A selection of Nick's scripts
Most of these scripts work just fine on most UNIX flavors, including
Apple's OS X. They were
developed and tested mainly under
not-too-old versions of perl and
Most scripts provide help and list their available options when invoked
with the -h option.
None of them is especially smart (and, by sufficiently bad abuse, most of
them can corrupt/destroy your data...), but a few friends and students find
add linearly extrapolated points to kinky-looking 2-column data files
a nifty tool to organize jpg pictures & avi/mpg movies into folders
according to their EXIF data
evaluate the average slope over slices of an extended 2-column data file
make a backup: to start it automatically every morning at 1:09,
put a line in crontab (edit it by crontab -e) such as:
9 1 * * 2-6 $HOME/bin/backup >> /tmp/backup.log 2>&1
converts energy units such as eV, J, cm-1, ...
rm, but far safer...
very handy to go through zillions of data files with linear fits
organize the filename of mp3 files to fit their tags in a regular format
convolve a spectrum with a Gaussian of given width
generate the histogram of a 1-column data file
minimal effort to get some quick approximation to a 1D integration
simple linear interpolation of 2-column data file
taking ω (rad/s) ε1 ε2 (real
& imag parts of dielectric function) in input, outputs real &
imag part of the refractive index, the normal reflectivity, and the
absorption coeff (in m^-1)
extracts a smoothed slope out of a "noisy" x-y data file
extract all local maxima of a function (typically a spectrum, or a time
series) given as a 2-column file
cleunup pointlessly long data files with numerically nearly equal points
fast remote login (
ssh -X), guess-completing the host name
from past hints
replace string1 string2 file(s):
does what its name promises... in any number of files!
evaluate a n-point running average of a 2-columns regularly-spaced data file
generate a sequence of equally-spaced numbers
visualizes a line spectrum
out of a list of energies (rather primitive, but contains smart pieces of
code, like the autoscale, better than a lot of professional software)
compute square difference of columns 2,3,...
Useful (together with integral) e.g. to compare quantitatively 2 functions
search for simultaneous occurrence of several patterns in many files, and
report the matching filenames
- swap, often better/safer than
2 PC's, e.g. a desktop and a laptop: make sure the same files are on both.
Interactive (you control what you copy where).
Based on rsync
cut a x-y-z box, optionally specifying a time interval
generate a .xyz file from a 3-columns numeric text file
extract the coordinates of 1 atom only to plot its trajectory
evaluate and report a specific bond length for all frames
change atomic nature of atoms in regions where the local density is higher
than a given threshold
change atomic nature of atoms depending on their relation to a 2D (xy)
periodic potential function of given lattice parameter
remove a specific atom from a xyz file
in a xyz file, remove atoms which have accidentally bumped into each
other, due to e.g. a merge of 2 xyz files, or to position randomization
(i) evaluate displacements between successive frames or (ii) change atomic
nature of atoms which move more than a given threshold
keep only first and last frame
evaluate the 2D or 3D g(r) correlation function averaging over all frames
and taking a periodic box into consideration
report all positions back into an arbitrary primitive cell
(the converse of xyz_lattice_translate)
replicate the atoms in a xyz file according to a Bravais lattice
(the converse of xyz_in_cell)
amplify displacements (by a user-set factor) relative to the linear
interpolation between the first and the last frame
rotate rigidly a xyz file around an arbitrary axis by an angle
scale the positions by a factor
choose 1 frame out of a loooong xyz file
keep only one frame every n and discard the rest
evaluate the average number density in a sphere (circle) for particles
given in a xyz file
generate statistical info
generate an xyz file stretched by a factor (optionally in a given direction)
translate rigidly a xyz file in any direction
Students should also check useful infos on
how to produce good pdf with LaTeX,
Errors in Technical Writing.
To use these scripts...
fix LaTeX accents according to Italian spelling rules
quick & efficient for converting LaTeX into html;
generate fairly clean and compact html;
useful also for LaTeX to .doc/.docx conversion,
since MS-Word /
can import html documents
- convertbiblio or the more
refined convertbiblio.py: wanna
use BibTeX to sort
\bibitem bibliography automatically, but never quite
dared to put them in the horrible bibtex format?
- latexcontinuo: Happy LaTeXing in 3
windows: editor + xdvi
or evince) + latexcontinuo
lyx and other complicated tools
n columns of data become a formatted LaTeX table, ready to be pasted
into a document
- fire up a command-line terminal;
create the appropriate folder [
mkdir $HOME/bin] (in case it already
exist, no worry, you get an error message, but can do no harm);
save the script "
scriptname" in there, or save it wherever
your browser wants and then move it in there, e.g.
mv ~/Downloads/scriptname ~/bin].
NB! never use cut&paste
from the browser screen to download any script: use the browser "Save As"
function under "File"!
let your computer know it is executable
chmod +x ~/bin/scriptname].
Hopefully, after your next login "
scriptname" becomes a
legitimate command, exactly like
whatever the folder you are currently working in.
If it does not, you need to fix your path, typically by adding an
instruction to your file
The simplest way to add this instruction is executing the command
echo '\nPATH=$HOME/bin:$PATH' >> $HOME/.bashrc .
(Please execute this command once only: repeated executions of this line
produce a dumb sequence of contradictory lines at the end of
In case you accidentally created them, please edit that file and keep only
After your next login your "
scriptname" should execute fine
from the command line.
Whenever it makes sense, these scripts throw their output to stdout,
i.e. output appears in the terminal window.
Their normal use is therefore
scriptname [options] inputfile [moreinputfiles] > outputfile
prunedata file_to_be_pruned.dat > pruned_file.dat
In case you forgot to redirect output with
> and your
terminal happens to be scrolling zillions of lines, then maybe you'd better
press CTRL-C to kill execution, and repeat your command line with an
Most script honor the -h option, which provides a list of acceptable
A note on redirecting output:
command > somefile replaces
whatever content was in
somefile with the output
command. This may lead to the unwarned irreversible
destruction of potentially precious data.
If this makes you feel unsafe, then
test scripts in folders where you keep no precious unduplicated data,
- think twice before pressing
Comments / debugging / patches are welcome!