Pymaclab

==========================

* Introduced indicator function feature in variable substitution section in order to manage discontinuities in FOCs.

* The magnitude and sign of shock can now be passed to IRF-generating method.

* Certain methods in pymaclab have been refactored and cleaned up considerably, making code more robust.

* Much work done on getting closer to having a pymaclab-to-dynare translator (but not quite there yet).

==========================

* Eliminated some small bugs in parser module as well as in other sections of code. This version is a significant step up from 0.90.1 !

* Implemented some work-around for PyMacLab to install/compile correctly under Enthought EPD on Macintosh.

* Eliminated some bug in model.updaters methods which did not update model.vardic at all when new item was assigned.

* When the PyMacLab library gets loaded, the correct number of CPU cores is now auto-detected and used automatically throughout.

* DSGE models can now be instantiated by passing an entire triple-quoted model file string to the model class, instead of using filepath.

* The DSGE model instance now has object model.template_paramdic which can be passed to pymaclab.modfiles.templates.wheezy_template to generate modfile.

* New steady state solution method introduced. Simply pass sstate="some Python dictionary" to pass values from outside model.

* New steady state solution method introduced. The USE_FOCS option can now also be declared from inside the model file.

* For SS calculations inside the numerical SS section it is now possible to compute smart starting values using symbolic maths, instead of just numbers.

* Inside variable substition section users can now employ new reserved command ALL{} to auto-generate large number of derivative declarations. See docs.

* The library now contains an extensive test suite using nose, just call pymaclab.test() to run it. Makes the library robust.

* Added a couple of experimental models in pymaclab.modfiles.models (Experimental RBC_Romer contributed by David Pugh at Edinburgh University. Thanks!)

* Updated documentation to describe new functionality/features.

* TODO: More DSGE models have to be added and existing ones modified.

* TODO: The documentation is still not complete. Especially the stats module is not explained.

* TODO: May write additional model templates using Jinja2 and Cheetah, instead of using only Wheezy.

* TODO: Use new templating funcionality to implement method which produces Dynare-compatible model files.

==========================

* Differentiation in the variable substitution section is now "smarter" and understands inter-temporal differentiation with special DISCOUNT variable

* Implemented the Christiano-Fitzgerald asymmetric band-pass filter to extract stationary cycle from simulated data.

* Eliminated some small bugs in parser module.

* Made new stats.VAR and stats.FAVAR experimentally available, but this is not documented yet !

* Also included the Fred API with the PyMacLab package to download data from the St. Louis Fed.

* The DSGE model filepath handles have been split into "models.stable.xmodel", "models.development.xmodel" and "models.testing.xmodel"

* The updater methods for dynamic runtime changes to DSGE models to are pretty much complete now.

* Have started adding more API documentation to PyMacLab Sphinx docs.

* The example/test files have been considerably updated. But need to implement Python nose cases next.

* Updated documentation substantially.

* TODO: More DSGE models have to be added and existing ones modified.

* TODO: The documentation is still not complete. Especially the stats module is not explained.

======================

* Re-introduced sympycore as default CAS. It is MUCH faster than sympy. Now comes packaged with PyMacLab.

* Also decided to always package the latest (working and tested) parallel python(pp) library with PyMacLab.

* PyMacLab's parser now also replaces substituted items in the closed form steady state section.

* A new parameter can be passed to pymaclab.newMOD(mesg=True|False) to generate diagnosis print-outs during instantiation.

* Another new parameter is pymaclab.newMOD(ncpus='auto'|INT) where INT is some integer. Number of CPU cores to be used. Default=1.

* Another new parameter is pymaclab.newMOD(mk_hessian=True|False). Should expensive Hessian be computed? Default=True.

* PyMacLab now supports 5 different ways of finding a DSGE model's steady state.

* One of the 5 SS methods uses externally passed arguments pymaclab.newMOD(use_focs=[],ssidic={}) to use FOCS to solve for SS.

* One-off and queued dynamic run-time DSGE model updaters objects have been added

* Improved various DSGE model templates. Monika Merz's unemployment RBC DSGE model is now working.

* Further improvements and additions to the online documentation

=====================

* PyMacLab now only depends on Sympy and no longer the outdated Sympycore.

* Significantly improved documentation, now on its way of being first draft.

* Translated some model files to make use of new functions.

* Deprecated alternative ways of specifying law of motion of exogenous states.

* Added new SS{EXPRESSION} steady-state conversion feature to Variable Substitution section.

* DIFF{EXPRESSION} now also works for steady state expressions w.r.t. steady state variables.

=====================

* Added Sphinx Documentation with API Section.

* Analytical differentiation using DIFF{EXPRESSION,x(t)} now possible in Substitution Section of modfiles.

* Forward and Backward operators FF_X{EXPRESSION} and BB_X{EXPRESSION} for X=1...N now possible in Substitution section of modfiles.