Weekend Codes

Django Admin on GAE

Did you ever wanted to run Django’s Admin interface on Google’s App Engine?
Now you can and we have made a live sample so you can clone, try and use it.

Here is what you want to know

This is running thanks to the projects listed below:
djangoappengine
django-nonrel

The real thing, a live sample

https://github.com/stockrt/djangoadminongae
http://djangoadminongae.appspot.com/admin

Starting your fresh project

Here you can find a djangoappengine bootstrap script to start a fresh
project like this one but with the newest possible versions of all deps. The
same script was used to create this sample project.

Resolvi traduzir alguns textos em lote, e para isso utilizei esse módulo chamado pytranslate.

Ele é um wrapper para o Google Translate e permite que você traduza textos com facilidade. O retorno dele sempre será a sentença completa (frase) traduzida.

Utilização:

import pytranslate

print pytranslate.translate('hello', sl='english', tl='portuguese')
print pytranslate.translate('hello', sl='auto', tl='portuguese')
print pytranslate.translate('hallo', sl='auto', tl='portuguese')
print pytranslate.translate('hallo', sl='auto', tl='french')
print pytranslate.translate('Bonjour', sl='auto', tl='dutch')

A instalação é simples:

wget --no-check-certificate https://github.com/stockrt/pytranslate/tarball/master -O pytranslate.tar.gz
tar xzvf pytranslate.tar.gz
cd stockrt-pytranslate*
./setup.py install

Você pode obter o código do pytranslate aqui

Quando estiver pensando em se hospedar em um hotel, ou utilizar redes de cafés, pergunte primeiro o quanto custa o minuto ou diária de acesso à internet.

Normalmente o custo é alto. Se este for o caso, opte pela conexão tarifada por minuto, e use o seguinte truque para economizar alguns trocados:

1. Logue no sistema do hotel normalmente, optanto pela menor tarifa para pouco tempo de utilização;
2. Execute um dos comandos abaixo, ssh ou putty;
3. Faça logoff do sistema de tarifação;
4. Configure seu browser e demais aplicativos para usar proxy SOCKS na porta local escolhida.

Usando linux

Se estiver usando linux no seu notebook, você poderá fazer o seguinte:

ssh -D 9999 user@servidor

Usando windows

Se estiver no windows, utilize o putty:

putty.exe -D 9999 user@servidor

Nota: Você precisa, claro, ter um “servidor” rodando um sshd e uma conta nele :) Talvez o seu linux que fica rodando no desktop, em casa? Ou aquele servidor da faculdade/trabalho?

Outra dica é configurar o putty para manter a sessão sempre aberta, com keepalive (Category / Connection / Seconds between keepalives / 5) ou manter a sessão ativa deixando um comando qualquer rodando no shell, que gere tráfego no terminal:

while true; do date; sleep 5; done

Depois de criado o proxy dinâmico

Agora você pode fazer logoff do sistema de internet tarifado e então apontar o firefox, ou qualquer outra aplicação (uma nova sessão de ssh, por exemplo), para o proxy dinâmico local, que agora ouve na porta local 9999.

Outros posts também descrevem essa técnica.

Com essas medidas você pode economizar no acesso, além de garantir uma navegação muito mais segura na internet, pois tudo será tunelado e codificado pelo ssh.

Enjoy the hacking.

In the post about emulating a browser in python with mechanize I have showed you how to make some basic tricks in the web with python, but I have not showed how to login a site and how to handle a session, with html forms, links and cookies.

Here I will show it all for you, let’s see it.

First, you must install some dependecies:

easy_install BeautifulSoup
easy_install html2text

Then, let the code speak:

import mechanize
import cookielib
from BeautifulSoup import BeautifulSoup
import html2text

# Browser
br = mechanize.Browser()

# Cookie Jar
cj = cookielib.LWPCookieJar()
br.set_cookiejar(cj)

# Browser options
br.set_handle_equiv(True)
br.set_handle_gzip(True)
br.set_handle_redirect(True)
br.set_handle_referer(True)
br.set_handle_robots(False)

# Follows refresh 0 but not hangs on refresh > 0
br.set_handle_refresh(mechanize._http.HTTPRefreshProcessor(), max_time=1)

# User-Agent (this is cheating, ok?)
br.addheaders = [('User-agent', 'Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1) Gecko/2008071615 Fedora/3.0.1-1.fc9 Firefox/3.0.1')]

# The site we will navigate into, handling it's session
br.open('http://gmail.com')

# Select the first (index zero) form
br.select_form(nr=0)

# User credentials
br.form['Email'] = 'user'
br.form['Passwd'] = 'password'

# Login
br.submit()

# Filter all links to mail messages in the inbox
all_msg_links = [l for l in br.links(url_regex='\?v=c&th=')]
# Select the first 3 messages
for msg_link in all_msg_links[0:3]:
    print msg_link
    # Open each message
    br.follow_link(msg_link)
    html = br.response().read()
    soup = BeautifulSoup(html)
    # Filter html to only show the message content
    msg = str(soup.findAll('div', attrs={'class': 'msg'})[0])
    # Show raw message content
    print msg
    # Convert html to text, easier to read but can fail if you have intl
    # chars
#   print html2text.html2text(msg)
    print
    # Go back to the Inbox
    br.follow_link(text='Inbox')

# Logout
br.follow_link(text='Sign out')

The basic flow is:

• Open the site and login;
• Session is handled by cookiejar, automatically;
• We list the first 3 mail messages;
• For each mail message, we open it and read it’s contents;
• We go back to the Inbox and to the next mail message;
• All done, we can logoff;
• The first 3 mail messages will have it’s status changed to “read” if you look at it in the gmail web interface.

You may ask, how do I know the name of the form fields to fill?

We can inspect it, before filling:

# Open the site
br.open('http://gmail.com')

# Forms
for f in br.forms():
    print f

The output contains the fields to fill in the form:

...
<TextControl(Email=)>
<PasswordControl(Passwd=)>
...

And the rest you already know.

Why gmail? It was only an example. I know we have libgmail , but again, it was only an example, with login forms and a session to handle :)

Every time (every few minutes?) I was looking at my git versioned projects (all my projects?) I encounter myself in doubt:

Am I at the branch I think I am?

And there was I, issuing a “git branch” command to check it…

That is enough! I said.

Looking at the github guides I found this tip very interesting, so I decided to bettered it, and to publish as a tip here.

It will show in your prompt which is your current branch, when your current work directory is a git initialized one.

To use it, just place this line inside your ~/.bashrc or into /etc/profile.d/git-branch.sh or /etc/bashrc or even /etc/profile, the choice is yours:

# git branch
parse_git_branch() {
    git branch 2> /dev/null | sed -e '/^[^*]/d' -e 's/* \(.*\)/(\1) /'
}
PS1="\$(parse_git_branch)$PS1"

This will give you a prompt like this one:

stockrt@jackbauer ~ $ cd Dropbox/stockrt/git/stockrt.github.com
(master) stockrt@jackbauer ~/Dropbox/stockrt/git/stockrt.github.com $

Here you can see a “normal” prompt and then, when I enter one of my git versioned directories, a “git branchned” prompt.

Way cool.