- Examples using the nAG Library for Python
- How to install the nAG Library for Python
- How to run the Jupyter notebook examples
- List of Chapters in the nAG Library for Python
- Useful links
This repository contains examples and demonstrations using the nAG Library for Python. The nAG Library for Python contains 1900+ functions spanning many areas of numerical computing and data science.
Designed to work alongside the open source Python packages, Numpy and Scipy, The nAG Library for Python can augment your computational workflow in many areas.
- Correlation and regression analysis
- Curve and surface fitting
- Dimensionality Reduction
- Global Optimization
- Linear Algebra
- Local optimization
- Multivariate methods
- Nearest Correlation Matrices
- Operations Research
- Random number generation
- Roots
- Time series analysis
In addition to those presented here, The nAG Library for Python ships with a set of usage examples. To see them all, run the following command
python -m naginterfaces.library.examples --locate
In this section we illustrate how to install the nAG Library for Python, request a Trial Licence and make sure the Library is working. Details and further information regarding the installation can be found here.
Note Before starting make sure you have access to a host that has Python 3 (3.4 or more recent).
Installing the nAG Library is done using the pip package manager, fire-up a terminal and create a Python 3 virtual environment where to install and test the nAG Library
guest@nag-37:~$ python3 -m venv nag3
guest@nag-37:~$ . nag3/bin/activate
(nag3) guest@nag-37:~$
Now use pip to install the nAG Library for Python
(nag3) guest@nag-37:~$ python -m pip install --extra-index-url https://www.nag.com/downloads/py/naginterfaces_nag naginterfaces
or if you prefer the version of the package that relies on Intel MKL for optimized linear algebra routines, then use
(nag3) guest@nag-37:~$ python -m pip install --extra-index-url https://www.nag.com/downloads/py/naginterfaces_mkl naginterfaces
The output should be similar to
Collecting naginterfaces
Downloading https://www.nag.com/downloads/py/naginterfaces_nag/naginterfaces/naginterfaces-27.1.0.0-py2.py3-none-linux_x86_64.whl (55.8MB)
100% |████████████████████████████████| 55.8MB 21kB/s
Collecting numpy>=1.15 (from naginterfaces)
Downloading https://files.pythonhosted.org/packages/45/b2/6c7545bb7a38754d63048c7696804a0d947328125d81bf12beaa692c3ae3/numpy-1.19.5-cp36-cp36m-manylinux1_x86_64.whl (13.4MB)
100% |████████████████████████████████| 13.4MB 70kB/s
Installing collected packages: numpy, naginterfaces
Successfully installed naginterfaces-27.1.0.0 numpy-1.19.5
The output indicates that the installation was successful.
The next step is to get the licensing info (product code and KUSARI ID) and use it to request a licence. From the same virtual terminal, try
(nag3) guest@nag-37:~$ python -m naginterfaces.kusari
The output should be similar to
The *n*AG Library for Python on this platform uses
underlying Library NLL6I271VL.
This Library has been installed as part of the package
and it requires a valid licence key.
No such key could be validated:
the key may not have been installed correctly or
it may have expired.
The Kusari licence-check utility reports the following:
User: guest
Directory: /home/guest
NAG_KUSARI_FILE=""
File /home/guest/nag.key does not exist
-------------------------------------------------------------------------------
Error: Licence not found; this product requires a key for NLL6I271VL
The above information has been generated on machine nag-37
For information on how to obtain a licence, please see
https://www.nag.com/numeric/py/nagdoc_latest/naginterfaces.kusari.html
KUSARI ID = "ADLXt-adEclJLmvnxlrU2sseteZoo,RopA-Ld"
The two important bits are the
-
product code shown as
underlying Library NLL6I271VL.which identifies the licence to request, and -
KUSARI ID shown as
KUSARI ID = "ADLXt-adEclJLmvnxlrU2sseteZoo,RopA-Ld"which identifies the host you are running the library on.
Note that the product code and KUSARI ID can be different from the previous example.
With these, you are set to contact nAG and request a trial licence.
The trial licence is a plain text chunk similar to
NLL6I271V TRIAL 2021/01/27 "RverXn0Pc-Ib?ctdgF=Wpis2j7I"
Save or copy the text into the file /home/guest/nag.key.
The final step is to make sure the licence is valid and the library is working as expected.
The last step is to make sure the licence was correctly stored and that the nAG Library is working correctly. From the same virtual terminal re-run the Kusari licence module
(nag3) guest@nag-37:~$ python -m naginterfaces.kusari
This time the output should be similar to
Licence available; the required NLL6I271VL licence key for this product is valid
TRIAL licence, 27 days remaining (licence from file)
Now let's try a more interesting example (list of optimization examples)
This command runs the example for the FOAS (First-Order Active set method) solver and minimizes the Rosenbrock 2D function.
(nag3) guest@nag-37:~$ python -m naginterfaces.library.examples.opt.handle_solve_bounds_foas_ex
Should generate an outputsimilar to
Trying:
main()
Expecting:
naginterfaces.library.opt.handle_solve_bounds_foas Python Example Results.
Minimizing a bound-constrained Rosenbrock problem.
E04KF, First order method for bound-constrained problems
...
Status: converged, an optimal solution was found
Value of the objective 4.00000E-02
...
ok
indicating that the example was successfully executed. The source code can be found here.
To display the full list of example source files on disk, but not run them, execute
python -m naginterfaces.library.examples --locate
All examples may be executed sequentially by running
python -m naginterfaces.library.examples
Run python -m naginterfaces.library.examples --help to see any additional usage.
This section briefly illustrates how to setup a host in order to open and run the Jupyter notebooks provided in this repository. Before running the notebooks make sure the nAG Library is installed and working. Before starting, it is advised to read Jupyter's installation page.
To install Jupyter, launch a terminal and activate the virtual environment used to install the nAG Library for Python
guest@nag-37:~$ . nag3/bin/activate
(nag3) guest@nag-37:~$ pip install notebook matplotlib
Collecting notebook
Downloading https://files.pythonhosted.org/packages/74/19/50cd38acf22e33370d01fef764355f1e3517f6e12b4fceb8d434ece4f8fd/notebook-6.2.0-py3-none-any.whl (9.5MB)
100% |████████████████████████████████| 9.5MB 115kB/s
Collecting argon2-cffi (from notebook)
...
Successfully installed jupyter-client-6.1.11 jupyterlab-pygments-0.1.2 ... wcwidth-0.2.5
This indicates that Jupyter and matplotlib were successfully installed. The next section shows how to start the notebok interface and open an example.
To run an example, grab a copy of the notebook of interest and start up the notebook interface.
For example, download the Rosenbrock 2D optimization example notebook rosenbrock2d.ipynb into the current directory
(nag3) guest@nag-37:~$ curl -O https://raw.githubusercontent.com/numericalalgorithmsgroup/NAGPythonExamples/master/local_optimization/FOAS/rosenbrock2d.ipynb
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 61961 100 61961 0 0 382k 0 --:--:-- --:--:-- --:--:-- 382k
and now open it using jupyter-notebook
(nag3) guest@nag-37:~$ jupyter-notebook rosenbrock2d.ipynb
[I 12:24:07.336 NotebookApp] Serving notebooks from local directory: /home/guest
[I 12:24:07.336 NotebookApp] Jupyter Notebook 6.2.0 is running at:
[I 12:24:07.336 NotebookApp] http://localhost:8888/?token=f1836a06799a92f25ef9966439bf3491b2f0960dcb51806d
...
This command will fire-up your web browser and open the rosenbrock2d.ipynb notebook, the window should be similar to
The following links take you to the relevant section in the official documentation
- library.anova - Analysis of Variance
- library.blas - Linear Algebra Support Routines
- library.blast - Further Linear Algebra Support Routines
- library.blgm - Linear Model Specification
- library.complex - Complex Arithmetic
- library.contab - Contingency Table Analysis
- library.correg - Correlation and Regression Analysis
- library.det - Determinants
- library.dot - Inner Products
- library.eigen - Eigenvalues and Eigenvectors
- library.fit - Curve and Surface Fitting
- library.glopt - Global Optimization of a Function
- library.ieee - IEEE Arithmetic
- library.info - Library Identification
- library.inteq - Integral Equations
- library.interp - Interpolation
- library.lapackeig - Least Squares and Eigenvalue Problems
- library.lapacklin - Linear Equations
- library.linsys - Simultaneous Linear Equations
- library.machine - Machine Constants
- library.math - Mathematical Constants
- library.matop - Matrix Operations, Including Inversion
- library.mesh - Mesh Generation
- library.mip - Operations Research
- library.mv - Multivariate Methods
- library.nonpar - Nonparametric Statistics
- library.numdiff - Numerical Differentiation
- library.ode - Ordinary Differential Equations
- library.omp - OpenMP Utilities
- library.opt - Minimizing or Maximizing a Function
- library.orthog - Orthogonalization
- library.pde - Partial Differential Equations
- library.quad - Quadrature
- library.rand - Random Number Generators
- library.rnla - Randomized Numerical Linear Algebra
- library.roots - Roots of One or More Transcendental Equations
- library.smooth - Smoothing in Statistics
- library.sort - Sorting and Searching
- library.sparse - Large Scale Linear Systems
- library.sparseig - Large Scale Eigenproblems
- library.specfun - Approximations of Special Functions
- library.stat - Simple Calculations on Statistical Data
- library.sum - Summation of Series
- library.surviv - Survival Analysis
- library.tsa - Time Series Analysis
- library.univar - Univariate Estimation
- library.wav - Wavelet Transforms
- library.zeros - Zeros of Polynomials