checking local net connections

$ watch ss -tp

openFOAM GPU

http://issuu.com/vratis/docs/speedit_plugin_to_openfoam_installation_guide

virtualisation note

I have worries abut graphics performance.  Anyway:

https://communities.vmware.com/thread/230428?start=15&tstart=0

https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&ved=0CEQQFjAD&url=http%3A%2F%2Fwww.vmware.com%2Ffiles%2Fpdf%2Ftechpaper%2Fvmware-horizon-view-graphics-acceleration-deployment.pdf&ei=zm4QUvmHHKyn0wW02YDwBA&usg=AFQjCNGAqJQsJ7e-6UNyvDXl5BcpiKjfGw&bvm=bv.50768961,d.d2k&cad=rja

/etc/exports

On update, restart with

$ exportfs -a

nbconvert notes

Changing to A4

Edit latex.py in the nbconvert directory and change, documentclass_options from `11pt,english' to `11pt,english,a4'

Efficient / elegant Python file reading

from ipython.org:

with open('workfile', 'r') as f:
        read_data = f.read()

ipython notebook animations. So far so bad.

Some problems here.

Attempt to use openshot to create video (failed)

Anyway, install openshot.  Openshot got upset with my Lucid installation codecs when I attempted export; claiming libx264 was not installed.  Some reading:  there are two versions, one closed, one open.  I most likely have the open version which is limited.


$ sudo apt-get install libavcodec-extra-52 libavdevice-extra-52 libavfilter-extra-0 libavformat-extra-52 libavutil-extra-49 libpostproc-extra-51 libswscale-extra-0

Then openshot hangs on export.

Attempt command line encoding with libav-tools (fail to install on Lucid)

Try command line:  I need the libav-tools package, not present in Lucid.  Add a personal package providing it:

$ sudo add-apt-repository ppa:kxstudio-team/ppa

$ sudo apt-get update

Still can't install it!  Manually add,

deb http://ppa.launchpad.net/kxstudio-team/ppa/ubuntu lucid main 

deb-src http://ppa.launchpad.net/kxstudio-team/ppa/ubuntu lucid main 

to /etc/apt/sources.list and update: $ sudo apt-get update

Still no joy.  Looks like I need to update my installation.

Workaround:

Attempt command line encoding with ffmpeg (Seg fault)

Older linux:  create movie from images with ffmpeg:

$ ffmpeg -qscale 5 -r 20 -b 9600 -i v9_j58_anim???.png v9_j58.mp4

FFmpeg version SVN-r0.5.9-4:0.5.9-0ubuntu0.10.04.3, Copyright (c) 2000-2009 Fabrice Bellard, et al.

  configuration: --extra-version=4:0.5.9-0ubuntu0.10.04.3 --prefix=/usr --enable-avfilter --enable-avfilter-lavf --enable-vdpau --enable-bzlib --enable-libgsm --enable-libschroedinger --enable-libspeex --enable-libtheora --enable-libvorbis --enable-pthreads --enable-zlib --disable-stripping --disable-vhook --enable-runtime-cpudetect --enable-gpl --enable-postproc --enable-swscale --enable-x11grab --enable-libdc1394 --enable-shared --disable-static

  libavutil     49.15. 0 / 49.15. 0

  libavcodec    52.20. 1 / 52.20. 1

  libavformat   52.31. 0 / 52.31. 0

  libavdevice   52. 1. 0 / 52. 1. 0

  libavfilter    0. 4. 0 /  0. 4. 0

  libswscale     0. 7. 1 /  0. 7. 1

  libpostproc   51. 2. 0 / 51. 2. 0

  built on Jan 24 2013 19:42:59, gcc: 4.4.3

Input #0, image2, from 'v9_j58_anim000.png':

  Duration: 00:00:00.05, start: 0.000000, bitrate: N/A

    Stream #0.0: Video: png, rgb24, 1276x934, 20 tbr, 20 tbn, 20 tbc

Segmentation fault

Pity.  Gave up here.#

Attempt with avconv (not in Lucid)

avconv looks simpler and is available on newer Ubuntus.  Try,

$ avconv -i *.png movie.mp4

or don't in Lucid:  it's not available.

Attempt with openmovieeditor

$ sudo apt-get install openmovieeditor

So far so good.  Navigate to the directory containing nothing but images and click render.

Result:  a non-zeros file, that displays a black mplayer screen.  That's a morning lost.

Attempt with gifsicle

I used to use this during the PhD.  It needs gif images so convert all these pngs to gif

Attempt with recordmydesktop

$ sudo apt-get install recordmydesktop

$ recordmydesktop -x 500 -y 250 --width 800 --height 700 --no-sound --delay 10 output.ogv

A good start.  Capture works and mplayer plays the output.  Now to crop.

Best approach (using xdotool, xmacrorec, screengran, python PIL and GIMP)

recordmydesktop worked but had a habit of leaving a white square on the desktop after closing, and

dropped some frames; and the result didn't look so good.  A better approach might be to write a macro

to iterate through the animation frames, saving them.  Crop the result in python and create the animation in GIMP.

Luckily, PafVu will save animations as a series of .png files, leaving only the cropping and creation of the animation.

If a frequency based animation is needed, use the other tools to create a macro to step through the displaced shapes.

imageRoot = '/home/me/screenshots/animSource/'
imageTarget1 = '/home/me/screenshots/animCroppedSolid/'
imageTarget2 = '/home/me/screenshots/animCroppedTrans/'
fontpath = '/usr/share/fonts/truetype/ubuntu-font-family-0.80/'  # system specific!


a = os.listdir(imageRoot)
a = [i for i in a]
f = reference.logscale(1, 1000, 192)
# It seems that PIL will not load normal fonts like those given away
# by Ubuntu, or those on the system already.  That's damn annoying.
# Seems they have thier own format:  pbm or .pil but I cannot find them.
# No:  truetype fonts do work but the command for use is truetype specific.
font = ImageFont.truetype(fontpath + 'Ubuntu-B.ttf', size=16) # true type
#font = ImageFont.load(fontpath + 'Ubuntu-B.ttf') # true type
# font = ImageFont.truetype("arial.ttf", 15) # truetype

ii = 0
imageTarget = imageTarget1
for i in sorted(a):
    print i, ii, f[int(ii/2)]
    b = Image.open(imageRoot + i)
    c = b.crop((740, 345, 1580, 880)) # use this when happy freqs are correct.
    # c = b  # use this to check f[ii/2] = f
    draw = ImageDraw.Draw(c)
    # draw.line((0,0) + c.size, fill=128) # an example.
    draw.text([300, 10], 'Frequency: %.2f Hz' % f[int(ii/2)], fill=128, font=font)  
    del(draw)
    c.save(imageTarget + i + '.new.png')
    if imageTarget == imageTarget1:
        imageTarget = imageTarget2
    else:
        imageTarget = imageTarget1
    ii += 1

ipython solution control. Create .geo from template, convert to PAFEC, parallel solve, import.

fn = 'v9_j55'
if solve:
    createAnalysisFile(dataRoot + fn + '_template.geo', dataRoot + fn + '.geo', 
                   prepend = 'dia_t = 1.067E-3;\n')
    output = !gmsh -2 {dataRoot + fn + '.geo'}  # this works!
    args = setArgs()
    args['FREQUENCIES.FOR.ANALYSIS'] = ffa(3, 15000, 50000, 192)
    args['restrain'] = [{'Heater':1}]
    convert.gmsh2pafec(dataRoot + fn, case = 66, verb = 0, args = args)
    solveJob(fn)
if loadData:
    res['v9_j55: heaterClamp'] = getResult(dataRoot + fn, Vnode = 68)

ipython 1.0 notes

The new ipython includes many new features:  so many I'll log some here.  IPython amazes me!  I wonder if I could up sections of my notebooks to blogger.....  anyway.  Much of the following is noted directly from the following address:

http://ipython.org/ipython-doc/stable/interactive/tutorial.html

% timeit range(1000)  # run this one line

%% timeit range(1000) # run this cell.
 max(x)

% run -d myprogram.py # step through the program!

% debug   # now works in the notebook!  As does raw_input() !

System commands!

output = !ping www.bbc.co.uk

.. to provide the command with ipython variables, use $ or wrap in {} as follows:

output = !grep -rF $pattern ipython/*

a = 'hello'
output = !grep -rF {a} ipython/*


Grab external python programs, edit and execute!

%load http://matplotlib.sourceforge.net/plot_directive/mpl_examples/mplot3d/contour3d_demo.py

Set high quality plots:  Put this at the start.

% config InlineBackend.figure_format = 'svg'

Undo :   Thank you!

%connect_info

to display connection info for other ipython consoles.

Setting up clients and workers:

worker:
user@worker: $ ipython kernel
[IPKernelApp] To connect to another client on this kernel, use:
[IPKernelApp] --existing kernel-12345.json

client:
user@client $ ipython qtconsole --ssh=worker --existing /path/to/kernel-12345.json

Manual ssh tunnels.  This is interesting even without ipython.

$ ssh :: -f -N

Blogger mangled.  Check the source at
http://ipython.org/ipython-doc/stable/interactive/qtconsole.html
for this.

python documentation / uml diagrams

To read convert.py and create documentation based on comment strings in a pdf:

$ epydoc --pdf convert -o convert_docs

To do something similar with html (including class diagrams)

$ epydoc convert -o convert_docs

connecting to ipython with another instance

%connect_info

or simply, 

$ ipython qtconsole --existing

ipython notebook formatting.

\begin{verbatim}

blah blah

\end{verbatim}

... generates a boxed display in the notebook but nbconvert uses the environment as hoped.

# title

## section

### subsection

#### subsubsection

behaves as expected.

Titles properly centred:

'title here \n good'  worked as hoped.

r'title here \\ bad' will left justify the second line.  Yuk.


tit = 'Effect of coupling PZT to diaphragm at edge only\n'
tit += r'with controlled $f_{0}$. $pcb_{t}=3.2mm$'

works as hoped. :)

linux security

John the Ripper:

$ sudo unshadow /etc/passwd /etc/shadow > mypasswd
$ cat mypasswd
$ john mypasswd

$ john --show mypasswd
0 password hashes cracked, 1 left

$ john --status
guesses: 0  time: 0:00:05:11 (3)  c/s: 456

http://www.openwall.com/john/doc/EXAMPLES.shtml

is useful.

Interrupting and running with $ bg  doesn't seem to work and left a john process and locked file.  Killing the process fixed this.  Running as,

$nohup john mypasswd &

then seemed to work ok but subsequent

$ john --status

returned a static,

guesses: 0  time: 0:00:00:02 0% (2)  c/s: 369

even though one of the machine cores was at 100$ with john using it all.  Don't do this then!  Run without nohup and &.