                         --with-gnu-ld                           \
>                         --mandir=$(debtmp)/usr/share/man        \
>                         --build $(arch)
>
> as can be seen here
> https://salsa.debian.org/edd/quantlib/blob/master/debian/rules
>
> Any idea?  Running build two now, trying to expect these two anymore.
>
> Dirk
>
> --
> http://dirk.eddelbuettel.com | @eddelbuettel | ed...@de...
>
>
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>

[Quantlib-dev] quantlib-benchmark and quantlib-test-suite no longer built?

[Dirk E. <ed...@de...>]

Shame on me for not testing the rc release, but on the first attempt to build
1.7 for Debian I got stopped in my tracks because the 'nice to have but not
essential' quantlib-benchmark and quantlib-test-suite binaries are no longer
made.

No change in setup on my, I configure with

	./configure 	--prefix=/usr 				\
			--enable-openmp   			\
			--enable-intraday			\
			--enable-examples			\
			--enable-benchmark			\
			--enable-shared 			\
			--enable-static 			\
			--with-gnu-ld 				\
			--mandir=$(debtmp)/usr/share/man 	\
			--build $(arch)

as can be seen here
https://salsa.debian.org/edd/quantlib/blob/master/debian/rules

Any idea?  Running build two now, trying to expect these two anymore.

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | ed...@de...