《幽兰操》 韩愈

一个君子是能处于不利的环境而保持他的志向和德行操守的

韩愈《幽兰操》：

兰之猗猗，扬扬其香。

不采而佩，于兰何伤。

今天之旋，其曷为然。

我行四方，以日以年。

雪霜贸贸，荠麦之茂。

子如不伤，我不尔觏。

荠麦之茂，荠麦之有。

君子之伤，君子之守。

Python print in the same line successively

The question is like in python code:
for i inrange(1,100):
    print i
And the output is like:
1
2
3
.
.
99
I want it to print in the same line like: 1 2 3 . . . 99
Solution: use ',' to prevent starting from a new line:

for i in range(1,100):
    print i,

Another thing is that if you want to start something from a new line, you should use '\n' in either 'print '\n' ' or '.write('\n')'

git basic

Normally, there is a classification of bare (empty) repository and non-bare repository. The bare repository is serving as a central media to connect any information to the other repositories. And there is always your working repository where you make change and improvment to your files inside.  Whenever you have made some change in your working repository, you should push it to the bare repository. And from your other computers other where or your account on cluster, you should 'git fectch' or 'git reset --hard' to the bare repository on your server.

Here is some basic procedures that one can build your repositories:

• Making a local "non-bare" working repository, add a folder and files inside under git version control:
  1. We start from mkdir an empty folder
  mkdir wire
  cd wire
  git init
  
  2. And then copy the files to wire
  3. Add all the files inside folder 'wire', and git commit them all
  
  git add *
  git commit -a
  # Note, it is very important not to include .pyc, .py~ files otherwise your will get much unnecessary and unexpected trouble in the future
  
  4. Edit git comment to provide some information about the version and descripiton comment. And also you can add user name and email information about it:
     git config --global user.name "Voornaam Achternaam"
     git config --global user.email "aaa@bbb.com"
     git config --global core.editor "my favorate editor" #e.g. vim, nano, less
     
• Make a bare repository, a remote copy of your local working reposiotry:
  
  1. making a remote bare repository using ssh to our server: ssh name@ssh.lorentz.leidenuniv.nl
     mkdir repo/wire.git
     cd repo/wire.git
     git init --bare
     exit
     
  2. Go back to your local working repository, and make remote add, and push your local repository to the remote bare repository wire.git at the server:
  
  git remote add origin name@ssh.lorentz.leidenuniv.nl:repo/wire.git
  git push origin master
  
  3. Note here "origin" is a name you can choose for the remote bare repository, you can name it for whatever you want, and you can check the all the remote repositories name or choose to delete them by using the command below
     
  git remote # check all the name of remote repository
  git remote rm aaa # remove the remote repository whose name is aaa
  
  4. After your git push from you local non-bare repository to the bare repository, you can go to your server and using "gitk --all" (if you have that) to check out the status, or using "git status". You can always use these command to check the status and branch.
• Get a copy of a remote bare repository:
  
  1. If it is from your own computer or server account
  
  git clone /repo/wire.git
  git clone name:ssh.lorent.leidenuniv.nl:repo/wire.git
  
  2. If it is from other computer or server account
  
  git clone ~/otheruser/repo/hie_or_her_project.git
  git clone name:ssh.lorent.leidenuniv.nl:~otheruser/repo/his_or_her_project.git
  
  Note, unless you want to make some change in this repository (but usually people do not make change here), do not first make this folder 'git init'. Here you just copy or fetch it from a remote repository. You will always working and make file changes in you local non-bare working repository, i.e. wire here, and then push the changed files to ~repo/wire.git. You will always keep up changes from wire.git by using 'git fetch'. Make sure you understand this information flow direction once you made some change.
• Keep up with changes from remote repositories:


1. Once you make some change in your local non-bare working repository, you git push it to the bare repository, and to keep up the change at other places from this non-repository you need to:
git fetch origin (or --all)
git reset --hard origin
Or you can delete the folder and make git clone, see above section
2. In order to make sure you have the same version, you go to your copied repository and check

git describe --always

This will return a series number which identify the version of your repository, which is made after each time you made change in your working repository and git commit them. Each version has a unique version series number. So this number here should be coincides with the same number you using 'git describe --always' at your local non-bare working repository.

How to install Numpy and Scipy in Ubuntu without root right

During my research working, I need to install 'matplot' in Ipython; however, the newest version of matplot requires latest version of Numpy and Scipy. Since I am using the university computer, Ubuntu 10.10, and Numpy and Scipy is very old. So I try to manage to update the later version of Numpy and Scipy.

1. Numpy install:

Here we basically follow the instruction of the instruction on the webpage"http://www.hongliangjie.com/2012/04/01/install-numpy-and-scipy-on-centos-without-root-privilege/"

Notice that here you have to install LAPAC and BLAS correctly before you install NUmpy. You can find some instruction about LAPAC and BLAS installation also from the link above.
1. Scipy install:

       Installation of Scipy is a little bit tricky. First, I failed because I cannot manage to change the installation directory. Then with the help of Michael yesterday, he did it.

• Copy the 'site.cfg' to the untared directory file folder of scipy, say in my case 'scipy-0.10.1'. 
• You have to make sure in which directory installed your python library. This is why I failed before. Michael somehow is experienced to that, and he find it out. So open the site.cfg and change the lines under '[Default]':

        [DEFAULT]
      library_dirs = /usr/lib
      include_dirs = /usr/include
      src_dirs = [BLAS]:[LAPACK]

•  Under the folder of 'scipy-0.10.1' in your terminal, you have to make sure to build with either gfortran or g77 complier, in my case I use gfortran, so you type:

       python setup.py build --fcomplier=gfortran

•  Then you can install scipy in your local computer:

        python setup.py install --home=~

Python: multiple sorted with keys

Today I encounter a problem of sorting an array or a list twice, first by index key1, then by index key2.

More specificlly:

ththings  = [("animal", "bear"), ("vehicle", "school bus"),("animal", "duck"), ("plant", "cactus"), ("vehicle", "speed boat")]

And I want it first be arranged by the first index alphabetically, and then within each class, i.e. 'animal', 'vehicle', 'plant', the second index also should be arranged alphabetically like:

[('animal', 'bear'),
 ('animal', 'duck'),
 ('plant', 'cactus'),
 ('vehicle', 'school bus'),
 ('vehicle', 'speed boat')]

The solution to this problem is to use sorted() function twice, but the tricky thing is that, we must sorted the by the second index inside the sorting of the first index:

c = sorted((sorted(things, key=lambda x: x[1])), key=lambda x:x[0])

where x[1] refers to the 2nd column of the array 'things'.

How do I install numpy without root right under Ubuntu

Today I tried to install numpy latest version 1.6.2 on my university computer, with Ubuntu 10.10

• check the version of your numpy using ipython:

     import numpy

     numpy.__version__

• download the latest numpy install package: tar -xvf numpy-1.6.2.tar.gz

-- to be continue soon

nanoscale views: Graphene, part II

nanoscale views: Graphene, part II: One reason that graphene has comparatively remarkable conduction properties is its band structure, and in particular the idea that single-pa...

Good Quantum Number

Now we are going to review a basic concept, i.e. 'Good quantum number'.

From "http://theworldofsmall.blogspot.nl/2009/01/what-are-good-coordinates.html", I quote:

In classical mechanics, we can assign any number of variables to specify the state of a system. For example, for a free particle, its position, momentum, etc,. can be known and can be assigned simultaneously to specify the state of the free particle. But the story of quantum mechanics is different. If the position of a particle at some instant is correctly known, its momentum will be highly/maximally uncertain. This means, we can't assign both the position and momentum to specify the quantum mechanical state of the particle simultaneously. In this sense, the quantities (observables) which can be assigned simultaneously to a quantum system at any time to specify its state at that time are called the good quantum mumbers of the system. A set of good quantum numbers is also called a complete set of commuting observables (C.S.C.O.). These coordiantes commute each other, which means that their measurements can be made simultaneously.

From wikipedia:

In quantum mechanics, given a particular Hamiltonian and an operator with corresponding eigenvalues and eigenvectors given by then the physical quantities are said to be "good quantum numbers" if every eigenvector remains an eigenvector of with the same eigenvalue as time evolves.
Hence, if:
then we require

for all eigenvectors in order to call a good quantum number.
A necessary and sufficient condition for q to be good is that commute with the Hamiltonian . Proof:
Assume . If is an eigenvector of , then we have (by definition) that , and so :

It can be said that a good classical analogue is that good quantum numbers are the equivalent of conserved quantities in classical mechanics, as their values do not change over time.In non-relativistic treatment,l and s are good quantum numbers but in relativistic quantum mechanics they are no longer good quantum numbers as L and S do not commute with H(in Dirac theory). J=L+S is a good quantum number in relativistic quantum mechanics as J commutes with H.
 

This is the first Blog text file in my Google blogger. The aim of this blog is to keep record as much as possible the things that happens during my daily life, research and study, and both rational thinking, argumentation, or emotional feelings.

But among all these, the first import aim is of this blog is to serve as a manuscript of my study and research. I once have a Notebook which I write everything what I have done, what I have learned during the my research and study. However, I find that the paper based the notebook is not continent enough. Some time I need to keep tracking of some information I have written before. Since it is only organized by date, so I find it hard to searching the information. By using the blog here, I wish I can use different labels and classify each blog so that it could be easy for me to search information.

Right now, I have thought about several categories:

Research and study: daily/weekly report (mainly on summarized what I have learned today), tomorrow/next week's plan

Life's thinking: emotional thinking and expressing, rational argument, feeling about life

And on the other hand, these categories can also have different labels, such as: Linux tech, Mathematica tech, Research project Problems and solving, Courses problem set and solutions, Latex learning, Journal paper reading and summarized, Booking reading summarized, etc.