RKkit 0.1.0__py3-none-any.whl

rkkit-0.1.0.dist-info/METADATA

@@ -0,0 +1,205 @@
+Metadata-Version: 2.5
+Name: RKkit
+Version: 0.1.0
+Summary: Playing with Runge-Kutta methods
+Project-URL: Homepage, https://github.com/Thierry-Dumont/RKkit
+Project-URL: Issues, https://github.com/Thierry-Dumont/RKkit/issues
+Author-email: Thierry Dumont <tdumont50@math.cnrs.fr>
+License-Expression: MIT
+License-File: LICENSE
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.13
+Provides-Extra: passagemath
+Requires-Dist: passagemath-categories; extra == 'passagemath'
+Requires-Dist: passagemath-combinat; extra == 'passagemath'
+Requires-Dist: passagemath-environment; extra == 'passagemath'
+Requires-Dist: passagemath-flint; extra == 'passagemath'
+Requires-Dist: passagemath-plot; extra == 'passagemath'
+Requires-Dist: passagemath-singular; extra == 'passagemath'
+Requires-Dist: passagemath-standard; extra == 'passagemath'
+Requires-Dist: passagemath-symbolics; extra == 'passagemath'
+Description-Content-Type: text/markdown
+
+
+
+# Playing with Runge-Kutta methods and SageMath. #
+
+## Installation
+
+You can install this package in sage and in passagemath.
+
+With sage do:
+
+`sage -pip install RKkit`
+
+With passagemath:
+
+1- Install passagemath in a virtualenv
+
+2- Once passagemath in installed, activate the virtualenv if necessary
+and then do:
+
+`pip install RKkit`
+
+
+
+ Then if you clone this repository you can 
+  run the python scripts in PythonExamples/
+   with sage or passagemath/sage and, il you use
+  passagemath you can run them with ipython or in a python script.
+  
+  To run the jupyter
+  notebooks of Notebooks/ with passagemath/sage you need to install
+  jupyter or jupyterlab (using `pip install`).
+  
+## Goal of the code:                        
+
+The goal of this code is to verify different properties of a given Runge--Kutta method, defined by it's Butcher array (all the notions used here are defined in the _Bible_, see below).
+
+* With SageMath, it is possible  to make *exact* computations in some
+  useful sets of numbers, and thus we make the hypothesis that the
+  coefficients of the methods live in the set of (real) algebraic
+  numbers (called *AA* in SageMath); this is not a really restrictive
+  hypothesis. But then, the result obtained are  *proofs* (if both my
+  codes and SageMath are correct!). 
+
+* To compute the order of a Runge-Kutta method, one use the so called
+  _rooted_ _trees_ for which we have a SageMath implementation, coded
+  by [Florent
+  Hivert](http://doc.sagemath.org/html/en/reference/combinat/sage/combinat/rooted_tree.html). 
+
+* We also provide a function to compute the Butcher array of a method
+  defined by collocation. A classical application is the set of
+  Gaussian Runge-Kutta methods. 
+
+Some sage/jupyter notebooks are provided: they probably provide  the
+best way to learn and to test this code.
+
+Note that the cost of the computation grows very fast with tne
+number of steps of the method; also, the rooted_tree machinery in
+sage is quite slow.
+
+#### References: ####
+
+* To learn about Ordinary Differential Equations solvers, you should read the
+_Bible_:
+
+1-   Solving Ordinary Differential Equations I, by Hairer, Nørsett,,
+Wanner (HNW),
+
+2-   Solving Ordinary Differential Equations II Stiff and Differential-Algebraic
+         Problems by Hairer and Wanner (HW),
+		 
+3-   Geometric Numerical Integration by Hairer, Lubich and Wanner (HLW).
+
+_If you want to learn and to understand rooted trees and B-series, I
+recommend to start first by reading reference 3. Being the latest
+book, things have become much easier to understand than in reference
+1._   
+
+* If you want to learn _SageMath_, you can read the book _Mathematical Computation
+with Sage_ (which now is available in French, English and German), and
+for which freely available pdf files can be [downloaded
+here](https://members.loria.fr/PZimmermann/sagebook/english.html) and
+[there](http://sagebook.gforge.inria.fr/). 
+
+
+## _Using the code_ 
+
+* __First__, you must define a Runge--Kutta method. To do this you
+  must write a (simple) python _class_: have a look at _formulas.py_
+  in methods/ to undestand what to do.
+
+	The class must derive from the "RungeKutta" class and  include a
+    constructor. 
+
+The constructor must:
+
+
+1- Define the arrays A and B of the Butcher array (the C part is generally
+    not necessary). 
+  
+2- Give a title.
+
+3-  Call the  base RungeKutta class constructor.
+
+Remember that the  "formula.py" file in methods/ give  examples of such classes.
+
+__Be careful:__ The coefficients of A and B *must* be algebraic numbers
+(AA or QQbar) or rational numbers (QQ). This is not absolutely trivial
+since, in sage:
+
+2/3, for example, is directly evaluated as a float (0.66666...). This
+is a Python intrinsic, and cannot be avoided. So as you need to enter
+an algebraic (real) number or a rational, you _must_ write QQ(2/3) or
+AA(2/3) or 2/QQ(3) or 2/AA(3) instead of 2/3. Again, have a look at
+methods/formula.py.
+
+Have a look at "GoodAndBad.ipynb": this notebook shows this more in
+details.
+
+
+* __Then__, the best is to look at the notebook _Example1.ipynb_ and
+  adapt it to your needs. For this, launch sage like this:
+
+`>sage -n jupyter`
+
+or
+
+`>sage -n jupyterlab` 
+
+and then, launch the notebook _Exemple1.ipynb_from Notebooks.
+
+
+
+## Gaussian formulas (and other methods obtained by collocation) ##
+
+Gausian formulae with n steps are obtained by collocation at the roots of the Legendre P polynomials of degree n, shifted from [-1,1] to [0,1].
+
+RKcolloc.colloc computes the Butcher arrays, given collocation points
+in [0,1] and returns a Runge-Kutta method class  (note that
+collocation points are not necessary Gaussian points). 
+
+See the notebook _Gaussian.ipynb._
+
+For the fun, formulas obtained by collocation at roots of Tchebychev
+polynomials are computed in "Tchebychev.ipynb" (they certainly have
+no interesting application).
+
+
+### Notebooks ###
+
+All are Sage/Jupyter notebooks. Launch Sage by typing:
+
+`sage -n jupyter`
+
+or
+
+`sage -n jupyterlab`
+
+if you use passagemath, you can install jupyter in the virtualenv:
+
+`pip install notebook` (once the vitualenv is activated).
+
+1. _Example.ipynb_ :  a tour of the system.
+
+2. _Gaussian.ipynb_:  construct and test some of the Gauss methods.
+
+3. _AllProperties.ipynb_: show how to compute all possible properties of a method.
+
+4. _Test.ipynb_ : just testing that everything works fine on a set of formulas.
+
+5. _RadauByCollocation.ipynb_ : Radau methods are collocation methods. Play with this.
+
+6. _GoodAndBad.ipynb_ : is supposed to show what to do and what not to
+   do when coding a Runge-Kutta formula.
+
+
+#### Implementation: ####
+
+* The code uses a decorator @_persistance to avoid recomputing known properties (which is often expensive).
+
+A precedent version was using  Sage's
+[@lazy_attribute](http://doc.sagemath.org/html/en/reference/misc/sage/misc/lazy_attribute.html)
+decorator, which could be disturbing. 

rkkit-0.1.0.dist-info/RECORD

@@ -0,0 +1,14 @@
+RKkit/README.md,sha256=aS9Rg6g115dKHNwHupiwx1erkboeYxlTFybMMm_WYW8,20
+RKkit/RKExceptions.py,sha256=qKT490ge3FJ6ZRiGUfuBU9J9Xb6oxegWapL-DLoySbU,2184
+RKkit/RKPolutilities.py,sha256=kR4OnYHz0gsIlAlKpoRE6SVzBpLPnQ-GVBeMSmZtKGQ,1458
+RKkit/RKRungeKutta.py,sha256=rsoeD9NyB0vUakwlNGJvK2o_obEQnJ4Lq5toPHCKZC8,1100
+RKkit/RKTrees.py,sha256=t9S3vD1j96VjyO5WRuZmI18iyQNlWL5XMWzQ6pIjiKY,3761
+RKkit/RKcolloc.py,sha256=tbhAO_NinzxUJgA-DaQDRY4f17AwXl-83KIi14TtSuA,2189
+RKkit/RKformula.py,sha256=nEsgoSr4cmJLtAUoIn9TjSEuwQHjLST_WijUtu5XtqI,13867
+RKkit/RKplot.py,sha256=BwO7eEdRYKWuz3YglfqIN5CxbmvkqLO7gZz12bPqB2E,3209
+RKkit/__init__.py,sha256=u87X3NrPWtZKHK4rFTEVJoEBJrNqz-YBSZqW_pvzPys,288
+RKkit/formulas.py,sha256=qYFXT55dmWYTlfVdX9pWxhAc7iKDqLxgLeYEfntfFXM,3124
+rkkit-0.1.0.dist-info/METADATA,sha256=yDsMSzgeLVw9-1f7Dr-x2-YeR5dXyXo8w0anJxzitWU,6881
+rkkit-0.1.0.dist-info/WHEEL,sha256=zOwg4jB6zX2kU910N-cMawjivD6tO8NEWvE12je1bVk,87
+rkkit-0.1.0.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+rkkit-0.1.0.dist-info/RECORD,,

rkkit-0.1.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.32.0
+Root-Is-Purelib: true
+Tag: py3-none-any