Skip to content

Latest commit

 

History

History
274 lines (228 loc) · 17.6 KB

README.md

File metadata and controls

274 lines (228 loc) · 17.6 KB
Raw

NAG Logo

Content

Examples using the NAG Library for Python

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.

Directory of GitHub examples

Examples that ship with the product

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

How to install the NAG Library for Python

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).

Step 1. Downloading and installing

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.

Step 2. Getting a trial licence

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 NAG 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

  1. product code shown as underlying Library NLL6I271VL. which identifies the licence to request, and

  2. 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.

Step 3. Testing the NAG Library

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.

Running more examples

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.

How to run the Jupyter notebook examples

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.

Installing Jupyter notebook

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.

Running the notebook examples

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

Notebook screenshot

List of Chapters in the NAG Library for Python

The following links take you to the relevant section in the official documentation

Useful links

Back to Top