Scientific Computing

For those working with real-world inverse problems, P.C. Hansen’s AIRtools and ReguTools are great ways to quickly try out inverse methods.

As in general in Python and Matlab, you need to be sure that your column vector of observations “b” is actually passed into the functions as a column vector. I’ll illustrate the issue by example.

Assume you have ill-conditioned problem Ax = b, with dimensions:

so to use the ReguTools function maxent.m from Python, implementing the Berg Maximum Entropy method, your myinv.py file would look like:

import oct2py
import numpy as np

def myinv(A,b,maxentLambda):
    oc = oct2py.Oct2Py(oned_as='column')
    oc.addpath('ReguTools')
    xhat = oc.maxent(A,b,maxentLambda)
    return xhat

PC Hansen AIRtools and Regutools in Python.

Assuming an OCR’d book or other text file that has line numbers embedded in file neat.py:

1 # cool program
2 import sys
3 def howneat():
4     sys.exit('Thanks for visiting')

Remove the line numbers by typing in Terminal:

perl -pe 's/^[ t]+d+//' neat.py  > neater.py

Regular expression:

^[ t]+d+

\^

beginning of the line

[ t]+

match space or tab (one or more)

d+

match one or more digits

-p

implicitly loop over (read each line of) the file neat.py.

-e

enable PCRE

And we replace with nothing. This leaves alone the indentation of the code (relevant for Python).

I could have used the -i flag to edit in place, but I instead redirected STDOUT to neater.py in case I made a mistake (inputting the wrong file, for example).

For a simple test (a horizontal linear shift) imwarp() was several times faster than imtransform()

bg = imread('myimage.jpg');
[nRow,nCol] = size(bg);

tform = affine2d(T);
RA = imref2d([nCol,nRow],[1 nCol],[1 nRow]);
data(:,:,i) = imwarp(bg, tform, outputView=RA);

tform = maketform('affine',T);
data(:,:,i) = imtransform(bg,tform,...
  Udata=[1 nCol],...
  Vdata=[1 nRow],...
  Xdata=[1 nCol],...
  Ydata=[1 nRow]);

This Python program extracts the LaTeX word count and equation count and several other counts such as floating element count from a LaTeX document set, using the Perl script texcount.

git clone https://github.com/scivision/texcounts.git
pip install -e .

Plot progress:

python texwcpost.py mydissertation.tex

Upload to the SFTP server:

python phdwc.py mydissertation.tex serverAddress username serverDirectory

John Bahcall 1977 paper advocates for the Hubble Space Telescope in accessible language.

The [Hubble] Space Telescope: Out Where the Stars Do Not Twinkle

A favorite vector symbol is a harpoon.

1. in document header

   \usepackage{harpoon}
   \newcommand*{\vect}[1]{\overrightharp{\ensuremath{#1}}}

2. in document body where desired.

   $\vect{x}$

Notes

• The harpoon package defaults to text mode, which will fail when math characters like ell are used in the math mode.
• reference

Manually put a USB HDD that doesn’t automatically spin-down after inactivity to sleep with sg_start

apt install sg3-utils

Determine which device ID by:

lsblk

Assuming the HDD to spin down was at /dev/sdc:

sg_start 0 --pc=3 /dev/sdc

The HDD wakes up automatically when writing/reading a file on it.

We spin down hard drives because HDDs put out substantial heat when spinning, but don’t spin up/down too frequently to avoid excessive wear.

Phoronix Test Suite is a well regarded benchmarking suite that supports most operating systems. This gives the ability to accurately benchmark operating system performance errors, and has been used to find Linux kernel performance regressions.

1. Download and install

2. run tests

   phoronix-test-suite benchmark c-ray
   phoronix-test-suite benchmark hpcc
   phoronix-test-suite benchmark scimark2
   phoronix-test-suite benchmark xonotic

3. List results

   phoronix-test-suite list-saved-results

4. Convert results to PDF

   phoronix-test-suite result-file-to-pdf  myTest

In Libreoffice Calc, to have the first two letters capitalized when the third letter is lowercase:

Tools → Autocorrect → Options → uncheck “Correct TWo INitial CApitals”

Sometimes for large number > 10000, Matlab/Octave axes ticks will appear as exponentials. Fix this to the desired number of digits by:

set(gca,'YTickLabel',sprintf('%7d|',get(gca,'YTick')))

This example is for 7 digits.