Spell checker in Vim

Vim's built in spell checker seems a little easier than externally using aspell. Firstly activate the feature in Vim

:set spell

Choose your language (for example)

:set spelllang=en_gb

You can add words that are spelled incorrectly (goes in something like ~/vim/en.utf-8.add) by simply going to the word that is highlighted and using the key sequence

zg

to add the word. See

:help spell

for more information.

dos2unix on Mac OSX

File conversion from dos to unix files on OSX has always been a bit flaky for me. Handily the tr command in one of two variants does the job:


tr '\r' '\n' < dosfile > unixfile


tr -d '\r' < dosfile > unixfile

Padding for superscript and subscript axis labels in gnuplot

By default gnuplot doesn't align superscripts and subscripts:

T_c^{/Symbol \245}

But, that leaves an ugly space between the symbol and superscript. You can force alignment with an @ character:

T@_c^{/Symbol \245}

 This works fine for standard characters, but for slanted variables, the align character makes the superscript overlap. To fix this we insert a partial before the infty symbol in the superscript. To do this we combine two text effects, blank characters &{a} and font size scaling {/* }:

T@_c^{{/*0.5 &{c}}{/Symbol \245}}

To give the desired result. You can change the factor 0.5 to whatever you need, with 1.0 being full character width.

Bash function to wget from user space in york

This can be placed in .bashrc and can be used to get a file from (your username abc500)

http://www-users.york.ac.uk/~abc500

(this is on unix0 in the web directory.

wgetyork ()
{
if [ -f "$1" ];then
rm "$1";
fi
wget http://www-users.york.ac.uk/~abc500/"$1"
}

it checks for the file because wget will save the file you download as filename.1 if you already have a file called filename. Call it using

wgetyork filename

cropping pdfs from terminal

From terminal you can use pdfcrop, which should be installed on most linux systems. As its terminal based it is easier to script.

example

pdfcrop --margins '5 0 5 10' --clip input.pdf output.pdf

Make tar.gz

tar czvf myarchive.tar.gz myfile

Multiple files in gnuplot

Seems obvious now but you can use paste

e.g.

plot "< paste file1.dat file2.dat" using 1:($2/$4) w points

Searching for tabs with grep

I was trying to search for an integer number, which had tabs after is
i.e grep 300 and it would come up with something like
300 -0.0137598 -0.00228325 0.930859 0.932037 -1.04561e+27 -1.73505e+26 7.07363e+28 7.08259e+28 -0.0129162 -0.00224769 0.874459 0.875568 0.00575037 0.00194563 -0.395375 0.396846 1.46282e+24
450 -0.00449249 -0.00663059 0.858427 0.869783 -3.41386e+26 -5.03861e+26 6.52322e+28 6.60951e+28 -0.00415084 -0.00613002 0.795084 0.805603 0.00124869 0.00187794 -0.257016 0.262097 1.46282e+24
etc....


i.e. where there is 300 within any number. To search for 300 i used
grep 300[[:space:]]

Wireless with Linux

This link shows how it is set up using the gnome GUI network manager

http://www.york.ac.uk/services/cserv/net/wireless/help/naslinuxnew.html

Getting xorg to recognise two GPU's

In Ubuntu 8.10, after installation of the nvidia driver, xorg is unable to decide which card to use. To fix this, login in to a terminal, and then type

sudo lspci grep VGA

On my machine this gave the following output:
02:00.0 VGA compatible controller: nVidia Corporation Device 05e2 (rev a1)
04:00.0 VGA compatible controller: nVidia Corporation Device 05e2 (rev a1)

So to get the xserver working, you need to add the following line oin the device section of /etc/X11/xorg.conf:

Section "Device"
Identifier "Configured Video Device"
Busid "PCI:2:0:0"
Driver "nvidia"
EndSection

Then run:

sudo service gdm restart

to get the xserver working as normal.

Bash extract function

Extract function for bash. Put in .bashrc

function extract() # Handy Extract Program.
{
if [ -f $1 ] ; then
case $1 in
*.tar.bz2) tar xvjf $1 ;;
*.tar.gz) tar xvzf $1 ;;
*.bz2) bunzip2 $1 ;;
*.rar) unrar x $1 ;;
*.gz) gunzip $1 ;;
*.tar) tar xvf $1 ;;
*.tbz2) tar xvjf $1 ;;
*.tgz) tar xvzf $1 ;;
*.zip) unzip $1 ;;
*.Z) uncompress $1 ;;
*.7z) 7z x $1 ;;
*) echo "'$1' cannot be extracted via >extract<" ;;
esac
else
echo "'$1' is not a valid file"
fi
}

epstopdf

It's epstopdf joe not eps2pdf

Fixing backspace on edred vim

To fix the backspace key on the vim on edred add this to .vimrc

set backspace=indent,eol,start

latex figures

To get rid of strange behaviour in latex where it does weird things with figure numbers its has to be done in a strict order.

\begin{figure}
\begin{center}
\includegraphics{...}
\end{center}
\caption{...}
\label{...}
\end{figure}

I think caption can go above \begin{center} but it must be outside of the centre section

Gnuplot PDF+Latex files

To create a pdf file from gnuplot with latex text do the following:

In the plot use the terminal:

set terminal epslatex

Then convert the eps to a pdf file.

Include the figure using

\input{figure.tex}

Printing variables from gnuplot to a file

To output variables such as fitting parameters from gnuplot use.

set fit errorvariables # creates a_err type variables for fitting errors

set print "tmp.dat"

print a,a_err,b,b_err

Labeling figures in Latex

If labeling of figures has gone wrong in latex it's probably because the \label{} must be at the end of the figure block after the includegraphics and caption.

Compiling OpenMPI with Intel Compilers on Ubuntu

Compiling OpenMPI with Intel compilers can be a bit of a nuisance, and so the following seems to work with version 11.074 Intel Compilers, OpenMPI v 2.8.1 and Ubuntu 8.10:

1) Install The Intel Compilers with defaults
2) Put iccvars.sh and ifortvars.sh into /etc/profile.d, and comment out the lines referring to non 64 bit binaries
3) source /etc/profile
4) Unpack the OpenMPI
5) Run ./configure --with-fortran CXX=icc FC=ifort LIBS=-lstdc++
6) type LANG=C
7) make all
8) sudo make all install
9) sudo ldconfig
10) type mpiCC --showme to verify the compiler wrappers are working correctly

Awk Scripting, combining columns

If you have multiple files named filename.< a number > e.g. filename.0, filename.1, filename.2,filename.3 etc then you can take an average of multiple columns using awk as below:

you need to know how many files you have, in this num_files=16, and you need to know the total number of columns per file in this tot_num_cols=13. Here I wanted to average colums 1 and 4. Using the paste command which pastes a number of files to the screen side by side and piping to awk the script will output the average of columns 1 and 4 to a file called anewfile.txt

paste filename.* | awk 'BEGIN{num_col=4;num_files=16;timecol=1;tot_num_cols} {c=0; for (j=0;j < num_files;j++){c+=$(timecol+tot_num_cols*j)}}; s="0;" i="0;i < ="num_files;i++){s+="$(num_col+tot_num_cols*i)};" > | anewfile.txt

Bash number formatting

Use printf as in c. For zero padding add a 0 infront of the field length, eg: %06i

So to delete more files than rm can handle

for i in `seq 0 100`; do file=`printf "%06i" "$i"`; echo $file; done