Compressing and Decompressing Files

To decompress tar.gz (tgz) do (tar combines multiple files in a directory to a single file and gz compresses it)

• tar xzvf filename

To compress to tar.gz do

• tar czvf filename.tgz directory/

To decompress bz2 files do

• bunzip2 filename.bz2
• if it is tar.bz2 then do tar -xjvf filename.tar.bz2

To compress into tar.bz2 do

• tar cjvf filename.tar.bz2 directory/

How to link external libraries in C/C++

If you want to link /usr/local/lib/liblbfgs.so and compile quadratic.cpp then you have several options.
(include libbfgs.h in quadratic.cpp first)
g++  -Wall quadratic.cpp /usr/local/lib/liblbfgs.so -o quadratic

Or if this library is on your LD_LIBRARY_PATH, then you can simply do the following
g++ -Wall quadratic -llbfgs -o quadratic

Flusing output in Python

When you use "print" statement to output messages to terminal during the execution of a python program, those messages might not get outputted promptly. This is true for time consuming programs that perform a lot of work in between print statements. To promptly flush the text to terminal invoke the python interpreter with "-u" option. For example, if your program is myprogram.py then do the following.

python -u myprogram.py

Alternatively, if you are using sys.stdout.write to print messages to the terminal, then you can do sys.stdout.flush() right after the sys.stdout.write() to do the same. This works even with file I/O.
file.flush() immediately writes the text that is in buffer. This is particularly useful when you are writing log files.

emacs useful commands

Here are some useful tips from this site.

1. C-s incremental search. Once you have reached the end of the buffer it will perform an overlap search from the start.
2. Use fn+left arrow to move to the end of the current line.
3. Use fn+right arrow to move to the start of the current line.
4. Use fn+up arrow to move to the start of the buffer.
5. Use fn+down arrow to move to the end of the buffer. (These are tested on Aquamacs)
6. M-x grep is extremely useful. It will bring the grep -n -e command to which you must append the filename and the 'string to match' and hit return. The grepped results will be shown in a separate buffer which you can click to move to the appropriate position in the source file.
7. M-x shell opens a remote/local shell on a separate buffer.
8. If a file is in version control then emacs shows that in the bottom buffer bar. (e.g. SVN).
9. C-x v = shows the modifications that you have done since the last commit to the current file.
10. C-x v v will perform the next appropriate version control action to the current file. If it is a commit then you will get a buffer to write a commit note and then hit C-c C-c to make the commit.
11. To enter/exit from full screen mode use: command+shift and return.

bash script to convert encoding of all files in a directory

The following bash script calls iconv on each text file in the input directory and converts them from shift-jis (sjis) to Unicode UTF-8 (utf-8).


#! /bin/bash

inputDir="danu-summaries"
outputDir="danu"

for file in ../$inputDir/*; do
    if [ -f $file ]; then
fname=`echo "$file" | cut -d '/' -f3`
echo $fname
iconv -f sjis -t utf-8 ../$inputDir/$fname > ./$outputDir/$fname
    fi
done

ANOVA and Tukey HSD in MATLAB

To perform ANOVA followed up by Tukey HSD in MATLAB do the following.

1. Create a table where each column represents a different system and each row represents a different domain (data source upon which the evaluations are performed).
2. Save this file in csv and import it into MATLAB using the data import wizard or dlmread function.
3. Now call one way ANOVA as follows.
   [p,t,st] = anova1(dataMatrix)
4. Now run Tukey HSD using st as follows,
   [c,m,h,nms] = multcompare(st)

TexShop

In TexShop to autocomplete a bibliography reference or a label just type some part of the reference and press F5 to get a list of candidates.